CONF-781046 ^ U.S. Department of Energy October 1978 Luray, Virginia Coordinated by the Lawrence Livermore Laboratory for the Assistant Secretary for Energy Technology Division of Energy Storage Systems Proceedings of the 1978 Mechanical and Magnetic Energy Storage Contractors' Review Meeting
05- v CONF-781046 U.S. Department of Energy October 24-26,1978 Dist. Category UC-94b Coordinated by the Luray, Virginia Lawrence Livermore Laboratory for the Assistant Secretary for Energy Technology Division of Energy Storage Systems Washington, D.C. 20545
Proceedings of the 1978 Mechanical and Magnetic Energy Storage Contractors' Review Meeting
Edited by: G. C. Chang T. M. Barlow
-NOTICE- Thil report was prepared 21 in accotnl of work sponsored by the United Stiles Government- Neither the United Stites nor the United Stites Depsnment of Energy, noi any of their employees, nor uiy of their contractors, subcontractors, or their employees, maKes any warranty, express or Implied, or awimes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. TABLE OF CONTENTS Page PREFACE v OPENING REMARKS - George C. Chang vi WELCOME - James H. Swisher vii FLYWHEELS Thomas M. Barlow, Lawrence Livermore Laboratory, "Mechanical Energy Storage Technology Development for Electric and Hybrid Vehicle Applications" 3 Robert F. McAlevy, Robert F. McAlevy & Associates, "The Impact of Mechanical-Energy-Storage Device Addition on the Performance of Electric Vehicles" 13 Arthur E. Raynard, Garrett-AiResearch, "Advanced Flywheel Energy Storage Unit for a High Power Energy Source for Vehicular Use" 27 Edward L. Lustenader, General Electric Company, "Regenerative Flywheel Energy Storage System" 33 D. W. Rabenhorst, Johns Hopkins University, "Low Cost Flywheel Demon- stration" 43 James A. Rinde, Lawrence Livermore Laboratory, "Materials Progr&m for Fiber Composite Flywheel: 55 R. G. Stone, Lawrence Livermore Laboratory, "The Laminated Disk Flywheel Program: A Rotor Development Project by LLL and G. E. Co 67 B. B. Smith, Union Carbide Corporation, "Flywheel Test Facility" 75 Robert 0. Woods, Sandia Laboratories, "Overview of Component Development... 81 E. David Reedy, Jr., Sandia Laboratories, "Sandia Composite-Rim Flywheel Development" 87 A. Keith Miller, Sandia Laboratories, "Structural Modeling of Thick-Rim Rotor" 93 Mel Baer, Sandia Laboratories, "Aerodynamic Heating of High-Speed Fly- wheels in Low-Density Environments" 99 M. W. Eusepi, Mechanical Technology, Inc., "The Application of Fluid Film Bearings and a Passive Magnetic Suspension to Energy Storage Flywheel s" Ill David B. Eisenhaure, The Charles Stark Draper Laboratories, "Low-Loss Ball Bearings for Flywheel Applications" 123 I. Anwar, The Franklin Institute, "Seal Studies for Advanced Flywheel Systems" 129 Flywheels (Conf d) Francis C. Younger, William M. Brobeck and Associates, "A Composite Flywheel for Vehicle Use" 141 P. Ward Hill, Hercules, Inc., "Progress in Composite Flywheel Development" 155 Donald E. Davis, Rockwell International, "Advanced Composite Flywheel for Vehicle Application" 163 D. L. Satchwell, Garrett-AiResearch, "High-Energy-Density Flywheel".. 171 SOLAR MECHANICAL B. C. Caskey, Sandia Laboratories, "Solar Mechanical Energy Storage Project" 179 David G. Uliman, Union College, "The Band Type Variable Inertia Fly- wheel and Fixed Ratio Power Recirculation Applied To It" 185 Arthur G. Erdman, University of Minnesota, "Cellulosic Flywheels".... 195 John M. Vance, Texas A&M University, "A Concept for Suppression of Non-synchronous Whirl in Flexible Flywheels" 205 Francis C. Younger, William M. Brobeck and Associates, "Conceptual Design of a Flywheel Energy Storage System" 211 Theodore W. Place, Garrett-AiResearch, "Residential Flywheel with Turbine Supply" 227 SUPERCONDUCTING MAGNETIC ENERGY STORAGE R. W. Boom, University of Wisconsin, "Superconductive Diurnal Energy Storage Studies" 237 S. van Sciver, University of Wisconsin, "Recent Component Development Studies for Superconductive Magnetic Energy Storage" 247 R. L. Cresap, Bonneville Power Corporation, "Power System Stability Using Superconducting Magnetic Energy Storage Dynamic Character- istics of the BPA System" 253 Paul C. Krause, Purdue University, "Hybrid Computer Study of a SMES Unit for Damping Power System Oscillations" 263 John D. Rogers, Jr., Los Alamos Scientific Laboratory, "Super- conducting Magnetic Energy Storage" 271 Carl Chowaniec, Westinghouse Electric Corp., "Superconducting Mag- netic Energy Storage for Power System Stability Applications" 283 UNDERGROUND PUMPED HYDROELECTRIC STORAGE Shiu-Wing Tarn, Argonne National Laboratory, "Underground Pumped Hydro Storage -- An Overview" 293 Underground Pumped Hydroelectric Storage (Cont'd) J. Degnan, Allis-Chaimers Corp., "Evaluation of One and Two Stage High Head Pump/Turbine Design for Underground Power Stations" 305 Alexander Gokhman, University of Miami, "Multistage Turbine-Pump with Controlled Flow Rate" 315 COMPRESSED AIR ENERGY STORAGE Walter V. Loscutoff, Battelle Pacific Northwest Laboratories, " Compressed Ai r Energy Storage Program Overvi ew" "331 L. E. Wiles, Battelle Pacific Northwest Laboratories, "Fluid Flow and Thermal Analysis for CAES in Porous Rock Reservoirs" 337 J. R. Friley, Battelle Pacific Northwest Laboratories, "Thermo Mechanical Stress Analysis of Porous Rock Reservoirs" 349 J. A. Stottlemyre, Battelle Pacific Northwest Laboratories, "Potential Air/Water/Rock Interactions in a Porous Media CAES Reservoir" 355 P. F. Gnirk, RE/Spec Inc., "Preliminary Design and Stability Criteria for CAES Hard Rock Caverns" 363 R. L. Thorns, Louisiana State University, "Preliminary Long-Term Stability Criteria for CAES Caverns in Salt Domes" 375 H. J. Pincus, University of Wisconsin, "Fabric Analysis of Rock Subjected to Cycling with Heated, Compressed Air" 385 Shosei Serata, Serata Geomechanics Inc., "Numerical Modeling of Beha- vior of Caverns in Salt for CAES" 393 Frederick W. Ahrens, Argonne National Laboratory, "The Design Optimization of Aquifer Reservoir-Based CAES" 403 George T. Kartsounes, Argonne National Laboratory, "Evaluation of Turbo-Machinery for Compressed Air Energy Storage Plants" 417 George T. Kartsounes, Argonne National Laboratory, "Evaluation of the Use of Reciprocating Engines in Compressed Air Energy Storage PI ants" 427 Walter V. Loscutoff, Battelle Pacific Northwest Laboratories, "Advanced CAES Systems Studies" 439 Gerrard T. Flynn, Massachusetts Institute of Technology, "Solar Thermal Augmentation of CAES" 449 Table of Contents (Cont'd) APPENDICES Alan Mi liner, Massachusetts Institute of Technology, Lincoln Laboratory, "The Application of Flywheel Energy Storage Tech- nology to Solar Photovoltaic Power Systems" 457 Program... 463 List of Attendees 469
IV PREFACE
The Mechanical and Magnetic Energy Storage Annual Contractors' Review Meeting was held in Luray, Virginia, on October 24-26, 1978, to review the current research and development activities being carried out in these programmatic areas by the Department of Energy's Division of Energy Storage Systems. Approximately 140 representatives from government, national labor- atories, universities and industry attended the three-day meeting, during which 45 presentations were made by major DOE contractors and subcontractors. Within the Division of Energy Storage systems, the responsibility for the development of mechanical and magnetic energy storage lies within the Advanced Physical Methods Branch. Included are five technology areas — flywheels, mechanical energy storage for solar/wind applications, compressed air energy storage, underground pumped hydroelectric energy storage, and superconducting magnetic energy storage. Applications for the technology include ground transportation, solar energy systems, and large-scale electric utility systems. The meeting was organized and conducted by the Lawrence Livermore Laboratory under the direction of Dr. George C. Chang, Chief, Advanced Physical Methods Branch of the Division of Energy Storage Systems. Session chairmen were selected from the active program participants for the eight designated technical sessions. Specific arrangements for the respective sessions were coordinated through the session chairmen. Credit for the success of this meeting is largely due to the active support of these session chairmen, as well as that of the individual speakers and of the Bradford National Corporation, through whom the administrative arrangements for the meeting were made. The technical papers which make up these proceedings were individually propared by the responsible contractors and subcontractors; the function of the "editors" in this case has been primarily that of reviewing the individual submissions for completeness and assembling the document for printing. Technical editing, in the formal sense, has been minimal.
Thomas M. Barlow, Chairman Mechanical and Magnetic Energy Storage Annual Contractors Review Meeting-1978 OPENING REMARKS
George C. Chang
As the Branch Chief for Advanced Physical Methods, I am happy to welcome you to the 1978 Mechanical and Magnetic Energy Storage Contractors' Review Meeting. We envision this to be the first of a series of annual meetings for which there are two purposes. The first is to make it possible for DOE Headquarters personnel to obtain, with a minimum of travel, a comprehensive overview of our programs and an acquaintance with the participants. The second purpose is to stimulate the exchange of information and ideas between different organizations and disciplines. To help accomplish the first purpose we have located this meeting in a pleasant place with easy access to Washington. To accomplish the second purpose we want to encourage all of you to be open and unin- hibited in making your comments and constructive criticism during the presentations to follow. Me also plan to publish the proceedings of this meeting so that the general public can have ready access to the information presented here. During the next few days you will hear reports from a wide variety of disci- plines, ranging from compressed air for utility peaking plants to flywheels for regenerative braking. You will hear about such esoteric subjects as superconducting magnets and more basic concepts such as underground pumped hydro. Each of these storage technologies has something special to offer. Perhaps the only quality they have in common is that all of them are advanced physical methods as opposed to chemical, thermal, thermo-chemical, or electro-chemical methods. Some of the technologies are for dispersed systems. The best examples are flywheels to store solar energy captured by photovoltaic collectors. Other technologies apply exclusively to centralized systems. Compressed air storage to be used in utility peaking plants is the outstanding example. Some of our programs such as magnetic energy storage have a long-term pay-cff. Others, such as compressed air, are almost ready for near-term exploitation. Flywheels for automotive application are a mid-term technology which we expect to be ready for commercialization in the mid-to-late 80's. We are pleased both with our mix of long-term and short-term projects and with our mix of dispersed and centralized technologies. We are also pleased with the progress that has taken place since 1975 when these programs were initiated. We hope that you, too, will be impressed as the details of our programs unfold in front of you during the next three days.
VI WELCOMING REMARKS I- >i James H. Swisher t ii I, too, would like to welcome you to the conference. For the past two years we have had annual contractors' review meetings for both the Thermal Storage Program and the Hydrogen Program, but until this year we have never had one for the Mechanical and Magnetic Programs. It is our intent to make this an annual event. In the past, we have found that we had a great deal of difficulty in sending all the Headquarters staff to annual meetings away from Washington. Hence, we \ would like to continue to have the meeting within a 150 mile radius of Washington. Also, we thought that we were missing out by not encouraging and enticing other government people to come to some of our reviews to learn more about our programs. We have made an attempt this year to see if some Congressional staff aides might like to come. We have also invited representatives from both DOE policy groups and other government agencies. By looking around the room I can see that we have been successful in obtaining the wider audience which we have been seeking. I am particularly pleased to be able to offer a special welcome to the new participants in our meetings. I hope that you will favor us with your comments and suggestions during the days to come.
vii SESSION I: FLYWHEELS PROJECT SUMMARY
Project Title: Mechanical Energy Storage Technology Development for Electric and Hybrid Vehicle Applications Principal Investigator: Thomas M. Barlow Organization: Lawrence Livermore Laboratory Mail Stop L-209 P. 0. Box 808 L-svermore, CA 94550 (415) 422-6434 Project Goals: The goals of the E&HV-MEST project are to develop and demonstrate mechanical energy storage technology for effective application to electric and hybrid vehicles in accord with the provisions of the Electric and Hybrid Vehicle Research, Development and Demonstration Act and to maximize the commercialization potential of this tech- nology. Project Status: Work was initiated in the final quarter of FY1977; three major subcontracts for the development of MES technology have been placed and are scheduled for completion in FY1979. Two of these, with Garrett AiResearch and General Electric CR&D, respectively, involve fully-contained flywheel energy storage systems with electric input/output machines. The third contract, with the Eaton Corporation, is for the design and evaluation of an elastomeric energy storage subsystem. Contract Number: W-7405-ENG-48 Contract Period: Sept. 1977 - Sept. 1979 Funding Level: $1,400,000 Funding Source: U. S. Department of Energy MECHANICAL ENERGY STORAGE TECHNOLOGY DEVELOPMENT FOR ELECTRIC AND HYBRID VEHICLE APPLICATIONS*
Thomas M. Barlow Lawrence Livermore Laboratory P.O. Box 808 (L-209) Livermore, California 94550
ABSTRACT
The Department of Energy authorized t'.a E&HV-MEST Project in September 1977 to provide a focus for its efforts to develop mechanical energy storage technology for application to electric and hybrid vehicles. Technical management responsibility for the project was assigned to the Lawrence Livermore Laboratory. LLL has, in the past year, contracted with industry for the development of two advanced flywheel concepts and one elastomeiic energy storage concept, all applicable to regenerative braking and designed to improve the performance and fuel economy of electric vehicles. Additional efforts include an experimental study of the effect of load leveling on battery life and analytical evaluations of mechanical energy storage technology. These activities are integrated in an overall plan and management structure designed to enhance the commercialization of electric vehicles.
INTRODUCTION BACKGROUND
The Electrical and Hybrid Vehicle FLYWHEEL TECHNOLOGY DEVELOPMENT Mechanical Energy Storage Technology (E&HV-MEST) Project was created in Sep- The U.S. Department of Energy's tember 1977 by the U.S. Department of flywheel technology program has shown Energy for the purpose of developing and steady growth and progress since its demonstrating the technology of flywheels inception in 1975 . As a result of an and other mechanical energy storage early technical and economic subsystems for application to electric and feasibility study,2 the main focus of hybrid vehicles. The responsibility for the program has been on the development conducting the project was delegated to of composite rotors which have high the Lawrence Livennore Laboratory (LLL). energy density and can be economically produced. Application of the This paper describes the background technology to the transportation which led to the establishment of the E&HV- sector has been emphasized. MEST Project and identifies the authority under which it is conducted. It also In the technical and economic presents the project plan and the organization feasibility study noted above, several by which the plan is carried out. A findings were reported which have summary of the progress made in FY 1978 in guided the DOE Program: each of the task areas describes the specific efforts undertaken, both within the Labora- o The cost of the flywheel is the tory and by subcontract. Other presentations major determining factor in at this meeting will provide additional the application of flywheel detail concerning the subcontracted efforts. technology.
o The development of high energy density is the dominant factor in reducing flywheel cost.
"Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livcrmore Laboratory under contract number W-7405-ENG-48." • The development of composite ro- practicability of electric and hybrid ve- tors with high energy densities hicles; and to promote the introduction will broaden the applicability of of electric and hybrid vehicles where flywheels in the transportation practical and beneficial. With regard sector and could provide oil sav- to energy storage technology, the Act ings in the range of 100 million provides for research and development; barrels of oil (0.56 quads) per regenerative braking is specifically year by 1995. identified as an area for development.
• Flywheel safety need not be regard- E&HV-MEST PROJECT SUMMARY ed as an unusually severe problem. Through the cooperation of the DOE • Flywheel systems with metallic ro- Divisions of Energy Storage Systems and tors are applicable to certain Transportation Energy Conservation, the near-term vehicle systems and would E&HV-MEST Project is carrying out the result in moderate oil savings. mechanical energy storage technology de- v.lopment specified by the Act. The pro- hire recent studies3"5 have reinfor- ject was authorized by DOE in September ced these findings and have further de- 1977 and is being conducted by the Univ- fined the potential of flywheels (and ersity of California's Lawrence Livermore that of other mechanical energy storage Laboratory in tivermore, California. The concepts) to provide electric and hybrid project emphasizes the development and vehicles with improved performance, in- evaluation of mechanical energy storage creased range, and extended battery life. technology on a subsystem level. While most of the experimental work is carried out by In addition to the technology studies subcontract to private industry, both uni- conducted under the DOE Program, technol- versities and the National Laboratories ogy development has also been addressed are involved in the analytical and project in the laboratory on both the component planning efforts. and subsystem levels. These efforts have been reported in the flywheel technology THE E&HV-MEST PROJECT symposia6'7 and elsewhere and include the development of advanced composite rotors GOALS AND APPROACH and other components, as well as the ad- vancement of complete systems for auto- The goals of the E&HV-MEST Project are motive application. to develop and demonstrate mechanical energy storage subsystems technology for effective THE ELECTRIC AND HYBRID VEHICLE ACT OF application to electric and hybrid vehicles T97o" in accord with the provisions of the Elec- tric and Hybrid Vehicle Research, Develop- The enactment of the Electric and ment, and Demonstration Act of 1976 and to Hybrid Vehicle Act of 1976 (Public Law maximize the commercialization potential of 94-413) gave added impetus to the DOE this technology through the continuing in- flywheel technology program. Within the volvement of the private sector. Act, Congress authorized the Department of Energy to conduct a program designed In carrying out the project, the po- to promote electric vehicle technologies tential for energy savings is investigated and to demonstrate the commercial feasi- through concurrent theoretical analysis and bility of electric and hybrid vehicles.8 laboratory evaluation. Emphasis is given The Act cites recognition of the facts to technology development efforts which are, that the Nation's dependence on foreign for the most part, carried out by subcon- oil must be reduced and that the intro- tract to private industry. Supporting an- duction of electric and hybrid vehicles alyses are conducted by the National Labor- into the transportation fleet could have atories, universities, and other organiza- a substantial impact on the use of pet- tions as appropriate. roleum in this country.- By means of the Act, Congress declared the policy to en- Efforts within the E&HV-MEST Project courage and support research and devel- are coordinated with complementary projects opment of electric and hybrid vehicles; within the Energy Storage Technology Pro- to demonstrate the economic and technical gram and the E&HV Program. Information E & HV-MEST project
Advanced Systems Technology Technical development component application management and evaluation evaluation assessment
• Flywheels • Mechanical • Energy storage • Project planning components requirements and analysis • Other MES concepts • Safety and • Technology • Project direction containment assessment
• Controls and data systems
Fig. 1. E&HV-MEST Project Organization
exchange is carried out routinely in a TECHNOLOGY DEVELOPMENT AND EVALUATION number of areas of mutual interest. The objective of this task is to ad- ORGANIZATION vance the current state-of-the-art of me- chanical energy storage technology by de- The E&HV-MEST Project consists of veloping and evaluating advanced MEST con- four principal tasks, with subtasks iden- cepts on a subsystems level. tified as new efforts are planned and in- itiated (Fig. 1). The four task areas, The approach followed in this task each of which is described in more de- is to establish subcontracts with industry, tail in later sections, are: on a competitive basis, for the develop- ment and laboratory testing of specific • Technology Develoment and Evalu- energy storage concepts. These develop- ation, including both flywheels ment efforts are designed to incorporate and other mechanical energy stor- new technology appropriate for commer- age (MES) concepts, such as hyrau- cialization in the early 1980's and to lic accumulators, elastic energy provide the capability, when applied to storage devices, and the like; electric and hybrid vehicles, to meet stat- ed improvements in vehicle acceleration, • Advanced Component Evaluation hill-climbing ability, and range. In ad- which emphasizes the line of dition, cost, reliability, and operational communication with the Energy simplicity are included as factors to be Storage Technology Program; considered.
• Systems Application Assessment; The individual subcontracted efforts and are limited to periods of 12 to 18 months, thus allowing an updated level of technol- • Technical Management, under which ogy to be included in subsequent contracts. the project planning and execu- This approach is designed to encourage max- tion functions are carried out. imum participation by private industry; to allow the development of a number of prom- As noted earlier, the Technology Develop- ising concepts; to assure that the technol- ment and Evaluation Task is emphasized. ogy is advanced at rhe maximum rate consist- ent with industry's capabilities; to permit the development of competitive sources; and, finally, to provide necessary flexibility Figure 2 summarizes the schedule of within the project. these activities, and indicates their con- tinuation into FY 1979. Included as a part of the technology development projects are a production and The General Electric and Garrett economic evaluation and a commercialization AiResearch efforts will be adequately ad- plan for the technology should it be suc- dressed in separate reviews by Lustenader cessfully developed and demonstrated. and Raynard, respectively. Because the Eaton contract was quite recently nego- The FY 1978 plan for the Technology tiated, and because it involves a rather and Evaluation Task specified the placement novel concept, some additional detail con- of subcontracts with private industry for cerning both the concept and the process the development and testing of advanced MES by which it was selected is provided here. concepts applicable to regenerative braking and combined (flywh&tVbattery or flywheel During the third quarter of FY 1978, battery/internal combustion engine) power proposals were solicited and evaluated systems. Accomplishments during the year for new MES technology development efforts. included the successful negotiation of a Thirty-three "Invitations For Proposal" contract with General Electric CR&D for were sent to companies having expressed the Phase II development of the inductor an interest in the effort. Thirteen pro- motor/alternator/flywheel energy storage posals (which included a variety of energy concept and with the Garrett AiResearch storage concepts) were received and eval- Manufacturing Company of California for ated; from these, the Eaton Corporation the development of an advanced flywheel and the AVCO Corporation were selected for energy storage concept which features a further action, with priority being directed rim-type composite flywheel rotor and an to Eaton. Current funding limitations have electrical input-output machine* A third limited contract action to Eaton; further contract was awarded to the Eaton Corpor- negotiations with AVCO will be undertaken ation in September 1978 as a result of as funding is available. a competitive proposal process. Each o£ the subcontracted technology development The concept proposed by the Eaton Corp- activities is designed to provide a re- oration involves the application of an elas- generative braking capability to electric tomeric energy storage concept to regener- vehicles and each will provide improve- ative braking in an urban vehicle. The ments in performance (acceleration), fuel concept is quite simple and involves a min- economy, and range. imal amount of stored energy (less than
Garrett AiResearch Contract negotiation Design study Hardware design and fab Test and evaluation General Electric CR&D Contract negotiation Design study Hardware design and fab Test and evaluation Eaton Corporation Proposal and negotiation Phase I design study Phase II fab and test
Fig. 2. FY 1978 Technology Development and Evaluation Task Schedule 100 wh), yet the application is one in slowly increasing the pressure (from which significant oil savings could be vacuum toward atmospheric) in the chamber. realized. The first phase of this effort At about 4500 microns (4.5 Torr) the outer invclves a design analysis and evaluation, two layers of the rotor failed. The which should be completed in January 1979. failure was completely contained, and If the Phase I results are favorable, no yielding of the housing was noted. a Phase II contract will be entered. The The maximum torque experienced was 1100 Phase II effort will involve a complete ft lbs., and the time required for the laboratory test and evaluation of the con- rotor to come to a stop was approximately cept. two seconds.
The AVCO Corporation proposal des- The test is significant in the fact cribed the development and evaluation of that the non-catastrophic failure mode a flywheel energy storage concept which of the rim-type composite rotor was dem- features a disc-type composite rotor and onstrated 3nd the integrity of the con- an infinitely variable mechanical trans- tainment concept was proven. mission. This concept could be applied either to regenerative braking or to a SYSTEMS APPLICATIONS ASSESSMENT "combined" power system and is particu- larly attractive because of the high- The objectives of this task are to efficiency, all-mechanical power trans- define the applicability of MES subsystems mission system. to electric and hybrid vehicles of var- ious types; to define the functional and ADVANCED COMPONENT EVALUATION developmental requirements for these sub- systems; and to provide technology devel- The objective of this task is to pro- opment criteria as required by other vide an evaluation of advanced component activities within the project. technology from inclusion in demonstration subsystems. A variety of components are As indicated by the statement of ob- involved, including flywheel rotors, bear- jectives, this task provides guidance and ings, and seals; safety and containment criteria for the Technology Development systems; and various transmission con- Task. It includes the consideration of cepts, both mechanical and electrical. complementary energy storage technology in order to provide perspective to the MES This task is carried out in coordin- efforts. Specifically, the following top- ation with the DOE Flywheel Energy Storage ics are addressed: Technology Project, for which the Sandia Laboratory, Albuquerque, New Mexico, has • Energy storage system functional responsibility. requirements,
A significant flywheel rotor burst • Identification of viable concepts, and containment test was conducted and re- ported by the Garrett AiResearch Company • Technology development requirements, in the third quarter of FY 1978. The test was carried out as a part of the • Economics and production evaluation, Garrett Company's development effort under the Near Term Electric Vehicle Program • Environmental and safety consider- (DOE Contract EV-76-C-03-1213). A rim- ations, and type flywheel was tested to failure at 26,000 rpm within a closed containment • Technology utilization. system. The purposes of the test were to demonstrate the system's capability The operating plan for this task in to contain the flywheel failure; to es- FY 1978 included an evaluation of the effect tablish flywheel failure data; and to of load leveling on battery life, a deter- obtain data pertinent to the design and mination of the potential impact of MES analysis of containment devices. technology on electric vehicle performance, and an analysis of the MES performance re- The failure test was conducted by quirements. bringing the flywheel rotor up to test speed in a vacuum environment and then The first of these activities is being addressed through the conduct of a coopera- vehicle braking and downgrade travel. tively supported battery load-leveling test Additional design flexibility is afforded, at the U. S. Army's Mobility Equipment Re- since the specific power of a flywheel search and Development Command (MERADCOM), is essentially Independent of its specific Ft. Belvoir, Virginia. The test includes energy. Flywheels can be designed with two sets of batteries - one is cycled to a a high specific power capability, which "normal" load profile experienced by a can be utilized to provide acceleration battery-powered transit vehicle, and the and regenerative braking capability not second includes a load profile which simu- otherwise achievable in electric vehicles. lates the effect of flywheel load leveling on the battery set. This test is currently TECHNICAL MANAGEMENT underway, and an indication of the effect of load leveling is expected when it is Activities within this task include completed next year. preparing and maintaining the project plan; implementing the plan by developing cri- The second and third activities are teria, preparing specifications, and estab- being addressed analytically in separate lishing technology development projects; efforts by Dr. R. F. McAlevy and by the directing contractor efforts and reporting LLL Transporation Systems Group. A report progress to DOE; and recommending companion of the LLL analysis is being prepared and research and development efforts. Each of will be presented in the future, and Dr. these activities has been addressed in McAlevy will report his findings later in FY 1978, with the Project Plan being sub- this meeting. mitted in April 1978; proposals for tech- nology development projects invited and con- As a general comment, several criteria tracts awarded to three companies; and reg- affect the design of the energy storage ular progress reports provided to the DOE system for vehicle applications: Divisions of Energy Storage Systems and Transportation Energy Conservation. In ad- • Specific energy is important for dition to the regular project reviews, range, long grades, high-speed project personnel participated in eleven passing, and other conditions workshops, review meetings, and symposia requiring considerable quantities during the year in order to facilitate co- of energy. ordination with complementary efforts car- ried out elsewhere. • Specific power is important for acceleration, braking, and high Figure 3 summarizes the resource allo- speed grades and passing. cations for the project (in Budget Author- ity) by task and by performing institution. • Cost is always Important, and both As the figure indicates, the majority of initial and life-cycle cost must the funding supports the Technology Devel- be considered. opment and Evaluation Task, and private in- dustry is, by far, the primary "performer." • The vehicle load cycle is import- Funding for the MERADCOM test is reflected ant in that it specifies the re- as "U. S. Army," while the university seg- tired acceleration and braking, ment consists primarily of faculty consul- stops per mile, and cruise speed tants to the project. and time - all of which impact the design of the energy storage As noted earlier, the E&HV-MEST system. Project interacts formally with a number of organizations which support DOE in both The unique characteristics of fly- the Energy Storage Technology Program pprt wheels (and, to some extent, other mech- the Electric and Hybrid Vehicle RD&D anical energy storage concepts) can be Program. Figure 4 schematically indi- used to advantage to extend the range of cates these interrelationships: electric vehicles, improve their acceler- ation characteristics, and provide them • The DOE Division of Energy Storage with increased battery life. These char- Storage Systems (STOR) is respon- acteristics include the unequaled ability sible for the Flywheel Technology to effectively recover and store, for Project, for which the Sandia Lab- future use, the energy normally lost in oratory in Albuquerque (SLA) has By Task By Performing Institution
Syitami Technical Application Universities U.S.Army Management,. .^Assessment 2% 5%
Fig. 3. Distribution of rY1978 Budget (Budget Authority)
Flywheel Vehicle Vehicle Propulsion technology MES systems system development technology technology technology
Fig. 4. E&HV-MEST Project Interactions
been delegated execution responsi- porting TEC in this regard are the bility. STOR provides the tech- Jet Propulsion Laboratory (JPL) of nical direction to the E&HV-MEST the California Institute of Tech- Project. nology in the vehicle systems area and the NASA-Lewis Research Center The DOE Division of Transporta- (LERC) in the propulsion systems . tion Energy Conservation (TEC) area. TEC exercises the execution is responsible for the Electric (budget) authority for the E&HV- and Hybrid Vehicle Program- Sup- MEST Project.
10 FY 1979 PLANS 2. Economic and Technical Feasibility Study for Energy Storage Flywheels, Current plans for the E&HV-MEST Pro- Energy Research and Development Admin- ject in FY 1979 include completing the istration Report ERDA 76-65, 1975. Garrett AiResearch and General Electric CR&D flywheel technology efforts; evalu- 3. E. Behrin, J. Bolger, C. Hudson, L. ating the Eaton Corporation's elastomeric O'Connell, B. Rubin, M. Schwartz, C. - energy storage concept, and initiating Waide, and W. Walsh, Energy Storage the Phase II development, if warranted; Systems for Automobile Propulsion. and completing the MERADCOM battery load- Lawrence Livermore Laboratory Report leveling test. Additional tasks, including UCRL-52303, Vols. I and II, 1977. the development of the AVCO flywheel energy storage concept, will be initiated as fund- 4. Determination of the Effectiveness ing permits. Supporting activities will be and Feasibility of Regenerative continued in the areas of applications an- Braking Systems on Electric and Other alysis and technical management. Automobiles. U. S. Department of Energy Report UCRL/W52306, Vols. I CONCLUSIONS and II, 1978.
The first year's progress of the E&HV- 5. R. F. McAlevy, R. F. McAlevy and As- MEST Project has been significant: sociates, Hoboken, IJJ, Private com- munication, September 1, 1978. • Three major technology development efforts are being carried out under 6. Proceedings of the 1975 Flywheel Tech- subcontract to private industry. nology Symposium. G. C. Chang and R. G» Stone, Editors, Energy Research and • The effect of load leveling on bat- Development Administration Report tery life is being experimentally ERDA-76-85, 1976. investigated. 7. 1977 Flywheel Technology Symposium Pro- • Mechanical energy storage system ceedings . G. C. Chang and R. G. Stone, performance criteria for vehicle Editors, U. S. Department of Energy applications has been determined. Report CONF-771053, 1978
• An analysis of the importance of 8. Introduction to the ERDA Electric and MES to electric vehicle performance Hybrid Vehicle Demonstration Project. has been completed. Energy and Research Development Admin- istration Report ERHQ-0008, 1977. Although the current state-of-the-art is adequate for some vehicular applications, 9. B. H. Rowlett, Burst and Containment further development is required, on both Test Report for Flywheel Power Systems. the component and subsystems level, in Near-Term Electric Vehicle Program order to improve the performance and eco- Phase II. Garrett AiResearch Manufac- nomics of mechanical energy storage tech- turing Co. of California Report 78- nology. 15148, 1978. NOTICE REFERENCES "This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States 1. G. C. Chang, J. H. Swisher, and G. F. Department of Energy, nor any of their employees, Pezdirtz, "DOE's Flywheel Program," nor any of their contractors, subcontractors, or their employees, makes any warranty, express or in 1977 Flywheel Technology Symposium implied, or assumes any legal liability or respon- Proceedings, G. C. Chang and R. G. sibility for the accuracy, completeness or usefulness of any information, apparatus, product Stone, Editors, U. S. Department of or process disclosed, or represents that its use Energy Report CONF-771053, 1978. would nol infringe privately-owned rights."
11 PROJECT SUMMARY
Project Title: The Impact of Mechanical-Energy-Storage Device Characteristics on Hybrid Vehicle Performance
Principal Investigator: Dr. Robert F. McAlevy III
Organization: Robert F. McAlevy III & Associates 1204 Bloomfield Street Hoboken, NJ 07030
Project Goals: Establishment of a rational basis for evaluating the impact on automotive vehicle performance produced by addition of a mechanical-energy- storage-device.
Project Status: Phase I work was initiated on May 25, 1973 and completed on September 1, 1978. Phase II efforts were funded on September 14, 1978, with an expected completion date of December 1, 1978.
Contract Number: University of California Purchase Order 2122409
Contract Period: May 25, 1978 - Dec. 1, 1978
Funding Level: $22,410
Funding Source: Lawrence Livermore Laboratory
13 THE IMPACT OF MECHANICAL-ENERGY-STORAGE-DEVICE ADDITION ON THE PERFCIUVIANCE OF ELECTRIC VEHICLES
Robert F. McAlevy III Robert F. McAlevy III & Assocs. 1204 Bloomfield St. Hoboken, New Jersey 07030
ABSTRACT Electric vehicle, EV, power and energy balances were used to estab- lish battery requirements as a function of vehicle mission specification. Normal freeway/urban commuter car driving patterns were found to require battery power capabilities in excess of those forecast for this century. However, Ni/Zn and other batteries should ptrmit EV's to follow normal urban commuter car driving patterns. A criterion for the optimum battery is evolved and used to delineate between power-determined (battery mass in excess of that needed to meet range specification) and range-deter- mined EV designs. The impact of adding a mechanical-energy-storage, MES, device to an EV depends strongly on the "baseline" EV design. For range- determined EV's, analytical relationships are developed for the maximum MES device mass permitted as a function of energy-efficiency increment produced, so that there will be no change in: (a) vehicle mass and (b) vehicle energy consumption,compared to the baseline EV. The limiting case of MES device mass approaching zero (e.g., infinite power-density and energy-density) while the increment of energy-efficiency remains is also examined. Therefore, the boundaries of beneficial MES device application to range-limited EV's has been established. Presently, the analysis is being extended to EV's of power determined design, where the payoff for MES device addition can be much greater, and to ICE vehicles.
BACKGROUND AND INTRODUCTION ent day batteries compared to a "tank" of liquid fuel. For example, By incorporating mechanical- the best lead/acid oattery has a energy-storage (MES) devices into gravimetric energy content less automotive vehicles it is possible than 1/100 that of petroleum-based to improve the performance of the fuels. Therefore, in order to vehicles. The performance incre- achieve a range of 60km or so, it is ment depends on both the performance necessary to allocate about 1/3 of of the original, or "baseline" ve- the EV mass to batteries (coiflpared hicle, and the characteristics of to about 1/30 for a fuel gas tank the MES device employed. in petroleum-fueled automobiles). In order to carry such battery mas- Electric vehicles, EV's,are ses in vehicles with reasonable the baseline vehicles considered in payload-fractions, it is necessary this document. In order to qualify to compromise their structural in- for participation in Federally sup- tegrity. For example, while using ported programs, they must exhibit the same materials and fabrication certain minimum performance levels . techniques as automobiles, EV's Recently, the performance of extant were found to have body/chassis EV's was critically reviewed^. Both mass-fractions that are, typically, the Federal Standards and actual only 0.45^ compared to 0.65 or so performance are substantially in- for present day automobile. ferior to that exhibited by present day automobiles of common experi- Battery research and develop- ence. This shortfall is a direct ment activities should result in result of the shortcomings of pres- improved batteries in the future. A panel of experts has forecast the the vehicle mass-fractions alloca- characteristics of future batteries ted for the several vehicle func- through the end of the century^. A tions that must be fulfilled. computer model was used to calcu- late the performance of EV's that The Vehicle Drive Cycle. Drive- use such batteries^. cycles are used to model vehicle driving patterns. A drive-cycle is The subject program involves merely the specification of the ve- an analytical model of baseline EV hicle's velocity-history over a performance. It is based on a time interval. The vehicle is pre- methodology developed and described sumed to repeat the pattern over elsewhere* and used previously for its entire range. comparisons amongst different kinds of alternate energy vehicles^. Here Integration of the velocity- it is used to establish the crite- history over the drive-cycle inter- rion for the optimum EV battery as val yields the distance traveled a function of EV mission requirement during the drive-cycle, d. The and to evaluate the potential of number of drive-cycles accomplished future batteries-^ to power EV's before depleting the battery below through different missions — spe- some reasonable level (say 80% dis- cifically, the urban commuter car charge) , n, times d yields the ve- and freeway/urban commuter car mis- hicle1 s practical range, R. That sions. It is' used to establish the threshold MES device characteris- tics required for "breakeven" per- R = nd (1) formance when the devices are added to baseline EV's of range deter- Knowledge of vehicle charac- mined design. teristics — aerodynamic-drag-co- efficient, frontal-area, and coef- This work is continuing and ficient of rolling-resistance — future reports will deal with the coupled with knowledge of the in- breakeven condition for baseline stantaneous velocity, permits cal- EV's of power determined design, as culation of the instantaneous road- well as the performance increments load imposed at the vehicle's possible by incorporation of MES wheels. Multiplication of the in- devices of known characteristics. stantaneous road-load by the in- It is expected that this work will stantaneous velocity yields the lead to a rational basis for evalu- instantaneous power. Inspection ating the impact of MES device ad- of the instantaneous power at each dition on the performance of EV's. point in the cycle reveals the max- The benefits that accrue will vary imum vehicle power demand, Pjnax- from situation to situation, but Integration of the instantaneous they include making it possible for power over the drive-cycle yields the MES/electric vehicle to perform the energy requirement at. the ve- missions from which the baseline EV hicle's wheels in order to achieve is precluded due to inadequate bat- one drive-cycle, Ev. tery power, improving vehicle structural integrity, and reducing These simple exercises have the ownership costs of electrically been performed assuming typical EV energized vehicles — which are characteristics? and a variety of currently about 5 times greater than drive-cycles or a level road. The petroleum-fueled automobiles*". Ap- results of these and other calcula- plication of the methodology to tions** are shown in Table 1.* VJ
16 maximum power level. Thus, Note that every term, except could oe reduced, or a more econom- (es)p_Q, depends on drive-cycle. ical battery of reduced (pTd.) [Ev/WTd] depends on vehicle aero- could be employed, etc. dynamic drag and rolling-resistance and their interaction with the Sizing the Battery System to Meet drive-cycle, np depends on power- the Range Demand. Due to losses in train characteristics and their in- the power-train, more energy must teraction with the drive-cycle and be supplied to it than can be de- np depends on battery system char- livered to the wheels. The propul- acteristics and their interaction sion efficiency, np# of the energy- with the drive-cycle and rip- conversion-device, which is strong- ly dependent on drive-cycle, ac- Nevertheless, despite the counts for these losses. For elec- somewhat complicated composition trochemical batteries, more energy of Rult» it nas a simple meaning. must be stored in them than can be It is the range achieved by a Uto- delivered to the energy-conversion- pian vehicle composed only of the device. The battery energy effic- energy storage-device, having an iency, rig, which is strongly de- energy-efficiency of ripig and fol- pendent on drive-cycle and rip, ac- lowing a drive-cycle with an en- counts for the internal dissipa- ergy-requirement of [Ev/WTd]. That tion within the battery. The ve- is, R •* Ruit when Wg •+ W^. hicle "energy-efficiency" is de- fined as n = ngnp. Note that this Real vehicles are composed of definition does not account for more than an energy-storage de- losses during charging of the ener- vice, so Wg < WT and therefore, As gy-storage-device, and properly so, R < Ruif previously mentioned since the charging process has Wg/WT for EV's are typically 10 nothing to do with vehicle perform- times greater than that of gaso- ance per se once the vehicle is put line-fueled vehicles, or Wg = 0.3 into operation. WT. Therefore, for EV's R = 0.3 Rult* The better EV lead-acid Defining the gravimetric-en- batteries exhibit (es)p=o = 45whr/ ergy-content of the energy-storage- kg and r)S = 0.65 over the SAEJ227a/ device, following charging and be- C-cycle. The better power-trains fore power is drawn, as (es)p_Q, exhibit rip = 0.7 over the cycle and then the balance between the ener- since [Ev/WTd] = 0.09 whr/kg-km, gy initially stored on board, Rult s 200 kln- Therefore, the Wg (es)p=Q, and the energy required range of an advanced lead-acid EV to execute n drive-cycles, can be over the SAEJ227a/C-cycle, assum- written as: ing WS/WT = 0.3, should be R = (0.3) (200) = 60 km. Experimental- [E /W d] ly-measured valves for R support 2
17 Therefore, incorporating an MES de- a vehicle with no R. Practical ve- vice increases n and therefore, hicles must have WS/WT > 0. Rulf As Ws/WT is increased in re- Distribution of Vehicle Mass-Frac- sponse to increasing range and tions. The vehicle total mass, Wrp, power demand, then WT/WPL must in- is composed of the sum of the mas- crease as well. The upper limit ses of its several components. for WS/WT is l-B-a-6 for ME:-/EV That is, hybrids and 1-B-a for EV's, be- cause as these finite levels are WT = WPL + WEC + WS WM + WB (8) approached W^/Vipi, •* °°, which also represents an impractical limit. where: W is the payload mass Practical vehicles must have Ws/WT PL < 1-B-a. W,'M is the MES-device mass, including that of its Since a, B, 6 and WS/WT can container, shaft, be assumed to be independent of WT gears, etc. for vehicles of the same generic type and design values, then WT/ WB is the body/chassis mass is independent of W^ as well. and the other terms have been de- It is interesting to note fined previously in this section. that if an MES device is added to an EV, while a + 3 are held con- Dividing Eq. 8 by W_ yields, stant, then Wr£./WpL would increase unless there was a compensating WP WS decrease in W5/W1J. (9) Operating Energy Consumption. The quantity of battery energy con- where: 6 is the MES-device mass- sumed per unit distance of vehicle fraction travel, C, is given in terms of B is the body/chassis the present notation as. mass fraction IE /W d] —Y.V 1 Re-arranging Eq. 9 yields, (12) nsnp This equation shows that w. C - WT- PL 1 - l-g-a-6 (10) Division of both sides of this equation by WpL yields In the case of 6 = 0, that is, an EV, Eq. 10 reduces to. (13) W, W PL nsnp "PL T (1-B-a) = —— (11) w.PL 1 - x is the energy consump- 1-B-a tion per unit distance of payload mass carried and is a useful pa- ppjj is an important ve- rameter for comparing the energy- hicle parameter. It is the amount performance of vehicles. of vehicle mass required per unit of payload and is a strong driver Since Ws/WpL is independent of vehicle capital costs. The low- of Wi>, then C/WPL will be inde- est possible value that Wj/Wpj, can pendent of WT if [Ev/WTd] and take is l/l-B-a-6 for an MES/EV are independent of WT. /T hybrid and 1/1-B-a for an EV. These decreases very slowly with in- values are reached in the impracti- creasing WT' and this variation cal limit of WS/WT •* 0. That is, will be ignored here. Also
18 should be independent of WT and Thus, for every set of mis- this will be assumed here as well. sion requirements (i.e., specifi- cation of R and drive-cycle), there The vehicle operating energy are minimum values of battery costs per unit of payload carried electrical characteristics that vary directly with C/WPT. Substi- must be exceeded if the EV contain- tution of Eqs. 10 and II into Eq. ing the battery is to be physically 13 shows how C/WpL increases with realizable. increasing Wg/WnT. Incorporation of an MES into an EV would produce Defining the minimum value of an increase in C/Wp^ unless there battery gravimetric peak-power- were compensating changes in VIQ/VI^, density as (pVa.)min, then, from and/or ngrip» Eqs. 4 and 14,
It is interesting to note that if ns = 0-65 and np = 0.7 are sub- (15) stituted into Eq. 12 along with min [Ev/WTd] = 0.09 whr/kg-km, which are appropriate for the lead/acid So, for physically realizable EV's battery EV following the SAEJ227a/ of power-determined designs, the C-cycle, Eq. 12 becomes C = 0.2 WT, decrease of Ws/Wy below (Wg/W ) whr/km, which is typical of the T max with increasing battery (pVct.) can better EV values found experimental- be expressed as, iyB. ANALYSIS (p. ws/wT (16) (ws/wT) (p.d.) In this section, the impact max on vehicle performance produced by the addition of an MES device to an Battery energy capacity is EV will be analyzed. The analyti- the product of (es)p=g and ns« cal approach developed in the pre- These appear as a product in the vious section will be employed to definition of Ruit (Eqs. 6 and 7), do so. In particular, the minimum and therefore, Ruit increases MES device characteristics required linearly with increasing ns(es)p=0* for no change in vehicle W^/Wp^ and Defining the minimum values of C/Wpj, — the "breakeven" situa- ultimate range and battery energy tions — will be established. capacity for physically realizable and n EV's as (Ruit)min [ s(es)p=olnin The nature of the impact de- respectively, then from Eqs. 6, 7 pends on whether the EV is of power or and 14, range-determined design, and this will be shown to depend on both (17)
19 EXAMINATION OP THE ABILITY OF FU- (Ws/WT)max/(Ws/WT), so (p?d\) = TURE BATTERIES TO MEET FUTURE EV (108)(0.37)/(0.23) = 173 w/kg. DRIVE-CYCLE POWER DEMAND This is the battery gravimetric peak-power requirement that must Urban/Freeway Commuter-Car Driving be met if the EV containing it is Pattern. The modest, for petrole- to follow the Scott cycle and be um-fueled autos, power demand of of prudent (WT/WPL = 7) design. the Scott cycle will be taken here as the model for urban/freeway com- Present day EV's, powered by muter car driving patterns. From lead/acid batteries, cannot even Table 1, to follow this cycle, an follow the SAEJ227"a"D cycle, so EV will have to have a road power there is no possibility for them capability of 32 w/kg (or 0.02 to follow the Scott cycle. How- hp/lb). Note that this is well be- ever, batteries having better elec- low the power of today's petroleum- trical characteristics hopefully fueled vehicles and only 2/3 of will become available in the future. that forecasted for future petrole- Recently, a panel of battery ex- um-fueled vehicles3. Further, it perts have projected, with opti- leaves no margin of "reserve power" mism, the electrical characteris- for use should extraordinary driv- tics of batteries that might be ing conditions occur during the available for EV applications in normal Scott cycle. Reserve power the 1990-2000 time frame3. The would be used if the EV, while (p7a.)'.<3 are: lead/acid, 94.6w/kg; traveling at high speed, had to ac- Ni/Fe, 103.8 w/kg, Ni/Zn, 139.9 celerate to avoid having an acci- w/kg; Zn/Cl2* 88 w/kg; Li-Al/FeS, dent. For the Scott cycle, a = 130 w/kg; Na/S (B-alumina), 130 0.18. w/kg; Na/S (glass), 160 w/kg; Li/air, 129.4 w/kg. Assuming it is necessary to hold EV £ to 0.45, and thus accept It is clear that none of these lower structural values than those batteries would be able to power exhibited by petroleum-fueled ve- an EV through the Scoti cycle. hicles,4 then 1-B-a = 1 - 0.45 - That is, all of the (pTd.)'s are 0.18 = 0.37. (Note that if the below (p.d.)mj.n = 173 w/kg. There- power train energy-density increas- fore, if there can be no improve- es in the future a will decrease ment, in fact, over these optimis- and 3 can increase while l-$-a re- tic forecasts, then EV's of pru- mains at 0.37.) Assuming^that np = dent design will not be able to 0.8, then from Eq. 15, (p.d.)min = follow the Scott cycle — or any 32/(0.8)(0.37) = 108 w/kg. driving-pattern with comparable peak power demands — in this When (pTct.) = (pTa.)min» %/ century1 WPL = "» an<^ this does not repre- sent a physically realizable ve- Consequently, only by employ- hicle. The actual Ws/WT must be ing an MES device to provide the less than 1-g-a. It will be as- peak-power demand of the Scott (or sumed 5 that prudent E.V. design comparable) cycle will electrical- calls for WT/WpL = 7 if the acqui- ly-energized vehicles be able to sition costs and operating energy compete with petroleum-fueled costs (per unit payload mass) are autos in the urban/freeway commu- to be reasonable. Substituting ter car market. This is a strong WT/WPL = 7 and 1-g-a =0.37 into imperative for developing appro- Eq. 11 yields WS/WT = 0.23. [Note priate MES devices. that the lower value of WS/WT com- pared to present practice 2 stems Urban Commuter-Car Driving Pattern. from the larger value of a re- The EPA urban cycle will be taken quired by the Scott cyclercompared here as the model for urban com- to a = 0.07 required by present day muter-car driving patterns (i.e., EV's following the low-power SAEJ with no freeway component). The 227(a)C cycle.] (Pmax/WT> is 23 w/kg, which means that a is 0.13 so 1-3-a = 1 - 0.45- From Eq. 16, (p.d.) = 0.13 = 0.42. Assuming that ^=0.8,
20 then from Eq. 15, (p.d.)min = 2V have, on the road within the next (0.8) (0.42) = 68 w/kg. decade, electrically energized ve- hicles that can follow normal ur- Assuming that prudent EV de- ban commuter car driving patterns. sign calls for WT/WPL = 7, then from Eq. 11, for 1-3-a = 0.42, Ws THE OTPIMALLY-DESIGNED EV BATTERY: Wrr= 0.28. From Eq. 16, (p"T5.) = INFLUENCE OF MISSION SPECIFICATION (p7a.)min (Ws/WT)max/(Ws/WT) = 68 (.42)/0.28 = 102 w/kg. The optimally designed EV bat- tery has electrical characteris- Comparing this value with the tics that result in a value of optimistic values projected3 for vehicle Wg/Wip which simultaneously batteries in the 1990-2000 time satisfied both the mission drive- frame, it appears that EV's employ- cycle and range requirements. ing lead/acid and Zn/Cl2 batteries will be unable to follow normal/ Eqs. 16 and 19, in order to urban commuter-car driving pat- be satisfied simultaneously re- terms in this century, and those • quire employing Ni/Fe batteries will be marginally acceptable. Should it •'min (20) be possible to develop the other J
In the 1985-90 time frame, the (pTd.) max P 1 only batterie^ that have adequate [Ev/WTd] R forecasted (pTd.)'s are possibly: ! I opt the Ni/Fe, 101.0 w/kg; Ni/Zn, 135.3 w/kg; Li-Al/FeSx, 115.1 w/kg; (21) and Na/S (g-alumina), 115.1 w/kg. But if the Ni/Zn battery, for ex- where[(p.d.)/ns(es)p_0]Ql:)t is the ample, cannot be developed to the value of the peak-power/energy- capacity ratio of the optimum bat- point where many hundred deep dis- tery. Note the strong influence charges are possible, then it prob- of mission on the characteristics ably will lead to EV annual expen- required for a battery to be op- ses that are too high to be eco- timum. nomically competitive with petrole- um-fueled vehicles; and thus it would not be used in EV's in the Conversely, given battery 1985-90 time frame. In addition, characteristics and a drive-cycle, should the Ni/Fe peak-power pro- it is possible to calculate a jection be too high by several per- value of R = Ro.b. that would make cent, and the Li-Al/FeSx and Na/S these battery characteristics op- ($-alumina) projections prove too timum. Thus, from Eq. 21, high by 15 percent or so, the EV's could not follow normal urban com- muter-car driving patterns. TEv/WTd] % Therefore, considering the "S* s P=0 many technical uncertainties in- volved in the development of these (22) batteries, it would be sagacious to hedge against the possible eco- For example, if an EV were to nomic/technical shortcomings of the follow the EPA^urban cycle and had batteries that are actually devel- np = 0.75 and ^p = 0.8^ then oped by using MES devices to supply (Pmax/Wij.)/[Ev/W>jid] np/TVp = (23) EV peak power requirements. Other- (0.75)/(0.09)(0.8) = 240 w/whr-km. wise, it might not be possible to For this situation, Eq. 22 becomes
21 of R. 240 (22a)
For the batteries having fore- casted^ values of (p?5.) > 102 w/kg (and therefore, able to power prudently designed EV's through the EPA urban cycle before the end of the century,) values of Ro.b. are listed in Table 2 (fol- lowing References). The forecast 3nr discharge capacity^ is used for ri s(es)P=0 in Table 2. An EV that can follow normal commuter car driving patterns, but that has a mission R < Ro.b. will be of power-limited design. Should 260 3'50 batteries of greater (]?Vcl.) than the values listed in Table 2 be R, km developed in the time frames indi- Fig. 1 W /W vs R cated, then the WT/Wp]j and C/WPL S T penalties associated with R < Ro.b. operation will be reduced. But shguld future battery actual (p.d.)'s be equal or less than the The potential benefit that forecasted values, then the only would accrue in Wg/Wy by substi- way to satisfy R < Ro.b. mission tuting an MES-device (Ni/Zn battery) with electrically-energized vehi- vehicle for a Ni/Zn battery EV in cles, while avoiding these penal- the region of R < Ro.b. can be ties, is to employ MES devices. estimated by comparing the gap be- tween Wg/WT = 0.21 and the projec- Graphical Representation of the tion of Eq. 11 into this region of Variation of Ws/tftp with R for an power determined design. Urban Commuter Car Using a Ni/Zn (1985-90) Battery. For these con- .. N1MUM THRESHOLD MES DEVICE CHAR- r| e n E ditions, Ruit = s( s)p=o p/[ v/ ACTERISTICS FOR BENEFICIAL APPLI- WTd] = (150)(0.75)/0.09 = 1250 km. CATION TO RANGE-DETERMINED EV's And if 1-B-a = 0.37 = Ws/WT, then WT/WPL -* «. Various improvements in ve- hicle characteristics (i.e., de- Assuming that for prudent de- crease in Wy/WpL and C/WPL) can be sign, Wtp/WpL = 7, then from Eq. 11, produced by MES device addition. Ws/W-r = 0.28, so the maximum R for However, the benefits will accrue a prudently designed vehicle is only if the MES device has char- (0.28)(1250) = 350 km. Thus, the acteristics above a certain thresh- value 350 km limits R in Fig. 1. old level. The threshold MES de- vice characteristics for "break- The (pTd.) for prudent design even" vehicle performance are was found to be 102 w/kg. There- analyzed here for a baseline EV of fore, to meet the power demand, range determined design (i.e., R > WS/WT = (0.28)(102)/135.3 = 0.21. This is the level of WS/WT that yields R = Ro b = 266 km (from Assuming that a remains un- Table 1). changed, which is a realistic as- sumption for the present level of W For R > Ro.b.' T/WPL is given analysis, then for both the base- by Eq. 11. This is the region of line EV and the vehicle with the range-determined design. For R < MES device to have the same Ro.b.' WT/WPL B 0.21, independently from Eqs. 10 and 11, 1-e-a-R 1-e-a-l 1-3-cx ult
(25) «>» - % (28)
Equation 25 is displayed in Fig. 2. The result of (<5)R > (6)R de- rives from the fact that improve- For both vehicles to have the ment in energy efficiency benefits same value of C/WpL, from Eqs. 10, C/WpL more strongly than WT/WpL. 11 and 13, Thus, for the same 1m a large vehicle mass-fraction can be de- [Ev/wTd]/nm voted to the MES device and still 1-a-B-R result in break-even Wc/WpL. l-g-P-(6)R- R R ult Assuming that n > i), then: m
(26) (1) for 6 < (6")R- (WT/WpL)m < (WT/ where (
WpL) and (C/WpL)m > (C/WpL)
23 Illustrative Example of MES Device 1 [0.37-0.24] Vehicle-Mass-Fraction Required for 1.3 [0.37-0.185] 0.54 Break-Even Performance: Baseline EV of Range-Determined Design. As- sume that an EV with Ni/Zn (1985- where Cm is defined as the energy 1990) battery has a mission of 300 consumption of the vehicle con- km following an urban commuter car taining the MES device. This re- drive-cycle. Therefore, 1-6-a = sult can be represented as [1 - (C)m/C]100 = 46%. That is, 0.37 and Ruit = 1250 km. Wg/WT = 300/1250 = 0.24. (See Fig. 1) 46% reduction of energy consump- tion is possible for the condi- Should the MES device have an tions assumed here. in/out efficiency such that, as a result of regenerative braking and For actual MES devices having battery peak-power-load-levelling, 0 < 6 < 0.055, WT savings will de- crease and approach 0 as S ap- ri/nm = 1/1.3 (which is reason- able3'6 ) then from Eq. 23, (6)R = proaches 0.055. If 0.055 < 6 < (0.24) [1 - 0.771 = 0.055, and from 0.085, the vehicle containing the MES device will have a greater mass Eq. 27, (6)R = (0.37) [1 - 0.77] = 0.085. than the baseline EV, but its en- ergy consumption will be below This calculation shows that, that of the baseline EV. Finally, for the same mission, a vehicle MES devices requiring <5 > 0.085 in order to produce nm/n = 1.3, can- incorporating an MES device will not be beneficially employed to re- have a lower Wy than that of the duce vehicle energy consumption baseline EV if the device can pro- for the situation considered in duce r\/r\t=lA-.'i while having a mass this example. less than 5.5% of WT. And the MES- device vehicle will consume less energy than the baseline EV if its RESULTS AND FUTURE WORK mass is less than 8.5% of W . T Analytical energy and power Illustrative Example of Best Possi- balances were used to produce equa- ble Performance of Vehicle With tions describing vehicle mass and MES Device: Baseline EV of Range- energy consumption as a function Determined Design. For the situa- of payload and mission specifica- tion considered in above section, tion. They form the basis for assuming that the gravimetric en- evaluating the impact on perform- ergy and power content of the MES ance produced by adding MES de- device is so great that 6 •* 0, then vices to vehicles. in the limit of 6=0 Eqs. 10 and 11 can be used to calculate the For EV's, it was found that corresponding (Wf/WpiJm,6=0 and no battery forecast for develop- (C/W ) ,6=o- ment in this century had a peak- PL m power capability sufficient for an = urban/freeway commuter car of pru- Since Rult = 1250 km; (Rult)m dent design. Only by incorpora- (1.3) (1250) = 1625 km, soR^(Rult)m= ting an MES device could EV's per- 300/1625 = 0.185. From Eq. 10, form this mission with these bat- (WT/WpL)m,6=0 = 1/(0.37 - 0.185) = 5.4 so a (7 - 5.4)/7 x 100 = 22.8% teries. reduction in WT is ideally possi- ble for the conditions assumed For missions that can be per- here. formed with EV's containing the forecast future batteries, (e.g., an urban commuter car mission,) the From Eqs. 10, 11 and 13 impact of MES device addition was found to depend on the baseline R [1-g-a- ] EV design. A criterion for the (C), optimum EV battery as a function of mission specification was evol- V [l-n-ct-rf- ved. Should an EV be used in a mission with a range less than
24 that derived from the criterion, Systems, Washington, D.C., April it is of power determined design; 17-22, 1977. if the range is greater, the de- sign is range-determined. 5) McAlevy, R.F. Ill, "Optimum De- sign of Automotive Vehicles Employ- ing Alternative Energy Sources of Time permitted only the range- Low Energy Density: Impact on determined case to be analyzed. The threshold MES device character- Selection of an Energy Carrier for istics for breakeven vehicle per- Future Urban Vehicle Transporta- formance was established as a tion Systems," Proceedings of the function vehicle energy efficiency International Conference on Alter- increment. The impact produced in nate Energy Sources, Miami Beach, the limiting case of MES device Florida, Dec. 5-7, 1977. gravimetric power and energy con- 6) Sandberg, J.J. and Leschly, L., tent approaching °° was examined. "User Experience with On-Road Electric Vehicles in the U.S.A. Future work includes a break- and Canada," Proceedings of the even analysis for EV's of power 13th Intersociety Energy Conver- determined design and use of the re- sion Engineering Conference: Aug. sults to examine the impact of 20-25, 1978, San Diego, CA, pp. actual MES devices on both power 644-654, SAE P-75, SAE Inc., War- determined and range determined rendale, PA, August, 1978. designs, and application of the methodology to establish the im- 7) Liles, A.W. and Fetterman, pact of MES-device addition to G.P., Jr., "Selection of Driving petroleum-fueled vehicles. Cycles for Electric Vehicles in the 1990's," Eleventh Intersociety Energy Conversion Engineering Con- REFERENCES ference Proceedings, Vol. II, Stateline, Nevada, Sept. 12-17, 1) Department of Energy, "Per- 1976. formance Standards for Demonstra- tion: Development of Energy's 8) Private Communication, E. Electric and Hybrid Veyicle Re- Behrin, May 23, 1978. search, Development and Demonstra- 9) O'Day, J., et al., "A Projec- tion Project,: Federal Register, tion of the Effects of Electric Vol. 43, No." 104, May 30, 1978. Vehicles on Highway Accident Sta- 2) McAlevy, R.F. Ill and Bedrosyan, tistics, " SAE Paper #780158, pre- L., "A Critical Review and Evalu- sented at SAE Congress, Detroit, ation of Published Electric-Ve- Michigan, Feb. 27-March 3, 1978. hicle Performance Data," Proceed- ings of the 13th Intersociety En- ergy Conversion Engineering Con- ference, Aug. 20-25, 1978, San Diego, CA, pp. 655-661, SAE P-75, SAE Inc., Warrendale, PA, August 1978. 3) Behrin, E. et al., Energy Stor- age Systems for Automotive Propul- sion, Vols. I and II, UCRL-52303, Lawrence Livermore Laboratory, University of California, Decem- ber 15, 1977.
4} McAlevy, R.F. Ill, "A Funda- mental Basis for Evaluating the Performance of Electric (and Other Energy-Storage) Automotive Vehi- cles and its Use in Energy Policy Analysis," Proceedings of the NATO/CCMS Fourth International Symposium on Automotive Propulsion
25 Table 1. Typical drive-cycle characteristics DRIVE-CYCLE
SAEJ227a/C 0.09 13 SAEJ227a/D 0.1 20 EPA Urban (no freeway) 0.09 23 Scott 0.12 32
Table 2. Forecasted (Ref. 3) Battery Electrical Characteristics and EV Range Required of Urban Commuter Car for Batteries to be Optimum whr (pTa.) w Battery (p.d.),£j- n (e R . ,km S s> P=0' kg ns(es)p=o'whr o.b. Ni/Fe (1985-1990) 101 59.4a 1.7 141 Ni/Fe (1990-2000) 101.8 98 1.0 240 Ni/Zn (1985-1990) 135.3 150 .9 266 Ni/Zn (1990-2000) 139.9 153 .5 0.91 263 Li-Al/FeSx (1985-1990) 115.1 100 1.2 200 Li-Al/FeSx (1990-2000) 130 110 1.2 200 Na/S (g-alum): (1985-1990) 115.1 82.5 1.4 171 Na/S (g-alum): (1990-2000) 130 123.2 1.1 218 Na/S (glass): (1990-2000) 160 150 1.1 218 Li/air: (1S90-200O) 119 357.1 0.33 727
aCap"acity at 3 hr discharge
26 PROJECT SUMMARY
Project Title: Advanced Flywheel Energy Storage Unit for a High Power Energy Source for Vehicular Use Principal Investigator: Arthur E. Raynard Organization: Garrett-Ai Research 2525 w. 190th Street Torrance, CA 90509 (213) 323-9500 Project Goals: The project goal is to determine the benefits of a light-weight, hermetically-sealed energy storage unit for vehicular applications. Project Status: The tradeoff study and detail design are completed. The test equipment and the test specimen are being fabricated for development and performance testing. The testing will verify functional compat- ibility and measure both parasitic losses and input/output losses. Testing will start in December 1978. Contract Number: University of California Purchase Order 9676603 Contract Period: May 1, 1978 - April 30, 1979 Funding Level: $470,182 Funding Source: Lawrence Livermore Laboratory
27 ADVANCED FLYWHEEL ENERGY STORAGE UNIT FOR A HIGH POWER ENERGY SOURCE FOR VEHICULAR USE
Arthur E. Raynard AiResearch Manufacturing Company of California A Division of The Garrett Corporation 2525 West 190th Street Torranee, California 90509
ABSTRACT The maximum benefits that may be gained by the incorporation of mechanical energy storage (MES) into vehicular propulsion systems are obtained by combining the concepts of load-leveling the prime energy source(s), and recovering vehicle kinetic energy by regeneration. These benefits are intuitively achievable and large, and have been indirectly demonstrated in past programs. There is a pressing naed for the verifica- tion of these benefits in a structured program designed for that purpose. The verifica- tion, or lack thereof, is needed to give direction and emphasis to future MES component and system development. This paper describes an experimental program that characterizes the magnitude of the benefits in a flywheel propulsion system. The first step in the program will be to experimentally determine the performance of a hermetically-sealed flywheel system that has been optimized for an electric or hybrid vehicle. The flywheel system has been designed and the design tradeoffs will be described.
INTRODUCTION PROJECT GOALS The need to reduce dependence on The principal goal is to provide petroleum sources for energy generation decision-making information regarding the has created a substantial interest in the benefits of a mechanical energy storage investigation and development of energy device as it applies to vehicle fuel con- storage devices, ihe flywheel energy sumption or vehicle range. Those benefits storage unit can provide substantial bene- must be of sufficient magnitude to reduce fits to transportation propulsion systems. the life-cycle cost of the flywheel pro- The flywheel can supply high power demands pulsion system below a conventional pro- and thereby provide a method for load pulsion system. The expected result of leveling the primary energy supply. This the development program is to stimulate energy supply can be an electromechanical industry to market such a s/stem and battery or a liquid-fueled heat engine thereby provide a method to reduce petro- power converter. In addition, the fly- leum dependence within the U.S. wheel can accept vehicle kinetic energy during braking (regeneration) at a rate BACKGROUND INFORMATION limited only by the transmission power capability. This method of vehicle energy A substantial number of programs have storage can also be applied to third rail, investigated flywheel systems for the pur- electrically-powered vehicles, or station- pose of load-1 eveI ing the energy, or power charged electrical Iy-powered vehicles. source, and using vehicular kinetic energy recovery to improve the system propulsive The one-year program that is efficiencies. Table 1 lists the programs described in this paper will involve that Garrett-AjResearch has managed or design, fabrication, and experimental contributed to that feature flywheel determination of the performance of an devices. A demonstrated savings of 33 advanced, hermetically-sealed energy percent was accomplished as a result of storage unit that has been sized for a braking regeneration only in a six-month typical 3000-lb curb weight vehicle. test program in The New York City subway
28 Table 1. Summary of load leveling design information to allow a state-of- demon strat i on s. the-art energy storage unit (ESU) to be evolved. The program goal is to provide PROGRAM RESULTS 1. ENERGY STORAGE CAR SUBWAY 33% ELECTRIC ENERGY an ESU with an overalI energy density of SAVINGS. NY. TEST 2. ADVANCED CONCEPT TRAIN SUBWAY DEVELOPMENT TARGET 3.0 w-hr/lb. This goal is to be achieved 33% SAVINGS for the indicated energy and power 3. UNIVERSITY OF WISCONSIN PASSENGER CAR 33% FUEL SAVINGS. F.U.O.C., OYNO TEST requirements. There were five main areas 4. HEAT ENGINE/FLYWHEEL PASSENGER CAR CALCULATED «5% FUEL ELECTRIC TRANSMISSION STUDY SAVINGS, F.U.O.C. that required investigation in order to 5. FLYWHEEL BUS STUOV URBAN BUS CALCULATED, 36S. define the ESU: FUEL SAVINGS POSTAL VAN DEVELOPMENT TARGET. 140H CYCLIC RANGE 7. BATTERY/FLYWHEEL CAR NEAR TERM DEVELOPMENT TARGET. (a) Flywheel tradeoffs ELECTRIC VEHICLE 140% CVCUC RANGE
8. HERMETICALLY SEALED VEHICLE DEVELOPMENT TARGET. ENERGY STORAGE UNIT PROPUL5ION 66% FUEL SAVINGS (b) Electrical input/output machine tradeoffs system. Two subway cars were included in the test. These cars were in revenue (c) Bearing tradeoffs service in 10-car trains. Studies have shown that it is possible to save as much as 65 percent of the fuel normally used (d) Vacuum tradeoffs in a passenger vehicle operating in a city driving mode. Other methods of improving (e) Lubrication tradeoffs fuel economy are being developed, but these other methods are complementary to The parameters investigated are indi- the flywheel system and do not replace cated in Fig. 2. the benefits associated with the use of a flywheel. • WEIGHT • RPM L ENERGY STORAGE UNIT - • WEIGHT • MATERIAL (ESUI •RPM The flywheel benefits, however, can • CONTAINMENT • EFFICIENCY CHARACTERISTIC become insignificant if the weight, the parasitic losses, and input/output effi- BEARING TRADEOFFS [/ VACUUM TRADEOFFS ciencies are non-optimum. This program •RPM * MOLECULAR PUMP •LOS5E5 PRESSURE VERSUS concentrates on t'nese factors. • SIZE LUBRICATION TRAOEOFFS FORE CHAMBER • MATERIALS PRESSURE PROGRAM SCHEDULE SUMMARY The program events are summarized in Fig. 2. Tradeoff study summary. Fig. 1. The design tradeoffs (Task 1) are complete and the results reported herein. The flywheel tradeoffs considered The hardware design (Task 2) is also com- weight, speed, material, and containment. plete and the hardware has been ordered. Two rotor rim materials were included, The results of Task 2 are summarized Kevlar and S-glass. The tip speed limits herein. were 2500 fps and 2200 fps, respectively.
MONTHS AFTER START OF PROGRAM The weight shown in Fig. 3 includes 197t t»79 MAY 1. 1»7» ' 'M j J A S o N D J F M A rotor, containment, molecular pump, and TASK 1 *lywheel housing. This weight is plotted DESIGN TRADEOFFS BBl m STUDVREPORT. against flywheel diameter and reaches a TASK 2 minimum at 42,000 rpm for the S-glass HAROWARE DESIGN - - - . "I TASK 3 12-inch diameter configuration. However, HARDWARE FABRICATION . urn B BBl BBl BBl Kevlar was chosen as the rim material TASK 4 — TESTING AND EVALUATION.. because the Kevlar rotor design has been TEST REPORT. developed and can be geometrically sealed TASKS PROGRAM MANAGEMENT u for this application. The weight differ- MONTHLY PROGRESS REPORT. QUARTERLY ORAL ence between the two materials is insig- l PRESENTATION. _ | I] \] nificant and the program risk is minimized u by utilizing an existing design concept. Fig. 1. Phase I program summary. ELECTRICAL INPUT/OUTPUT MACHINE TRADEOFFS TRADEOFF STUDY SUMMARY Five types of machines were paramet- rically designed in terms of weight, speed, The purpose of the tradeoff study range, and part-power efficiencies. They (Task 1) was to develop the parametric are identified in Table 2. Each machine
29 Table 3. Bearing tradeoffs.
BEARING POWER BIO bEARlMG BEARING BORE LOSS LIFE LOCATION MM WATTS HOURS FLYWHEEL ZOO to 30 6.680 ASSEMBLY 201 2 3B 9.770 LOWER m b S3 17.S3O WEIGHT, 7 76 "5.000 LBS 204 •0 135 I7.9JU* 302 b 90 *.no UPPER 303 i 120 _ 304 20 163 22.450 •MAT I M50C£VM
VACUUM SYSTEM DESIGN
0 10 12 14 16 18 20 22 24 The ESU has a duaI chamber vacuum FLYWHEEL RIM DIAMETER, INCHES system during operation. The flywheel chamber is run at a pressure between 1.4 Fig. 3. Flywheel tradeoffs. and 50 microns while the electrical machine chamber is at approximately 1000 microns. The low pressure is maintained Table 2. Electrical input/output in the flywheel cavity by a molecular pump machine tradeoffs. with a rotor diameter of 6 in. The design characteristics are summarized in Fig. 4. This design provides a pressure cat'\o of BEST CHARACTERISTIC MACHINE WEIGHT 1C0% SPEED EFFICIENCY 700/1 at 100 percent speed and 20/1 at 50 TYPE LB RPM % percent speed. The energy storage unit SALIENT POLE 28.5 25.000 89 SYNCHRONOUS crossection is shown in Fig. 5. ROUND ROTOR 23 36.000 83 SYNCHRONOUS INDUCTION 16 42.000 92 H0M0P0LAR 25 60.000 89 INDUCTOR
IRONLESS NO SOLUTION 14,000 - PERMANENT MAGNET best-weight, which corresponds to its weight at the indicated speed, is Iisted along with its characteristic efficiency. The induction machine, in combination with the flywheel, provides the optimum ESU from a weight, size, and efficiency Standpoint, and therefore was selected as Fig. 4. Vacuum system design. the ESU input/output machine.
BEARING TRADEOFFS
Bearings were investigated in terms of losses, sizes, and materials. The design loads criteria were imposed along with the speed that was determined from the electrical machine and flywheel opti- mization. The study results are summar- ized in Table 3.
A 202-size-bear ing made from 52100 steel was selected for the lower bearing. A 204-size-bearing made from M50 CEVM steel was selected for the upper bearing^. The M50 material yields a 17,975 hr BtO life. The total predicted maximum bearing Fig. 5. Energy storage unit toss is 188 w. cross-section.
30 LUBRICATION TRADEOFFS rotor heating caused by the waveform of the inverter. The current conductors and Five candidate oils are identified end rings are cast copper alloy with a in Table 4. As of this date, the Fyrquel conductivity related to pure copper of 60 150 appears to be the best candidate oil. percent. A flow of oil is pumped through Material compatibility tests are currently the inner diameter of the hollow rotor taking place. Results will be, discussed shaft to provide cooling. The rotor has in the next reporting period. a maximum peripheral speed of 420 fps, which is consistent with other high speed Table 4. Lubrication tradeoffs. rotor designs.
VAPOR VISCOSITY PRESSURE CENT'STOKES TORR CORROSION The stator consists of a stack of CANDIDA TES TVPE 021D°F LUBRICITY STABILITY silicon steel laminations punched to MILL 23699 SYNTHETIC 5 , 2 EXCELLENT EXCELLENT IMOBIL.STD. OIL BASE receive the stator windings. The stator PLUS OTHERS! COhAV. GRADE 32 MINERAL 46 2X1»3 GOOD POOR insulation system wilI be class 220°C and EXXON BASE OXIDATION NAPTHENIC RESISTANCE will have HML-coated magnet wire and TERESSOJ79 5 X ID 3 EXXON Nomax-Kapton-Nomax laminated slot-liner BRAYCOB15Z PER 40 1 XIO8 EXCELLENT EXCELLENT insulation. The stator will be vacuum BRAr OIL co FLUORIDATED POIV-ETHER impregnated with ML varnish. The lead FYRQUEL 150' TR1-ARYL 45 3.5X102 GOOD GOOD STAUFFER PHOSPHATE • 380OF wires will have double Teflon insulation. CHEMICAL CO. The stator 0D wiI I be ground for a FLYWHEEL ROTOR CONSTRUCTION shrink fit into the motor housing. The motor housing is machined from aluminum The flywheel rim is a filament- alloy and contains oil passages around wound composite material consisting of the motor stator to provide for stator concentric outer rings wound of Kevlar cool ing. roving with an epoxy binder, and one inner ring wound using S-glass roving with an VACUUM SYSTEM epoxy binder. The Kevlar rings have a higher strenqth-to-density ratio and are The housing for tie entire motor, superior for enerqy storage and fatigue flywheel, and pump is evacuated to life. The inner ring of S-glass provides approximately one rorr (one millimeter of for a dead loading of the outer rings and mercury). In the unit shown, which is the for higher compressive strength in the laboratory demonstrator, this process may area of contact with the hub spokes. have to be repeated occasionally because of the 0-ring seals in the housing. A The rim is mounted on a 4-spoked second evacuation may be necessary to hub, machined from 7075-T7351 aluminum facilitate disassembly for development plate stock. There is a short protrusion changes. The flywheel cavity is brought on one side of the hub to receive the to an even lower pressure of approximately bolts that mount the flywheel to the 0.001 torr by means of a molecular pump. shaft. This device, with its rotor adjacent to the flywheel hub, propels molecules up a CONTAINMENT/VACUUM CHAMBER spiral path in the stator plate, thereby causing the molecules to move from the The flywheel is mounted in an air- flywheel housing to the motor housing. tight enclosure to provide for a vacuum No rubbing seals are required. environment surrounding the flywheel. The walIs of this enclosure are of steel and LUBRICAT 10N/C00LING SYSTEM are of sufficient thickness to provide for containment of the flywheel in the Oil is circulated through the motor unlikely event of a burst failure. cooling passages, the bearings, and the oil cooler by means of a centrifugal MOTOR/GENERATOR CONSTRUCTION pump. The pump is gear driven from tne lower end of the flywheel shaft. Oil is The motor/generator is of the squir- picked up from the oil sump fn the lower rel cage induction type. The rotor con- end of the unit and discharged under a sists of a stack of silicon steel lamina- pressure of approximately 10 psi into tions pressed onto a magnetic steel shaft. passages in the aluminum housing. From The rotor laminations are of special here it passes out through the oil cooler design to minimize harmonic losses and and back into the motor housing. One
31 passage leads through an external tube to FUTURE WORK the jet in the oil sump cover. This jet shoots oil into the center hole in the The completion of the current contract rotating flywheel shaft. The high speed is based on the following objectives: of rotation causes the oil to flow on the ID of the central hole and move to areas • Prove composite rotor in of greater radius. This oil is then operating environment drained bacx into the oil sump through the outer annul us of the rotor shaft. • Design integration of all elements After the oi I reenters the motor housing from the oil cooler, it enters • Determine parasitic losses the motor stator cooling passages. After cooling the motor stator, the oiI is • Determine I/O losses returned to the sump by gravity. The split of the oil flow into the two possi- A test program will measure the ble paths is controlled by an orifice in energy storage unit performance and the housing that can be adjusted during determine the compatibility of the com- the development program. ponents with the cooling and lubrication system. The effects of the vacuum level BEARING ARRANGEMENT will be observed and measured. After these tests are completed, a series of The fIywheeI rotor i s cant iI ever- limited environmental tests will evaluate mounted on the end of the motor shaft. the energy storage unit in terms of a A 204-size ball bearing is located at the vehicle environment and operational cycle. lower end of the shaft. The bearing inner races and other elements of the rotating Follow through work to be considered assembly are clamped tightly axially for will investigate the total propulsion increased shaft cross section to provide system. This work will determine and the required stiffness. The upper bearing optimize the benefits associated with the is resilient Iy mounted to control the flywheel concept. The work must be sys- range of critical speeds, reduce noise, tematic since the results will be strongly and to increase bearing life. The lower influenced by the intended application. bearing has an axial preload spring with force that is also transmitted to the A second element of work to be con- upper bearing. Both bearings are shaft- sidered will determine the added weight cooled and mist-lubricated. and cost of the total flywheel propulsion system. ESU INPUT/OUTPUT MACHINE PARAMETRIC PERFORMANCE
For trade-off consideration, some of the machine parameters are plotted in Fig. 6. The left-hand curve shows a plot of machine efficiency versus output power. Slip values are shown for the correspond- ing power outputs. Both the high-speed case (in solid line) and the low-speed case (in dotted line) are presented.
15 30 45 15 30 45 POWER OUTPUT (KW) POWER OUTPUT |KW) HIGH SPEED LOWSPEEO Fig. 6. ESU I/O machine parametric performance.
32 PROJECT SUMMARY
Project Title: Regnerative Flywheel Energy Storage System Principal Investigator: E. L. Lustenader Organization: General Electric Company Corporate Research and Development P. 0. Box 43 Schenectady, NY (518) 385-3084 Project Goals: Laboratory test an improved flywheel energy recovery system sized for a 3000 pound class battery/flywheel electric vehicle. Tests to simulate an electric vehicle operating under the SAE J227a Schedule D driving cycle will establish the range improvements attributed to the flywheel. Project Status: The design and detail drawings are complete, and component manufacturing is underway. The flywheel drive motor/alternator will be a 20,000 rpm six-pole synchronous inductor type machine powered by an 8 SCR loan commutated inverter. The motor/alter- nator is coupled directly to a small steel disc flywheel designed to recover vehicle braking energy. The laboratory set up includes a 109 volt lead-acid battery bank, a new separately excited DC electric vehicle propulsion motor, the flywheel energy storage system (motor/flywheel, PCU and control) and a load flywheel to simulate vehicle inertia. Facility modification is underway. Contract Number: University of California Purchase Order 8990503 Contract Period: Mar. 1978 - Mar. 1979 Funding Level: $450,905 Funding Source: Lawrence Livermore Laboratory
33 REGENERATIVE FLYWHEEL ENERGY STORAGE SYSTEM
E.L. Lustenader General Electric Company Corporate Research and Development P.O. Box 43 Schenectady, New York 12301
ABSTRACT
This paper describes the progress to date on the laboratory development and evalu- ation of a regenerative flywheel energy storage system. The system has been designed specifically for a battery/flywheel electric vehicle in the 3000 pound class. Planned laboratory tests will simulate this electric vehicle operating over the SAE J227a Schedule D driving cycle. The range improvement attributed to the use of the flywheel will be established. The flywheel energy storage system will consist of a solid rotor, syn- chronous inductor-type flywheel drive machine electrically coupled to a DC battery elec- tric propulsion system through a load commutated inverter. The motor/alternator unit is coupled mechanically to a small steel flywheel which regenerates the vehicle's braking energy. The laboratory simulation of the battery/flvwheel propulsion system will include a 108 volt lead-acid battery bank, a separately excited DC propulsion motor coupled to a flywheel which simulates the vehicle's inertia, and the flywheel energy storage system comprised of the motor/flywheel unit, the load commutated inverter and its control.
INTRODUCTION AND BACKGROUND • Novel force commutated circuit for In 1974, a flywheel energy storage starting the synchronous machine from system was conceived by the General Electric standstill; Company and proposed for demonstration to the Department of Energy, Office of Energy • Novel control system not requiring the Technology, Division of Energy Storage Sys- use of rotor shaft position sensors. tems. The overall objective of the program was to demonstrate new technology associated This new technology was successfully with a flywheel energy storage system con- demonstrated on DOE Contract EY-76-C-02-4010. sisting of a composite flywheel directly coupled to an AC synchronous motor/alter- A follow-on to the program was initi- nator. The motor/alternator/flywheel unit ated on March 17, 1978. The new program was hermetically sealed with the rotating has the specific objective of laboratory assembly operating in a low pressure helium testing an improved flywheel energy recov- atmosphere. The motor/alternator received ery system sized for a battery/flywheel its power from a solid state inverter/recti- hybrid electric vehicle in the 3000 pound fier unit designed to provide the necessary class. Laboratory tests will be conducted frequency control from a constant DC battery to simulate the electric vehicle operating power source. under the SAE J227a Schedule D driving cycle. The objective is to simulate the The objective of the DOE program was vehicle operating over this duty cycle and to demonstrate the following new technology: to determine the range improvement that can be attributed to the use of a small fly- • Solid Rotor, inductor type synchronous wheel in combination with the battery bank. motor/alternator; Analysis has indicated that the fly- • Direct coupled composite flywheel of wheel/battery system can isolate the battery new "cross-ply" construction; from the acceleration power demands of the • Sealed rotor assembly operating in low vehicle and can also recover a substantial pressure helium; portion of the braking energy. Thus, the range of the flywheel/battery electric • Load commutated inverter power supply vehicle is projected to be greater than that to couple the flywheel energy package of an all-battery electric car when used in to a DC input/output power source; a repetitive start-stop driving cycle. SCOPE OF CURRENT PROGRAM OVERALL VEHICLE SYSTEM STUDIES
In achieving the above objective, an In analyzing the performance of a improved flywheel energy recovery system flywheel/battery powered vehicle operating has been designed and is being fabricated. on a duty cycle such as the SAE J227a Sched- There are nine major technical tasks to be ule D, a multitude of operating nodes can accomplished. The scope of the overall be assumed. On this contract, five modes program can be summarized as follows: of operation were selected for consideration. These operating modes are illustrated in • Establish specifications for a regen- Fig. 1. erative flywheel energy storage system and prepare a test plan. These speci- In Mode 1 all of the braking energy fications would be based on tradeoff stored in the flywheel is used to supply studies between system weight, compo- the drive motor armature power during the nent efficiency, and performance. initial stage of acceleration. This results in power not being required from the bat- • Design a new, improved, all steel, tery by the motor armature until some point inductor motor/flywheel energy storage beyond the "cornering point." "Cornering package for a 3000 pound class electric Point" is defined as the vehicle speed at vehicle. Test results from the first which the back EMF of the propulsion motor phase of the DOE program would be uti- equals the battery voltage. This elimin- lized in arriving at this design. ates the need for an armature chopper in the battery circuit, thus resulting in a • Test the regenerative flywheel energy cost and weight saving as well as elimina- storage system in the laboratory. In ting the losses ssociated with the chopper. order to simulate an actual driving In this mode of operation, as well as in cycle, a load flywheel would be de- signed to simulate the inertia of the 3000 pound vehicle. This would be coupled directly to a new separately excited DC propulsion motor.
The fourteen SCR inverter/rectifier circuit which was demonstrated on the first contract was completely redesigned and replaced by a new 8 SCR unit with power con- tactors. Field reversal in the separately excited propulsion motor obviates the need for the six additional SCR's originally proposed for braking.
The simulated propulsion system will consist of a 108 Volt lead-acid battery bank, a DC propulsion motor with a load flywheel, and the regenerative flywheel energy storage system. Tests will be conducted on the equipment to simulate the vehicle in oper- ation. The equipment will be operated to establish the performance of the regenera- tive flywheel energy storage system as if it were performing in a 3000 pound elec- tric vehicle. It will be operated over simulated driving cycles, and measurements will be made to determine the energy re- quired per cycle. Data will be reduced and the results analyzed and compared to the predicted performance. Results will pro- vide an estimated range and power consump- MOOES tion for a 3000 pound class electric vehicle with this type of flywhee1/battery propul- Note: O denotes "cornering point" sion package. Fig. 1. Power-Time Profiles for Various Driving Cycles
35 other modes studied/ it is assumed that the • Aerodynamic drag coefficient; batteries will, at all times, supply the • Vehicle frontal area, drive motor field power and all other aux- iliary power requirements. • Number of transmission speed ranges (gears), Mode 2 is similar to Mode 1 except • Speed ratio for each gear, that once the cornering point is reached, the remaining available flywheel power is • Rolling resistance drag coefficients, used uniformly over the remainder of the • Wheel inertia, accelerating period with the additional required armature power being supplied by • Transmission/final drive efficiency, the battery. In this way '.he need for an • Motor design parameters (as defined armature chopper is still elminated but the by the motor model), peak battery current is reduced with a re- sulting improvement in battery output. • Battery design parameters (as defined This is accomplished, however, at the ex- by the battery model), pense of added control complexity. • Electrical losses in drive train (as defined by field and/or armature In Mode 3 all of the braking energy chopper models), stored in the flywheel is used to reduce the peak battery current. In this case, • Auxiliary power losses (including an armature chopper is required in the bat- accessories and ventilating blowers). tery circuit since battery power is used prior to reaching the cornering point. COMPUTER PROGRAM MODIFICATIONS However, since peak battery current is FOR THE FLYWHEEL/BATTERY VEHICLE lower, total battery output will increase. Figure 2 shows schematically the over- In the last two modes of operation. all system modelled for computer evaluation. Mode 4 and Mode 5, load levelling is used. The flywheel package was considered as a This is accomplished by using the battery single subsystem, containing the motor/ to charge the flywheel during the idle and/ alternator/flywheel and power conditioning or cruise segments of the cycle. This equipment. As such, only one parameter, results in minimizing the peak battery cur- the combined in/out efficiency, was required to define this subsystem. This value became rent. The tradeoff in this case, however, an input to the program. It was assumed is between the improved battery output due that the size and speed range of the fly- to lower peak current and the losses asso- wheel would be designed to be sufficient to ciated with charging the flywheel from the store the braking energy. battery. With the exception of the load levelling feature, Mode 4 is similar to Mode 2, and Mode 5 is similar to Mode 3. RESULTS OF THE VEHICLE SIMULATION
PERFORMANCE SIMULATION The battery/flywheel vehicle assumed for these simulation runs was taken as a In order to select the optimum mode of modification of the all-battery electric operation and to carry out tradeoff studies vehicle which General Electric is develop- between vehicle weight, battery weight, ing under a DOE/JPL contract. The vehicle flywheel weight, flywheel in/out efficiency, has an empty weight of 2000 lbs, a battery etc., computer simulation runs were made. weight of 1,100 lbs, and a passenger load of To accomplish this, four digital simulation 600 lbs for a gross total weight of 3700 lbs. programs were set up; each one was derived Load Leveling CMM Only - Mode* 4 and 5 by making extensions to an existing computer program currently being used for design optimization performance predictions of the electric vehicle being developed by the General Electric Company under DOE Contract. The modified programs are flexible enough to accomodate a variety of vehicle para- inertl* meters including: • Vehicle gross weight. • Wheel rolling radius, AtroDrao • Final drive ratio, Fig. Overall System Model
36 In assessing the results of the vehicle A typical set of results is shown in simulation runs, two criteria were used: Figs. 3 and 4. These computations are First, the actual range for a repetitive based on the Mode 3 type of operation indi- J227a Schedule D cycle, and second, how cated in Fig. 1. Since the flywheel energy this range canpares with that computed for storage system had not yet been defined, the pure battery electric vehicle. In the its weight was unknown. Therefore, com- latter case, no regeneration into the bat- puter runs were made with the assumption tery was assumed for a referenced vehicle that the flywheel propulsion system would since the amount of regeneration possible weigh betweeen 0 and 200 pounds more than is not precisely known at this time. the pure battery (chopper controlled) pro- pulsion system. Originally, it was proposed that the total weight of the vehicle be kept con- Calculations showed that for any fixed stant even though a flywheel energy storage weight increase due to the flywheel package package may weigh more than the chopper differential the vehicle performance was control system used in a conventional bat- slightly better if the battery weight were tery electric vehicle. The constant vehicle kept constant as opposed to the gross weight weight would be accomplished by removing of the vehicle being kept constant. If the one or more of the propulsion batteries. vehicle gross weight were held constant, a Computer runs for constant vehicle weight slightly higher flywheel in/out efficiency were made, but runs were also made in which was required in order to break even with a the vehicle gross weight was allowed to pure battery vehicle. The same conclusion increase by the amount by which the fly- was found for other modes of operation. wheel package weight exceeded the weight of the conventional propulsion system. In , The shortcoming of Mode 3 and Mode 5 these latter runs, the comparison with the operation is that an armature chopper or pure battery electric vehicle was made on some other means of controlling the DC volt- the basis of equal gross weight with the age to the traction motor is required. increased weight of the battery vehicle Realistically, this is also a shortcoming _being made up of additional propulsion of Mode 1 and Mode 2 operation since little batteries. energy may be left in the flywheel after the
J227a Schedule D Driving Cycle MODE 3 Operation Vehicle Gross Weight Constant at 3700 Pounds J227a Schedule O Driving Cycle 100 Battery Weight Constant at 1100 Pounds
90 3700 Ib Vehicle (0 wt flywheel) 3800 Ib Vehicle (100 Ib flywheel) 3900 Ib Vehicle (200 Ib flywheel)
SO
40 0 20 40 60 80 100 0 20 40 60 80 100 In-Out Efficiency of Flywheel Energy Storage Package (%) In-Out Efficiency of Flywheel Energy Storage Package (%) Fig. 3. Effect on Vehicle Range for Fig. 4. Effect on Vehicle Range for Constant Vehicle Weight Constant Battery Weight
37 cornering point is reached even under ideal conditions. For lower flywheel package efficiencies or for interrupted driving 110 cycles, there may be insufficient flywheel energy to reach the cornering point. Mode 4, however, avoids this shortcoming by pro- 100 Zei > Weight Flywheel Propulsion Package Differential viding for the transfer of energy from the 3700 Ib Total (vehicle and battery to the flywheel at a low current passengers) level during the idle period and as a further option during the cruise period. 90 In this way, the flywheel has sufficient energy at the start of the cycle to reach v200 Ib Flywheel Propulsion the cornering point and make a substantial 80 Package Differential reduction in peak battery current under all 3900 Ib Total (vehicle conditions. and passengers)
70 - SELECTED OPERATING MODE Mode 4 - 7.307 Gear Ratio Optimum Battery Current During Idle Operating Mode 4 with load levelling during the idle period was, therefore, 60 - selected as the operating mode to be used - Basic 3700 Ib Vehicle w/o Flywheel Package or Battery Regeneration for the remainder of this study. Based on this operating mode, a computer simulation study of the J227a Schedule D driving cycle 50 -O- 60 80 100 was made to predict the vehicle range for a single battery charge. In/Out Flywheel Efficiency (%)
As a further output of the computer Fig. 5. Flywheel Augmented Vehicle Range runs, detailed calculations of voltages, currents, etc., at discrete time intervals were made, thus allowing subsequent design assumed 50 amps idle recharge current, a studies of the various system components. flywheel stored energy of 105 watt hours, The in/out efficiency of the flywheel pack- and a gear ratio of 7.307. age was assumed to lie somewhere between 60 and 100%. The recharging of the flywheel FLYWHEEL ENERGY STORAGE PACKAGE during the vehicle idle period was at a constant current for a 25-second interval. ELECTRICAL DESIGN The idle period charging current was varied parametrically to determine the optimum As far as the flywheel drive machine flywheel recharge. Range was determined for is concerned, the most difficult operating two propulsion motor/gear ratios, 5.48 and point of the J227a driving cycle occurs 7.307; the former corresponding to a vehicle during vehicle braking and flywheel motor- speed of 60 mph at full rated motor speed, ing. From an electrical point of view, the and the latter to 45 mph at the same motor most severe machine load occurs at 145% of speed and thus is a "lower" gear. The 7.307 the base (or minimum operating) speed, gear ratio (lower gear) proved to be better where the commutation requirements are 230 adapted to the Schedule D driving cycle as amperes at 108 volts in the DC link. This the cornering point of the traction motor particular point is, therefore taken as the is reached at a lower vehicle speed. critical design point for electrical trade- off studies. Initial computer studies show that the 7.307 gear ratio achieved approximately 5 miles better range than the 5.48 ratio. NUMBER OF MAGNETIC POLES Therefore, the remaining computations were done for a 7.307 gear ratio. The inductor machine which was built on the first DOE Contract (EY76-C-02-4010) Plotting the calculated range vs. the had 8 poles. That design was chosen as a idle time battery current showed a maximum compromise between mechanical and elec- range occurred with between 50 to 60 amps trical requirements. Maximum power con- fed to the flywheel package. This maximum ditioning frequency was limited to 1000 Hz range obtained at optimum idle recharge and because of flywheel weight, the maxi- current is plotted in Fig. 5. On the basis mum rotor speed was set at 15,000 rpm. For of these results, the detailed design study this application, a small steel flywheel is
38 the tradeoff between weight and motor/ alternator efficiency, electrical loss reduction was assessed in terms of overall vehicle weight. Vehicle system studies showed that an energy storage system efficiency improvement of 1 percentage point during braking regeneration was equivalent to approximately 25 pounds of vehicle weight This factor was considered in the overall tradeoff.
The electrical machine finally chosen from the tradeoff is a 6-pole machine operating with a maximum speed of 20,000 rpm, with a rotor radius of 3.6 inches and a rotor length of 3.44 inches.
MECHANICAL DESIGN
The basic mechanical design of the flywheel energy storage package now under construction is shown in Fig. 6.
The inductor machine portion of the rotating assembly, including the six poles and the central shaft, is machined from a magnetic steel billet (AISI 4340). Steel shafts of non-magnetic, austenitic steel are welded to both ends of the pole bearing section in order to minimise flux leakage, which tends to saturate the magnetic iron and magnetize the ball bearings, making them traps'"for magnetic wear particles. The stub shafts are hollow to reduce weight.
The rotor shaft is vertically oriented in order to minimize gyroscopic effects which would occur if the unit were operated in a vehicle. Angular contact ball bear- ings are used. The upper bearing supports the weight of the rotor while the lower bearing acts as a guide with a spring con- trolled preload. Bearing lubrication and cooling is achieved by an oil jet directed Fig. 6. Machine with Maximum Design on the inner race. Most of this oil is Speed of 20,000 RPM thrown free of the bearing and returns to the sump by a parallel path so as to mini- used since less stored energy is required. mize churning losses. The lower bearing loads in this case allow higher rotor speeds. Thus, still within The oil is circulated by an internal the constraint of maximum frequency of 1000 gear pump driven through a gear reducer Hz, it is possible to design either a 6-pole from the bottom of the main shaft. Because machine operating between 10,000 and 20,000 of the low atmospheric pressure in the rpm or a 4-pole machine operating between rotor enclosure, the pump will be operated 15,000 and 30,000 rpm. at low speed to prevent cavitation at the pump inlet. At full motor speed, the pump Design tradeoff studies were performed will operate at approximately 300 rpm. The for 4-, 6- and 8-pole machines. The load pump and gear reducer housing also serves point at 145% of base -speed for the flywheel as the oil sump and reservoir. machine during motoring was taken as a design point. Results showed that the 6- pole machines provided the lightest weight for this application. In order to assess
39 FLYWHEEL WINDAGE AND BEARING LOSSES POWER CONDITIONER & CONTROL
The outer sides of the machine poles POWER CONDITIONER are shrouded to reduce windage losses. Windage losses due to the flywheel can be The power conditioner, which is a load computed with reasonable accuracy. However, conunutated inverter/rectifier, is based on the windage loss due to the pumping action the system provided by General Electric of the lobed rotor can only be estimated on under the previous Contract (EY-76-C-02- the bases of test results from the previous 4010). However, the new unit will be machine. For the 6-pole machine, the pres- lighter and smaller than the original load sure within the housing will be set at 0.01 coounutated inverter (LCI) and will require atm helium. At this pressure, the direct only six power thyristors instead of twelve. flywheel windage loss is predicted to be 38 Reversal of power flow is provided by a watts at the maximum rotor speed of 20,000 hybrid reverser (2 diodes and 2 contactors) rpm. This reduces to an average loss of 18 rather than a second set of six thyristors. watts over the flywheel duty cycle. A block diagram of the entire system is shown in Fig. 7. The four major components The average windage loss of the lobed for the electrical portion of the drive rotor will be approximately 35 watts. When system (battery, propulsion motor, hybrid the average rotor loss (35 watts) is added reverser, and load commutated inverter/ to the average flywheel loss (18 watts) the rectifier) are shown together with the total average estimated windage loss is 53 electromagnetic contactors that connect the watts. various components in the several modes in which the system operates. A contact DESIGN STRESSES closure sequence for each mode is shown in Table 1. The designation Vj in Table 1 is The energy stored in the inductor rotor the motor rpm corresponding to the con- is approximately 15 watt-hours over the dition when the propulsion motor voltage equals the battery voltage. That speed is speed range of 10,000 to 20,000 rpm. An a function of the state of charge of the additional 90 watt hours will be stored in battery and does not correspond to a fixed the flywheel. This flywheel will be a speed for all driving cycles. multidisk shrunk on design made of vacuum melt AISI 4340 steel. It will operate at a relatively modest stress level in order Table 1. Contact Closure Sequence to produce a factor of safety of 2.0 rela- tive to the 10? cycle curve for alternating Contactor stress. .Although a higher stress design Mode 1 2 3 4 would be possible, it was not considered worthwhile in this case since doubling the Motoring (Stop-V^) X design stress would only remove approxi- Motoring (V^-Top) X X mately 10% of the total package weight from the flywheel. The flywheel weight reduc- Braking (Top-Stop) X X tion would be counteracted by the neod for Recharge at Stop X X a heavier containment ring. The hig;ier stressed wheel would also be larger dia- Initial Starting of meter and occupy more vehicle space. Flywheel X X X = contactor closed Contactor numbers are shown in Fig. 7
Figure 8 shows the power and auxiliary thyristors of the load commutated inverter/
Commutating CapKllor
Six flower Two Auxiliary Thyristora Thyristors Fig. 7. Schematic Diagram of Simulated Fig. 8. Load Commutated Inverter/Rectifier Propulsion System Power and Coiranutating Circuit 40 rectifier. In this system, power flow is accomplished by reversing the direction of J227» Schedule D the DC voltage while maintaining the current flew from the thyristors in the same direc- tion. The two auxiliary thyristors and the single commutating capacitor connected to the synchronous motor neutral terminal for starting are similar to that provided in the t previous contract. 10
This configuration of the power circuit 10 20 30 40 SO 70 80 90 100 110 120 was chosen in order to provide the capa- bility to recharge the flywheel from the Time (seconds) battery during conditions of either zero speed operation or full speed operation. Fig. 10. Flywheel Per Unit Speed vs Time
The electrical requirements of the load commutated inverter (relating to its J227a Schedule D DC side) when operating over a simulated SAE J227a Schedule D driving cycle are given in Figs. 9, 10, and 11. Positive.current indicates power flow from the flywheel to the propulsion motor. Negative current indicates power flow from the propulsion motor, acting as a generator during braking, to the flywheel or from the battery to the flywheel. During the condition in which the battery is supplying power to the pro- pulsion motor or to the flywheel motor, the 10 20 30 40 50 60 70 80 90 100 110 120 voltage can be in the range of 75 to 108 Time (seconds) —»» volts depending on the state of battery charge. Figs. 9 and 11 provide the informa- Fig. 11. Inverter DC Voltage vs Time tion necessary to select the power semi- conductors for the inverter/rectifier and each major component and measure the energy their cooling requirements when repetitively storage efficiency of the flywheel energy operating over the simulated driving cycle. storage package so as to determine the energy improvement to the electric vehicle. 300 250 A The control system will consist of an 200 10 sec inverter current regulator with an internal 125 A feedback loop using motor operating angle 100 to maintain synchronism of the inverter to 28 sec the motor.
10 20 30 40 50 60 70 80 90 100 110 8 In addition to the primary control Time —•- system, two field current controls are S -100 required. One is for the flywheel motor, which will simply vary the flywheel motor -200 field current as a fixed function of in- verter DC link current, flywheel speed, and -300 DC link voltage. Fig. 9. Inverter DC Link Current vs Time The second is for the traction motor field which will remain fixed at low vehicle CONTROL speeds but must be varied at high speeds to regulate the armature current during the The system control is required to period of time that the battery supplies regulate the operation of both the energy all the power to the propulsion motor. storage flywheel system and the vehicle traction motor. A number of variables in The major part of the system control the power circuit will be sensed to accom- will be implemented using the laboratory plish this. In addition, it is desirable hybrid controller. It allows for easily to establish the operating efficiency of made changes in control system configura-
41 tions and parameters as the test program REFERENCES develops. 1. R.H. Guess, & E.L. Iiustenader, SYSTEM TEST "Development of a High Performance and Lightweight Hybrid Flywheel/Battery Powered An existing General Electric laboratory Electric Vehicle Drive," Fourth Inter- test facility is currently being modified national Electric Vehicle Symposium, for use on this contract. This laboratory Dusseldorf, Germany, August 1976. setup will include a new separately excited DC propulsion motor, a flywheel simulating 2. E.L. Lustenader, "Flywheel Energy vehicle inertia, an electrical load machine Storage System Development," Flywheel to provide road loss, and a torque trans- Technology Symposium, San Francisco, CA., ducer to measure the propulsion motor October 1977. : torque. This equipment together with a 108 3. E.L. Lustenader, G. Chang, E. volt battery bank, the inverter/rectifier, Richter, F.G. Turnbull, J.S. Hickey, the inductor motor/alternator/flywheel pack- "Flywheel Module for Electric Vehicle age, and the hybrid controller will be Regenerative Braking," 12th Intersociety assembled in the laboratory for component Energy Conversion Engineering Conference, and system testing. Washington, DC, August 1977. The system will be operated in 1979 to 4. A.B. Plunkett and F.G. Turnbull, determine the performance of the regenera- "Load Commutated Inverter/Synchronous tive battery/flywheel energy storage pro- Motor Drive Without a Shaft Position pulsion system as if it were operating in Sensor," 1977 IEEE/IAS Annual Meeting a 3000 pound battery electric vehicle. Conference Record, October 1977, Los Performance will be established for the SAE Angeles, CA., IEEE Publication No. 77 J227a Schedule D driving cycle. CHI246-6-IA, pp. 748-757. 5. R.L. Steigerwald and T.A. Lipo, "Analysis of a Novel Forced Commutation Starting Scheme for a Load Commutated ACKNOWLEDGEMENTS Synchronous Motor Drive," 1977 IEEE/IAS Annual Meeting Conference Record, October The author wishes to acknowledge the 1977, Los Angeles, CA., IEEE Publication technical contributions made to this pro- No. 77 CHI246-8-IA, pp. 739-747. gram by the following General Electric Corporate Research and Development 6. E.L. Lustenader and E.S. Zorzi, personnel: "A Status of the 'Alpha-Ply' Composite Flywheel Concept Development," Society for Mr. I.H. Edelfelt, System Analysis and the Advancement of Material and Process Computer Simulation Engineering, 1978 National SAMPE Symposium. Mr. D.W. Jones, Flywheel Energy Storage, 7. A.B. Plunkett and F.C, Turnbull, Mechanical Design "System Design Method for a \oad Commutated Inverter-Synchronous Motor Drive," IEEE Dr. A. Plunkett, Control System Design Industry Applications Society Annual Dr. E. Richter, Electrical Design of Meeting, October 1-5, 1978, Toronto, Canada. Inductor Type Synchronous Machine 8. R.L. Steigerwald, "Characteristics Mr. F.G. Turnbull, Load Commutated inverter of a Current-Fed Inverter With Commutation and Drive System Design. Applied Through Load Neutral Point," GE Report 78CRD162, August 1978.
October 1978
42 PROJECT SUMMARY
Project Title: Low Cost Flywheel System Demonstration Principal Investigator: D. W. Rabenhorst Organization: Applied Physics Laboratory The Johns Hopkins University John Hopkins Road Laurel, MD 20810 (301) 953-7800 Project Goals: Develop and evaluate a flywheel capable of storing 20 watt hours per dollar. Demonstrate this flywheel in a complete 115 vac system (1 KUH, 2 KM). Develop and evaluate low drag, long life bearing systems. Project Status: All principal objectives are expected to be met. Approximately 60 percent of the scheduled work has been completed in accordance with the revised schedule of May 1978. Contract dumber: EC-77-C-01-5085 Contract Period: Oct. 1, 1977 - Mar. 31, 1979 Funding Level: 5355,190 Funding Source: Department of Energy, Division of Energy Storage Systems LOW COST FLYWHEEL DEMONSTRATION
D. W. Rabenhorst The Johns Hopkins University Applied Physics Laboratory Johns Hopkins Road Laurel, Maryland 20810
ABSTRACT The current Applied Physics Laboratory/Department of Energy Program involves the demonstration of a very low cost flywheel ($50/kwh) in a complete energy storage system. The program also includes the develop- ment and evaluation of low loss, long life bearing systems. The energy storage system, operating at 115 VAC, has a capability of storing one kilowatt-hour of energy and of accepting or delivering this energy at an average rate of two (2) kilowatts power level. A final output of the program is the extrapolation of the flywheel design to a full size unit with a storage capability of 10 to 100 kwh. In addition to meeting the program objectives, this program has resulted in the following notable achievements in related flywheel technology: - A flywheel configuration has been demonstrated which allows the exploitation of a wide variety of applicable low cost materials in a very inexpensive fabrication process. - A flywheel suspension system has been developed which reduces bearing loads by more than 90%, thus permitting the use of smaller bearings at lighter loads and longer projected life. Resulting bearing performance approaches that of a three plane magnetic suspension—without the disadvantages of the magnetic system. - Several novel bearing systems were evaluated which have exhibited further reduced losses while having an order of magnitude longer predicted life. - A number of novel low cost flywheel materials were evaluated which promise future flywheels having even lower cost.
INTRODUCTION kilowatt hour). The Applied Physics Laboratory. 2. Evaluate the character- The Johns hopkins University (APL) istics of this low cost flywheel. is currently engaged in a 15-month program with the Department of 3. Demonstrate the flywheel Energy (DOE) which has as its in a complete home type energy primary objective the feasibility storage system having one kilo- demonstration of a flywheel energy watt hour storage at a rate of storage system utilizing a very approximately two kilowatts power. low cost flywheel. The principal objectives of this program are as 4. Develop and evaluate long follows: life low drag bearing systems for use in flywheel systems. 1. Demonstrate a flywheel capable of storing one kilowatt 5. Provide design projec- hour at a cost of less than 20 tions into energy storage systems watt hours per dollar ($50 per of interest to DOE (e.g. 10 kilo- watt hour, etc.). a considerable amount of matrix weight. 6. Investigate and evaluate potential low cost flywheel The hundreds of flywheel and materials. material spin tests which followed the original bare filament fly- In addition to the final wheel concept have demonstrated report on this program covering that there are a number of addi- the foregoing subjects, the tional advantages to the APL bare principal product of the program filament flywheel configuration. will be a complete energy storage The problems of expansion and con- system operating at 115 volts traction of the wound filaments input capable of receiving one as the flywheel rotational speed kilowatt hour of energy, storing is increased or decreased are this energy for an unspecified reduced to an absolute minimum. period of time, and delivering In the bare filament configuration this energy to a load at 115 volts the bulk of the wound filaments AC. This energy storage system are free to expand and contract will demonstrate the general without interfering with one feasibility of nighttime energy another or transferring loads to storage in an individual home. one another. This expansion and contraction is apparently done FLYWHEEL DEVELOPMENT very evenly, in view of the ex- tremely large number of filaments The Applied Physics Laboratory involved. A further advantage has been engaged for a number of arising from the relatively inde- years in the development of fly- pendent action of the individual wheel configurations which not filaments is the fact that it is only permit optimal use of the theoretically impossible to have filamentary materials but will a simultaneous failure of all of permit their use in configurations the filaments in the flywheel. having an absolute minimum fabri- A catastrophic failure of the bare cation cost. A detailed descrip- filament wound flywheel has never tion of the APL bare filament fly- been experienced in any of the APL wheel configuration is contained spin tests to date. In contrast, in the referenced report,1 where the typical failure pattern is it was concluded that the bare that the outer fibers will fail, filament configuration offers the leaving the inner fibers compris- highest possible energy per unit ing the major portion of the fly- weight, energy per unit volume, wheel assembly, intact. It has and energy per unit cost of any also been demonstrated that the known flywheel configuration. failure of the outer wound bare The principal reason for this is filament fibers in the APL fly- that the performance of any wound wheel does not necessarily mean flywheel is a function of the that the flywheel will go out of strength to weight ratio of the balance and be destroyed. On material used in its construction. several occasions a relatively The typical wound multi-ring fly- large proportion of the fibers wheel construction involves a was destroyed, but the rest of filamentary high strength material the flywheel remained intact, and in a polymer matrix, with the was brought satisfactorily to ratio of filament to polymer rest without further damage. On seldom exceeding 70%. In the APL numerous occasions the flywheel configuration, on the other hand, hub assembly was reused two, the ratio of high strength fila- three, or even four times follow- ments to polymer matrix is gener- ing a flywheel spin test to ally of the order of-98 to 99 destruction. percent. Thus, the performance of the APL flywheel can be a It would appear, in view of function of the strength to weight the foregoing, that the APL bare of the high strength filament filament configuration, in addi- itself, without the detraction of tion to its performance potential
45 offers exceptional safety charac- serve as a lubricant in the fabri- teristics compared with other cation and processing operations. types. Although the hose wire is marginal in energy to cost potential, it MATERIALS EVALUATION was nevertheless selected as the primary material in the APL pro- Of all the materials investi- gram, in view of the fact that its gated for use in the subject pro- strength and cost characteristics gram, three were selected as having are easily determined, whereas the the appropriate characteristics in corresponding characteristics of terms of performance, cost, fabri- the other materials considered cation simplicity, availability involve a considerable amount of and safe failure mode. Although theoretical projections. the three materials selected appear to have satisfactory char- The third material selected acteristics in all of these areas, for the bare filament flywheel they represent respectively very configuration is Metglas®. Gener- different material configurations. ically. Metglas has been termed The vinyl impregnated fiberglass as a metallic alloy with glass- strand material, hereafter refer- like properties. The properties red to as vinyl-glass, actually is of most interest to the flywheel a bundle of as many as 1000 fibers application, however, are the fact each having a diameter of .0003 that the Metglas can be made to inch. The presence of the vinyl have a high strength to weight in this configuration serves to ratio at a relatively low projected prevent abrasion among the fila- cost. It also offers the prospect ments, and also, to some degree, of providing these capabilities excludes water vapor from the in a configuration which also glass filaments. In its optimum qualifies on the issue of safety configuration the vinyl-glass in the failure mode. The Metglas would be processed immediately used in the program to date has after drawing the molten glass been a one-half inch wide ribbon from the platinum crucible, even having a thickness of .002 inch. before the glass strand has This ribbon form gives the Metglas received any aqueous lubricant, a distinct advantage over the which normally protects it as it other two materials mentioned, in is routed over numerous pulleys that the winding time to fabricate in the manufacturing process. a flywheel of a given size can be as much as a hundred times less In the flywheel application than the corresponding winding the vinyl impregnated glass times of the other materials. appears to have a considerably Eventually this could have a con- higher usable strength than ordi- siderable effect on the fabricated nary uncoated fiberglass, and this cost of the flywheel in question. leads to higher flywheel perfor- mance and/or lower flywheel cost. VINYL-GLASS FLYWHEEL Usable strengths of the order of 50% of the ultimate tensile The vinyl-glass used in the strength are believed to be pos- subject program was provided at sible with this material configu- no cost to the program by PPG ration. Industries, Inc., Glass Research Center, Pittsburgh, Pennsylvania. The second material which The material used up until the qualified for the low cost fly- present time is 50% vinyl and 50% wheel bare filament configuration glass, primarily because this is steel wire, such as that used material in that form was avail- in pneumatic and hydraulic hose able from other current applica- reinforcement. This wire, called tions. PPG has indicated, however, hose wire, has a diameter of .012 that it will be a relatively to .015 inch, and is usually simple matter to produce this plated with a very thin costing of material in a configuration con- brass to protect the wire and to sisting of 10% vinyl and 90% glass
46 especially for the flywheel appli- process, it may result in effective cation, and this material is partial bonding of these strands expected to be made available in the flywheel, which is an unde- later in the current program. sirable feature of this flywheel configuration. This effect will A typical vinyl-glass test be evaluated in more detail in the flywheel is illustrated in Fig. 1. balance of the current program. METGLAS FLYWHEEL The typical Metglas test fly- wheel is illustrated in Fig. 2
Fig. 1 Vinyl-glass flywheel.
Here it can be seen that the vinyl-glass is wound in a very dense pattern, and gives the Fig. 2 Metglas flywheel. appearance of being bonded throughout. Actually the only The Metglas provides an extremely areas of polymer bonding are at dense, clean and attractive wound the four narrow band radial wrap configuration, where the Metglas positions, and everywhere else on occupies essentially 100% of the the flywheel the vinyl-glass wound structure. This is in con- strands are in the so-called bare trast to the fiber and wire con- filament condition. The vinyl- figurations, where the material glass appears to be an extremely occupies only about 80% of the tough material. On at least one wound structure. occasion the flywheel has been re-spin-tested with only minor The Metglas is a proprietary repairs, after having been inad- material in experimental production vertently dropped to the bottom at the Allied Chemical Corporation of the spin chamber at 20,000 rpm, of Morristown, N. J. All of the and was bounced around inside the Metglas used in the current pro- chamber for several minutes. gram was provided at no cost to the program by Allied Chemical. One characteristic of the The material provided to the vinyl-glass which represents a present time is a ferrous alloy distinct advantage during the whose general strength character- winding and handling processes, istics correspond to an equivalent may actually turn out to be a steel. The mcnufacturing process disadvantage in the final analysis. involves extremely rapid chilling This property is the "tackiness" of the molten material at a rate of the vinyl-glass strand as one of approximately two million turn is laid on top of another in degrees centigrade per second. the winding process. While this The result is a smooth metallic condition greatly facilitates the ribbon which has surprisingly
47 uniform dimensions and physical at the Applied Physics Laboratory characteristics. The configura- was received at no cost to the tion of the Metglas currently program from two separate sources. being processed into flywheels is The first source is the Central called the "As-Cast" condition. Special Studies Group of Industrie In this condition the Metglas Pirelli S.p.A. in Milan, Italy. apparently has microscopic cracks The second source is the National along the edges, which prevent Standards Company in Niles, achievement of the ultimate physi- Michigan, U.S.A. cal characteristics of this material. A process has been The fabrication of the steel developed at Allied Chemical wire bare filament flywheel has wherein these microscopic cracks involved the development of special are removed without appreciably techniques to accommodate the adding to the production cost of peculiar characteristics of this the material. Following this steel wire. Early attempts to process the Metglas ribbon is wind the bare filament steel fly- expected to have usable strengths wheels were unsuccessful because of the order of 60% higher than of these characteristics; however, the usable strengths in the."As- the problems have be*~>n resolved, Cast" material received to date. and successful configurations are Nevertheless, the experience now a standard achievement. Three gained in processing and spin principal problems were as follows. testing flywheels of the "As-Cast" First, a satisfactory wire termina- Metglas material has been tion scheme had to be developed extremely valuable, and has which would allow the termination allowed the development of fabri- of the final end of the wire on cation techniques which will apply the inside of the flywheel, in to the ultimate Metglas material. order to permit the maximum utili- zation of the tensile strength of STEEL WIRE FLYWHEEL the wire. The method developed was to pre-machine a helical slot The typical steel wire fly- occupying one quarter of a revolu- wheel is illustrated in Fig. 3 tion around the periphery of the flywheel hub. Thereafter, upon completion of the winding of the steel wire, a final revolution was hand wound to the center of the wound structure, whereupon the end of the wire was fed through the pre-machined slot and bonded securely. This method has proven to be quite satisfactory, and, in fact, was later adapted for use in the vinyl-glass flywheel described above. The othex* two problems concerned the "slickness" and wire- cast conditions. It was found that these two problems working together resulted in a wound flywheel structure which was statically unstable, when removed from the winding mandrel. After a series of experiments, and careful con- Fig. 3 Steel wire flywheel. sultation with the National Standards Company, appropriate The wire is helically wound on to wound geometries and processes the flywheel hub in a manner sim- were established which eliminated ilar to that used for the vinyl this problem. glass flywheel described in the foregoing. The steel wire used in the current flywheel testing
48 FULL SIZE FLYWHEEL The material selected for the hub was the most effective studied The characteristics of the in terms of structural capabili- full size flywheel to be used in ties, damping qualities, minimum the final energy storage demonstra- fabrication cost, and minimum tion system are illustrated in material cost. Ic. this case the Fig. 4. material is multi-ply Baltic Birch plywood. Although this material typically has plugs and Tapered hub plates Flat section is 1.5% occasional voids internally, it \Subcircular = 9.6 in. radius has proven to be a very consistent Fiberglass caps material, probably because of the Radial wrap Hub nominal fact that in the total hub struc- over these radius = 9.81 in. ture, there are 27 plys of material. In fabrication a rigid compression ring is formed of the winding material itself. The steel wire is wet wound for the first one quarter inch radial dimension, which forms a stiff ring bonded Wet wind to the plywood hub material .-*JuTr throughout its periphery. There- Steel wire .75 in.-*J' after, the winding is continued (2.25 x 2.25 in.) 1? in _ I J without resin except for the very thin radial bands in the four Fig. 4 Full size one kWh rotor. radial wrap positions. Although the steel hose wire was selected as the primary material for this full scale fly- wheel, it will be of interest to The physical dimensions of this compare the properties of this flywheel are those required to flywheel with those of similar satisfy the rotational speed ones made of vinyl-glass and requirements of the motor genera- polished Metglas. This comparison tor and control system. In this is made in Table 1. case a maximum rotational speed of 14,400 with a four to one Table 1 speed ratio resulted in a minimum rotational speed of 3600 rpm. Low cost flywheel materials comparison. Then, knowing the maximum design 10/90 Polished Steel hose peripheral velocity of the vinyl-glass metglass wire desired flywheel material, it 0.02" x 0.5" was a straightforward calculation to determine outside diameter of Tensile strength 264.000 450,000 368,000 the flywheel. In this case, Intrinsic energy 40.8 25.2 20.4 using the steel wire as the fly- density - Wh/lb Usable energy 17.8 15.8 12.7 wheel wound material, the outside density - Wh/lb diameter was 24 inches, and the Assumed material 50 50 60 cross section dimensions of the cost — c/lb flywheel are simply determined Material performance 35.6 31.6 21.2 from the amount of weight in the Wh/$ wound flywheel necessary to permit Winding weight — 60 67 84 the storage of one kilowatt hour Ibs-for 1 kWh of energy at the maximum rotational Relative volume 2.1 0.67 1 speed. Actually, the total amount Relative winding time 2.1 .02 1 of energy in the flywheel at this condition is approximately 1100 watt hours, in order to permit the Considering first the steel wire storage of one kilowatt hour at a flywheel, the relatively low four to one speed ratio. usable energy density and result- ing high winding weight are
49 relatively inconsequential for the than the steel wire flywheel. stated application. Of particular However, because this material is importance, however, is the fact used in a relatively wide tape that the material performance is form rather than wire form, its somewhat marginal at 21.2 watt relative winding time is approxi- hours per dollar, indicating that mately one fiftieth of the wind- the steel wire flywheel cost may ing time of the steel wire and be somewhat higher than desired. l/100th of the winding time of the vinyl-glass. This could be The relative volume and rela- an important factor in favor of tive winding time are arbitrarily the Metglas flywheel over the •set at unity for the steel wire other types, inasmuch as the flywheel, in order to best compare winding time is a significant this with the other types. The cost factor in the flywheel fabri- relative winding time is a second cation. It should also be pointed order indication of total cost of out that the Metglas can appar- the flywheel. ently be made available in one inch widths, as opposed to the The vinyl glass bare filament one-half inch widths currently flywheel, on the other hand, has being used, which should result about 40% more energy density and in further reductions in fabrica- nearly a 70% advantage in material tion cost. performance cost. However, both of these items are based upon the FLYWHEEL BEARING AND assumption that the vinyl-glass SUSPENSION DEVELOPMENT configuration will permit utiliza- tion of 50% of the ultimate tensile The design of the flywheel strength, compared with 70% in the bearing system and suspension case of the steel wire. While system are inextricably tied this may be a perfectly valid together. In the case of the APL assumption, based on the prelimi- system an attempt has been made nary testing to date, a consider- to minimize the axial and radial able amount of additional testing loads on the bearings, and thereby will be required in order to con- permit the use of much smaller firm this assumption. It can be bearings at a much lower than seen that the vinyl-glass flywheel normal load rating. Also the occupies more than twice the bearings themselves, have received wound volume of the steel wire considerable attention, and flywheel, and it is largely designs have been developed which because of this that the relative appear to offer somewhat lower winding time is also greater than drag but an order of magnitude in the case of the steel wire. longer projected lifetime. These bearings together with the APL The polished Metglas flywheel suspension system are described appears to combine the best in the following. advantages of the other two, with some additional advantages of its BEARING TEST EQUIPMENT own. Usable energy density is somewhat comparable to that of The test rig illustrated in the vinyl-glass flywheel, as is Fig. 5 was used to evaluate the the material performance cost. various bearing configurations But this latter factor is depen- being considered. dent upon a very important assumption that the projected cost of the Metglas material will be $.50 per pound, which is two orders of magnitude lower cost than it is at the present time. Because of its very high usable strength, the Metglas fly- wheel occupies 50% less volume
50 SERIES BEARING CONFIGURATION Figure 7 illustrates the APL series bearing concept which is, in effect, one bearing rota- ting inside of another bearing, with the resulting rotational speed of each being reduced by 50%.
Stationary housing
Identical high speed bearings
Fig. 5 Bearing test equipment. Fig. 7 Series bearing concept.
The losses were determined by respective measurements of bearing The test assembly of this concept torques using the sensitive torque is illustrated in Fig. 8, where it watch device illustrated in Fig. 6. can be seen that the bearings are compactly arranged so as to be axially adjacent rather than radially adjacent.
Fig. 6 Torque watch for measuring bearing.
This device has an apparent accuracy of 1/100th of an ounce Fig. 8 Series bearing test unit. inch. The bearing test equipment is so arranged that tests can be conducted either at atmospheric In addition to providing an order pressure or under vacuum condi- of magnitude increase in projected tions. lifetime by virtue of the greatly reduced rotational speeds, this bearing arrangement also provides improved reliability (since it is unlikely that both bearings would fail simultaneously), and reduced
51 bearing losses. 140 Magnet — Barium ferrite no. 5 The use of the series bearing li20 O.D. = 3.950" configuration in the APL flywheel I.D. = 1.292" energy storage system results in llOO Thickness = 0.425" such a low rotational speed on the Steel shell = 4.683" O.D. bearings, that more exotic schemes I 80- -Nominal design range such as the Draper Retainerless = 80 ± 5 Ib Bearing Concept, are not believed I 60- to be required. 40 2 4 6 8 10 12 14 16 18 MAGNETIC SUPPORT SYSTEM GAP—thousandths of an inch Since the flywheel spin axis Fig. 10 Performance of magnetic support system. in the APL energy storage system is in the vertical plane, it is practical to consider relieving the gravity loads from the fly- The magnetic attractive mode was wheel and motor generator with a selected for the system rather passive magnet system. For than the repulsive mode for a example, if the rotating machinery number of reasons. First, in the weight were 85 pounds, and a attractive mode the magnet is magnetic support system were stable in two planes and unstable utilized having a capability of in only one plane (the axial 8C pounds then the actual load plane). On the other hand in the from gravity on the bearings would repulsive mode the magnet system be only 5 pounds. In other words, is stable in one plane (axial) the load on the bearing would be and unstable in the other two. the same order of magnitude as But perhaps even more important the desired preload on the bearings. is the fact that in the repulsive The overall size of the rotating mode the optimum system would mass is such that a permanent magnet require two magnets; whereas only system could be considered, and one magnet is required in the such a system is illustrated in attractive mode. The effect on Fig. 9. Its magnetic force magnetic system cost is obvious. properties are illustrated in Fig. 10. The Jobmaster Corporation in Randallstown, Md., who designed and built the APL magnetic support system, conducted an in depth study of the advantages and dis- advantages of the rare earth magnets versus the simple ceramic ferrite magnets. These studies indicated conclusively that the ferrite magnet by far offers out- standing performance per unit cost. It has been projected that the magnetic support system illustra- ted in Fig. 9 and having the per- formance approximately the same as Fig. 10 would, in mass produc- tion, cost approximately 75£. Thus, the addition of the magnetic support system to the low cost stationary energy storage system flywheel represented an insigni- ficant increase compared with the Fig. 9 Magnetic support component advantages gained.
52 SUSPENSION SYSTEM DEVELOPMENT one which is being used in the current APL system, is the bonded The basic APL concept for hub arrangement, where the fly- the stationary flywheel energy wheel hub structure is bonded to storage system suspension is the metal hub assembly with a thin illustrated in the Fig. 11 sketch. rubber slab located in between, in order to permit accommodation <.-* differential expansion, and also to "^-Energy absorbing provide damping. Drive elastomer motor A number of schemes have been tested to provide driver damping. One successful driver damping Relatively arrangement is illustrated in stiff shaft Fig. 12, and this is the arrange- ment currently used in the APL Flywheel spin test facility.
Energy absorbing elastomer Fig. 11 Flywheel suspension system concept.
The principal advantage of this flexible shaft suspension system is that the radial loads (or imbalance loads) on the flywheel bearings are reduced to an absolute minimum, and actually are virtually eliminated. A secondary advantage of this arrangement is that the system ae.ommodates a relatively large amount of imbalance without trans- mitting imbalance loads to the Fig. 12 Demonstration system suspension system. bearings. SYSTEM DISCUSSION Here the turbine drive assembly is isolated from the support Two types of damping are structure through a solid rubber employed in the APL flywheel coupling, which allows consider- suspension system. There is hub able motion between the turbine damping of the motion between the and the support structure, while flywheel plane and the spin axis, at the same time providing the and there is driver damping which necessary damping between these damps motion between the overall structures. A functionally sprung system and the support similar arrangement is that em- structure. Hub damping has been ployed in the full-size energy applied in the previous APL sub- storage system as a part of the systems by means of a rubber current program. Here the entire coupling between the flywheel and flywheel container assembly is the flywheel shaft. In the past mounted on suitable shock mounts such a coupling has been in the to provide the same effect as the form of a typical Lord Corporation driver damping. shock mount located at the fly- wheel hub, in order to permit These flywheel suspension gimballing, as well as isolation arrangements coupled with an of the flywheel from the rotating appropriately designed flexible system. A second type, and the shaft having critical speeds far
53 in excess of those expected to be 15,000 rpm. Its design is based encountered, permitted relatively upon an off-the-shelf unit having smooth flywheel operation through- moderately high performance. out the design rotational speed range. The control system receives 115 volt, one phase, 60 hertz ENERGY STORAGE SYSTEM DESIGN input power, rectifies it to a controlled level DC voltage, and Two complete energy storage inverts it to controlled frequency systems will be fabricated in the three phase AC voltage. The con- current program. The first of trol system is arranged to provide these systems will be the so-called constant current per phase, so "battleship model" made with that the motor produces a constant ruggedized components for the torque up to the point where the purpose of evaluating the electri- line voltage reaches 230 VAC at cal and thermal problems in 120 Hertz. At this point the advance of the final system com- motor is producing in excess of ponent availability. The princi- four HP. Over the entire rota- pal components of this energy tional speed range, the average storage system are illustrated in power input and output is of the Fig. 13. order of two kilowatts.
The control system has been designed to accept the energy from a 115 VAC source, transfer this energy into the flywheel, allow storage of the energy in the fly- wheel for an unspecified period, and finally to transfer the energy from the flywheel into a 115 VAC electrical load. REFERENCE 1D. W. Rabenhorst and T. R. Small, "Composite Flywheel Develop- ment Program: Final Report" APL/JHU Report No. SDO-4616A dated April 1977.
Fig. 13 Energy storage demonstration system.
The flywheel in this system is a laminated aluminum disk which, while storing only about half of the final flywheel assembly, nevertheless operates at essen- tially the same rotational speed range as the final flywheel system. The motor generator and control system are essentially the same as those which will be used in the final system, and this equip- ment will be used to evaluate the overall characteristics of this motor generator and control system. The rotating machine is a squirrel cage induction motor designed to operate between 3600 and about
54 PROJECT SUMMARY
Project Title: Materials Program for Fiber Composite Flywheels Principal Investigator: J. A. Rinde Organization: Lawrence Livermore Laboratory University of California P. 0. Box 808, L-338 Livermore, CA 94550 (415-422-7077, FTS 532-7077) Project Goals: The goals of this project are threefold: (1) to accelerate the widespread use of the fiber composite flywheel by developing the necessary engineering design data on fiber composite materials, (2) to demonstrate the high energy kinetics attainable with fi' f* composite materials, and (3) to transfer the technology thus qained to the private sector. Project Status: Our fiber composite materials program for flywheels is divided into the following areas: matrix resins, static engineering properties, stress rupture (lifetime tests at constant load), and dynamic fatigue. During the past year, we characterized a rubberized epoxy resin that offers improved fracture toughness and suitable performance at moderately elevated temperatures (up to 70 C). We also evaluated six epoxy resins for service at 150 C. In addition, the rubberized epoxy resin was used as a matrix in Kevlar 49 composites and engineering design data were generated. We initiated stress rupture tests on E-glass com- posites at load levels of 60 to 85% of short-term strength; these tests will continue for several years. We also began dynamic fatigue tests on Kevlar 49 composites in the tension- tension mode. In addition, fatigue tests on a Kevlar 49 composite rinq specimen at 50 to 75% of ultimate strength are in progress. We anticipate that these data will provide a direct estimate of flywheel performance. Contract Number: W-7405-ENG-48 Contract Period: Continuing Funding Level: $350,000 (includino rotor development) Funding Source: U. S. Department of Energy, Mechanical Energy Storage Division
55 MATERIALS PROGRAM FOR FIBER COMPOSITE FLYWHEELS* J. A. Rinde Lawrence Livermore Laboratory, University of California Livermore, California 94550
ABSTRACT
Our fiber composite materials program for flywheels is divided into the following areas: matrix resins, static engineering properties, stress rupture (lifetime tests at constant load), and dynamic fatigue. During the past year, we characterized a rubberized eposcy resin thtL offers improved fracture toughness and suitable perfor- mance at moderately elevated temperatures (up to 70°C). We also evaluated six epoxy resins for service at 150°C. In addition, the rubberized epoxy resin was used as a matrix in Revlar 49 composites and engineering design data were generated. We initi- ated stress rupture tests on E-glass composites at load levels of 60 to 85% of short- term strength; these tests will continue for several years. We also began dynamic fatigue tests on Kevlar 49 composites in the tension-tension mode. In addition, fa- tigue tests on a Kevlar 49 composite ring specimen at 50 to 75% ultimate strength are in progress. We anticipate that these data will provide a direct estimate of flywheel performance.
DJTRODOCTION Revlar 49 and E-glass composites. During the past year, we concentrated our The fiber composites materials pro- efforts specifically on new matrix res- gram at LLL was begun in 1975 with the ins, E-glass composite stress rupture, primary goal of providing meaningful and engineering properties of a composite reliable engineering design data on com- made with a rubberized resin, and com- posite materials specifically intended posite fatigue lifetime. Each of these for flywheel rotors. The two main crit- areas is discussed below. eria for material selection for this fly- wheel program are high performance and MATRIX RESINS* low to reasonable cost. Because the high-performance composites used by the We studied two types of epoxy re- aerospace industry are very costly, we sins this year, a rubberized resin with could not consider these materials, even improved fracture toughness, and high though thorough design data exist. temperature resins for service at 150°C. Complete details of our work are given in Therefore, since 1975, we have: Refs. 1 and 2. • Generated engineering design data for Kevlar 49, S2-glass, and E-glass com- RUBBERIZED EPOXY 3ESIN posi tes. • Spin tested thin-rim flywheel rotors of A rubberized epoxy resin is a nor- Kevlar and glass fiber composites and mal epoxy resin that has been modified by correlated the results with the fibers' the incorporation of a soluble carboxy- static strength properties. terminated butadiene acrylonitrile (CTBN) • Developed and engineered new fiber com- rubber. During the curing process, the posite flywheel design—a quasi-isotro- CTBN rubber is forced out of solution to pic, laminated, solid disk rotor with a form a second phase of l-to-10-um dia- tapered profile. meter rubber particles. These particles • Characterized flexible and rubberized modify the fracture process of the resin matrix resins and provide improved fracture toughness. • Investigated the transverse tensile In our work we formulated a resin system properties as a function of matrix resin Six table for wet filament winding, con- modulus for a series of S2-glass compos- ducted fracture tests to demonstrate the ites. • Conducted stress rupture tc~ts on •Principal Investigator, J. A. Rinde. *This work was performed under bhe auspices of the 0. S. Department of Energy by Lawrence Livermore Laboratory under contract No. W-7405-Eng-48.
56 resin's improved fracture toughness and the resin systems on five criteria: (1) then characterized the mechanical and low viscosity (1.0 Pas at 25°C), (2) physical properties of the resin. long gel time (220 h for a 30-g mass at 25°C), (3) high glass transition tem- Table 1 summarized the properties of perature (Tg 2 180°C), (4) high ten- this resin system. The epoxy resin sile strength with nigh modulus, and (5) XD 7575.03 (Dow Chemical) contains 10% good retention of mechanical properties CTBN rubber. This rubberized resin sys- upon accelerated aging (7 days at 175°C). tem offers the advantages of low viscos- Two of the resin formulations (1 and 2) ity and long gel time for easy processing have tensile strengths above 85 MPa and as well as a moderately high tensile glass transition temperatures above 205°C strength of 76 MPa. However, since we in both and as-cured and aged conditions. completed this work, both the XD 7575.03 The resin systems formulated and some of and XD 7714 (Dow Chemical) have been their key characteristics are presented in Table 2. withdrawn from the commercial market. An equivalent product for the XD 7575.03 can be obtained by using 23% Kelpoxy All resin systems were cured for 3 h G-293 (Spencer Kellog Division of at 70°C plus 2 h at 120°C plus 2 h at Textron) plus 77% DER 332; XD 7114 can 180°C to achieve a high Tg and to cure be replaced by Wilmington Chemicals the resin at a temperature above the in- Helox 68. tended-use temperature of 150°C. With the exception of the resin systems cured RESINS FOR HIGH-TEMPERATURE SERVICE with APCO 2347 (modified imidazol), these cure conditions produced a Tg greater We have formulated and evaluated six than 185°C. Both the APCO 2347 systems epoxy resin systems suitable for wet fil- (resins 5 and 6) exhibited additional ament winding of high-performance fly- curing upon heat aging. wheels. We determined their processing characteristics as well as their mechani- The six resin formulations were sub- cal properties. Specifically, we judged jected to two mechanical-properties tests
Table 1. Properties of a rubberized epoxy resin system. Components Parts by weignt Resin: Dow XD 7575.03 100 Diluent: Dow XD 7114 65 Curing agent: Tonox 60-40 33.9 Cure cycle,h/°C 1.5/90 Tensile properties: + 2/130 Viscosity at 25°C, 0.95 Stress at maximum, MPa 76.1 Pa*s Gel time for 30-g mass 29.3 Stress at failure, MPa 72.5 at 25°C, h Cured density at 25°C, 1.19 Strain at maximum, % 5.7 Mg/m3 Water absorption (ASTM Strain at failure, % 8.4 D-570-63), % gain/h 2.02/8, Modulus, GPa 2.43 3.13/24 Glass transition temperature, Compressive properties: °C 104 , Thermal coefficient of linear Maximum strength, MPa 88.7 expansion from 298 to 377 K, in./in./°C 71 x 10"6 Strain at maximum strength, % 5.7 IZOD impact (ASTM D-256-73), Secant modulus, GPa 2.9 J/m: Method A 47.7 Torsional properties: Method E 645.4 Maximum shear stress, MPa 54.9 Flexural Strength at 5% strain, Strain at maximum stress, % 16.4 MPa 106.4 Tangent modulus, GPa 2.85 Tangent modulus, GPa 1.59
57 Table 2. Performance of the six epoxy resin systems formulated for elevated- temperature service.
Resin Resin Gel Tensile Properties system components Viscosity, time, Modulus, Stress, Strain, No. (parts by weight) Pa-s h GPa MPa % Tg,°C
Ciba 0510/RD-2/APCO 2330 1.05 21.1 (100/20/43.1): Cureda 5.4 89.5 2.0 205 Agedb 4.8 84.7 2.0 220 Ciba 0510/Tonox 60-40 1.33 25.1 (100/49.8): Cured 4.1 86.8 3.2 210 Aged 4.1 73.4 2.2 226 APCO 2447/ERL 4206/ Tonox 60-40 (100/20/ 1.08 38.6 39.9): Cured 3.9 83.5 2.4 190 Aged 3.9 71.4 2.0 215 Ciba 0510/DEN 438/RD-2/ Tonox 60-40 (75/25/25/ 1.48 19.6 49.9): Cured 4.0 81.0 2.5 185 Aged 4.2 78.5 2.8 213 Ciba 0510/DEN 438/RD-2/ APCO 2347 (75/25/20/ 0.88 34.4 12.2): Cured 3.1 74.8 3.7 115 Aged 3.4 73.2 2.8 170 DER 332/RD-2/APCO 2347 0.90 48.5 (100/15/9.5): Cured 2.8 59.8 2.7 115 Aged 2.6 53.7 2.5 145
a3 h at 70°C plus 2 h at 120°C plus 2 h at 180°C. b7 d at 175°C.
and to a thermal gravimetric analysis in both the as-cured and aged conditions. Resin samples were aged in a constant- temperature, forced-air oven at 175°C for 7 days. In general, aging caused the samples to darken and become brit- tle. The results of these tests are summarized in Table 2. 5
The tests reveal that resin system 1, cured with APCO 2330, meets our process sing requirements of low viscosity and long gel time, and has the highest ten- 300 sile strength and modulus, the second smallest loss in tensile strength upon Temperature, °C heat aging, and the second highest Tg Pig. 1. Dynamic shear modulus for the in both the as-cured and aged conditions. six epoxy resin systems formulated Resin system 3 also meets the processing for high-temperature service (tested requirements, but has a lower Tg, a at 0.1 Hz). All resin were cured lower tensile strength, and a larger for 3 h at 70°C plus 2 h at 120°C reduction in strength upon heat aging plus 2 h at 180°C. Resin 1 is than resin 1. Ciba 0510/RD-2/APCO 2330, resin 2 is Ciba 0510/Tonox 60-40, resin 3 We conducted dynamic shear modulus is APCO 2447/ERL 4206/Tonox 60-40, measurements as a function of temper- resin 4 is Ciba 0510/OEH 438/RD-2/ ature; the results for all six resin Tonox 60-40, resin 5 is Ciba 0510/ systems in the as-cured condition are DEN 438/RD-2/APCO 2347, and resin 6 shown in Fig. 1. We used these measure- is DER 332/RD-2/APCO 2347.
58 100 1 flywheel operating under various stresses 1 • i i ii 150°C_ associated with the input, storage, and 98 - T55= output of energy. We use. the results of a? 96 — — stress rupture tests at constant load 175°C -* 200°C ~ levels as the baseline benchmarks. Such M tests are required because even a nominal .i variation (typically less than 5%) in 3 92 - 225°C - static strength can lead to a large 250°C scatter in stress rupture life, often in 90 ,275°C excess of 100% (see Fig. 3). Therefore, 1 1 • 1 i 88 to provide the necessary statistical 20 40 60 80 100 parameters for the reliable design of Time, min composite components, large data samples from long-term testing are being accumu- lated in testing facilities designed for Fig. 2. Isothermal weight loss at simultaneous testing of dozens of various temperatures in air by ther- samples. Figure 4 illustrates the type mogravimetric analysis for resin of data being generated. From such system 1, Ciba 0510/RD-2 APCO 2330. curves, we can determine the amount of stress-level derating that is required to obtain the desired degree of reli- ability in the component's operating
ments to determine the Tg of the cured lifetime. resins as well as to show the variation of modulus with temperature and to pro- vide a measure of cross-link density. We are conducting the first stress From these results, we conclude that re- rupture tests on E-glass/epoxy; E-glass sins 1, 2, 3, and 4 would perform well at is a low-cost, large-bundle fiber of in- 150°C in short-term application. terest for flywheel applications. For these tests, the glass yarn was inpreg- The effect of heat aging these re- nated with resin (Dow DER 332/Jeffamine sins is exhibited by the change in Tg T-403) on a filament-winder and the re- and tensile properties (see Table 2). In sultant composite strands were given a all cases, the Tg increases with aging mild heat cure. Strands were cut to but the tensile strength and failure length, fixed with end tabs, and loaded strain decrease with aging. in the test apparatus at 60 to 85% of their average ultimate failure strength. In Fig. 4, the stress rupture results are Figure 2 presents the isothermal plotted as precent load vs log time. weight loss curves for resin system 1 at Curves for Kevlar 49/epoxy and S-glass/ temperatures from 150 to 275°C as epoxy composites also are included for determined by thermogravimetric analy- comparison. sis. At 150°C, resin 1 is thermally stable after an initial weight loss of about 2%. Above about 200°C, the sample exhibits an accelerating rate of decomposition as a function of the in- 1 . • . i crease in temperature. Therefore, we •at 550 rr expect that flywheels using this resin cv< system would have a lifetime of several 500 years at 150°C and be able to with- stand shorter exposures (~30 min) to 450 temperatures up to 200°C (Tg = 205°C). y cv >iro%\ STRESS RUPTURE* 400 1 1 1 1 Xl. 0 100 500 1000 We are also characterizing the time- dependent strength of several candidate Time to failure, h fiber composites being considered for flywheels. These data are needed to pre- Fig. 3. Nominal scatter in static dict the probability of failure for a strength data that can result in large scatter in lifetime predic- •Principal investigator, L. Penn. tions .
59 1h
I ' I I I I I I 'I *>l. I I I I I
50 i il l i 11 i i i 11 i i i 11 i i i 11 i \ i i i 10 -3 10- 10r1 10° 101 102 103 1O4 105, Lifetime, h Pig. 4 Stress-rupture lifetime data for several composite materials. The 2-to-100% bands for S-glass and Kevlar 49 composites are displayed; data points are for E-glass composites.
The most noticeable feature in Composites specimens 60-to-70-vol% Fig. 4 is the much broader distribution Kevlar 49 fiber (1420 denier without of break times in the E-glass composite sizing) were filament wound. The resin as compared to the S-glass and the system was XD 7575.03/XD7114/Tonox 60-40 Kevlar 49 composites. This was expected (100/65/33.9). Composites were cured for because E-glass is a low cost fiber and 4.5 h at 60°C plus 3 h at 120°C. The is known to have a higher variability fabrication and testing methods are re- than the more expensive, aerospace grade ported elsehwere.3'4 S-glass fiber. Some of the variability in E-glass results may have been caused Elastic constants and ultimate by an inaccurate load applied to the strength properties of composites of 60-, sample. Such inaccurate application of 65-, and 70-vol% fiber are presented in load is possible because of the lever Table 3. The stress-strain curves of the arm arrangement of the apparatus. We 65-vol% composite are shown in Fig. 5. will perform some dead-weight load tests The performance of this composite is com- to verify or disprove this hypothesis. pared to two other Kevlar 49 composites Because so few data are available at the using different resin systems (DER 332/ lower load levels at this time, we can Jeffamine T-403, and XD 7818/Jeffamine draw no conclusions from these data con- T-403) in Table 4. This comparison re- cerning the useful stress levels for veals that the rubberized epoxy yields E-glass composite flywheels. lower shear properties than the other two matrices; other mechanical proper- ENGINEERING PROPERTIES* ties, however, are comparable.
We have also conducted an extensive FATIGUE LIFETIME* evaluation of the physical and mechani- cal properties of a Kevlar 49 composite Knowledge of the time-dependent de- using the rubberized epoxy resin system formation and strength properties is described above. The primary advantage essential for the design of flywheels to of this composite over previously tested ensure minimum dimensional change, maxi- Kevlar composites is the higher Tg of mum dynamic stability and long-term the rubberized matrix resin. safety of operation. Among the candidate
•Principal investigator, L. C. Clements. *Principal investigator, E. M. Wu.
60 Table 3. Mechanical properties data for a composite of Kevlar 49 in a rubberized epoxy resin, XD 7575.03/XD 7114/Tonox 60-40. Fiber content, vol% Property 60 (CV8) Hb 65 (CV) N 70 (CV) N Elastic Constants0 Longitudinal young's modulus, GPa 80.52d (8.2) 22 Transverse Young's modulus, GPa 7.82 (15.5) 5 Shear modulus, GPa 1.7 (2.0) 8 Major Poisson's ratio 0.367 (3.96) 22 Minor Poisson's ratio 0.037 (7.33) 5
Ultimates Longitudinal tension: stress, MPa 1685 (5.9) 5 1814 (4.1) 6 1920 (3.0) 4 strain, % 2.6 (5.2) 5 2.6 (3.5) 6 3.3 (2.0) 4 Longitudinal compression: stress, MPa 220 (5.8) 6 strain, % 0.35 (7.2) 6 Transverse tension: stress, MPa 8.7 (8.8) 5 6.4 (10.5) 5 4.8 (10.6) 5 strain, % 0.17 (11.4) 5 Transverse compression: stress, MPa 40.3 (9.1) 5 strain, % 1.4 (15.4) 5 Shear at 0.2% offset: stress, MPa 19.2 (2.8) 7 16.2 (3.8) 8 16.2 (4.3) 8 strain, % 1.4 (3.7) 7 1.2 (4.7) 8 1.2 (7.9) 8 Shear at failure: stress, MPa 26.8 (2.4) 7 22.55 (1.2) 8 21.36 (1.5) 8 strain, % 2.7 (12.3) 7 2.0 (3.6) 8 2.0 (8.9) 8 ^V coefficient of variation, in percent. TJumber of specimens tested. cElastic constants are assumed to be valid for both tension and compresion. Estimated from tests at 60 to 70 vol% fiber, normalized to 65 vol%.
2000 flywheel composite materials, only com- posites of Kevlar 49 exhibit time-depen- dent deformation in both the fiber-con- trolled and the matrix-controlled proper- 250 ties. Therefore, Kevlar 49 composites 1600 are being studied extensively because of their apparent high engineering poten- 200 tials. Longitudinaij We are characterizing the time-de- 1200 tension pendent properties of Kevlar 49 com- posites in the fiber direction by con- 150 ventional operations using linear visco- elasticity. Tests being conducted are 800 illustrated in Fig. 6. Creep compli- '65% fiber volume ances are obtained from static-weight 100 tests. The onset cf nonlinearity is dependent on both stress level and load- ing rate and is being identified by com- 400- paring creep tests and ranp loading tests 50 at several loading rates. We are using - j»Longitudinal "ompression low-frequency fatigue tests (1000 s/ ^Transverse tension _. cycle) to identify the effects of cyclic 5tf i i- snear stress history on the composite deforma- 0 12 3 tions and lifetimes. By comparing the fatigue data to our stress rupture.data, Strain, % we will obtain the design curves used to Fig. 5. Mechanical properties of estimate composite lifetimes in accor- 65-vol% Kevlar 49/epoxy composite dance with current engineering practice in tension, compression, and shear: (e.g., S-N or strain vs number of cycles rubberized epoxy was XD 7575.03/XD to failure curves). We are also explor- 7114/Tonox 60-40. ing a theoretical correlation via the
61 Table 4. Mechanical properties of 60 vol% Kevlar 49 composites in three resin systems, two rigid matrices and the rubberized matrix.a
Resin matrix Property Der 332/T-403 XD 7818/T-403 Rubberized
Elastic Constants'3 Longitudinal Young's modulus, GPa 81.8 75.1 80.5 Transverse Young's modulus, GPa 5.10 4.56 7.8 Shear modulus, GPa 1.82 1.89 1.7
Ultimates Longitudinal tension: stress, MPa 1850 1400 1685 strain, % 2.23 1.7 2.6 Transverse tension: stress, MPa 7.9 12.4 8.7 strain, % 0.16 0.28 0.17 Shear at 0.2% offset: stress, MPa 24.4 33.67 19.2 strain, % 1.55 1.98 1.4
aRigid matrices: DER 332/T-403 (Dow Chemical/Jefferson Chemical), XD 7818/T-403 (Dow Chemical/Jefferson Chemical). Rubberized matrix: XD 7575.03/XD 7114/Tonox 60-40 (Dow Chemical/Dow Chemical/Oniroyal). ^Elastic constants are assumed to be valid for both tension and compression.
damage function formulation in an attempt this cyclic test simulates the ring-type to increase the utility of these fatigue flywheel and that the test results will data in generalized applications. be useful as a direct estimate of fly- wheel structural performance. A second part of our fatigue program is to test small Kevlar 49 rings by load cycling in the tension-tension mode at In conventional engineering prac- the stress levels expected in operating tice, the time-dependent strength of a fiber composite flywheels (i.e., 50 to composite is characterized in terms of 75% of ultimate stress). We believe that the input loads {e.g., the S-N curves and the stress rupture curves). In our work, however, we are relating materials response properties to the input loads. In this manner, we are not only producing immediately usable engineering data, we are also laying the groundwork for a generalized failure theory for fiber com- Creep and creep rupture posites. Much of our effort and re- «• ••• ^M w» m^^ «HW M^^B av sources have been directed toward the 1 • • • necessary instrumentation and data aqui- sition systems to measure the strain re- •* .• Ramp loading sponses of the composite. Examples of CO the creep and fatigue responses are shown in Pig. 7. The experimental results ac- / / ..•** Fatigue cumulated to date are shown in Fig. 8: the open data points are from creep rupture tests and the solid points are from fatigue tests. In Fig. 8a, we pre- sent the failure points in the conven- Time —»• tional format of stress level (i.e., in- put load) vs the time to failure {i.e., Fig. 6. Mechanical test performed to number of load cycles). In Fig. 8b, we determine the time-dependent be- present the failure points in terms of havior of Kevlar 49 composites in the materials response (i.e., the failure the tension-tension mode. strain) vs the time to failure.
62 250 U.Ul • ' • ' • ' ' (b)
0.008 -
Strain , % - P
0.002 -
n i i 0 15 30 45 60 2.4 4.8 7.2 9.6 12 Strain response to creep, h Strain response to fatigue, h Fig. 7. Strain response to creep (a) and to fatigue (b) of Kevlar 49/ epoxy strands. Fatigue testing was conducted with square wave cycling.
10"1 10 102 103 104 90 1 1 • > • 1 i W'20% ' 1 30'% Failed' (a)
a 85 - ^^ ^-A ^==:A:::^^5«!!
80 1 1 1 11 . i i • i i I 1 1 1 1 i 1 i i , . 1 , i i i 840I ' 1 r 1 I I i £ 820
•S 800
780
(m a 760
740
n t o fail i g } Creep
Stra i 720 • Cyclic fatigue
70a ± ± • • .1 • • • 1 i10 10" 1 10 10£ 103 104 m-2 Time to break, h Fig. 8. Creep and fatigue of Kevlar 49/epoxy strands tested at 25°C: (a) stress (i.e., input load), and (b) strain (i.e., materials response).
63 Plots of stress vs time to failura REPORTS, PUBLICATIONS, ADO PRESENTATIONS (Fig. 8a) ace useful as design charts with which to estimate the lifetime re- 1. C. C. Chiao and T. T. Chiao, Aramid duction of the composite component due to Fibers and Composites, Lawrence creep and fatigue loads. On the other Livermore Laboratory, Rept. hand, plots of fatigue strain (Fig. 8b) UCRL-80400 (December 1977); to be a suggest the generalized observation that chapter in 27K Handbook on Fiber- composite strain compliance at failure glass and Plastic Composites. is weakly dependent on stress level and strongly dependent on load history. 2. T. T. Chiao, J. H. Rinde, and E. T. Mones, Epoxy Resins for Fiber Com- FUTURE WORK posite Flywheel Rotors, Lawrence Livermore Laboratory, Rept. We are continuing our stress rupture OCRL-79573 (October 1977), presented and dynamic fatigue programs to obtain at the 1977 Flywheel Technology more baseline data for predicting the Symposium, San Francisco, CA. lifetimes of fiber composite flywheels. Kevlar 49 and E-glass composites are now 3. T. T. Chiao, Material Properties being tested. New stress rupture and of Composite Flywheels, Lawrence cyclic fatigue tests also will be con- Livermore Laboratory, Rept. ducted at elevated temperatures and under UCRL-80515 (April 1978), presented vacuum. Some of this work will be done at The Middle Atlantic Regional under contract at the Oak Ridge National Meeting of the American chemical Laboratory. We plan to continue char- Society, Hunt Valley, MD. acterizing the matrix resin formulated for service at 150°C. We also will in- 4. T. T. Chiao, Some Interesting vestigate resins with improved bonding to Mechanical Behavior of Composite Kevlar 49 in an attempt to increase the Materials, Lawrence Livermore transverse tensile and shear properties Laboratory, Rept. UCRL-80908 (April of the composite. Engineering properties 1978), presented at the VS-VSSR will be determined on S2-glass/high-temp- Seminar on Fracture of Composite erature-service resin composites at room Materials, Riga, USSR. temperature and at elevated temperatures. Some work may also be done on low-cost 5. R. M. Christensen, J. A. Rinde, and graphite fiber/epoxy composites. E. T. Mones, Transverse Tensile Characteristics of Fiber Composites Using Flexible Resins, Lawrence Livermore Laboratory, Rept. REFERENCES UCRL-80241 (October 1977), presented at the 1977 Flywheel 1. J. A. Rinde, E. T. Mones, R. L. Technology Symposium, Moore, and H. A. Newey, An Epoxy San Francisco, CA and accepted for Resin-Elastomer System for Filament publication in J. Polymer Eng. Sci. Winding, Lawrence Livermore Laboratory, Rept. UCRL-81245 (1978) 6. L. L. Clements, Fiber Composites Flywheel Program - Filament Wound 2. J. A. Rinde, E. T. Mones, and H. A. Composite Data Sheets, Lawrence Newey, Filament Winding Epoxy Resins for Livermore Laboratory, Rept. Elevated Temperature Service, Lawrence UCID-17874 (August 1978). Livermore Laboratory, Rept. UCRL-52577 (1978). 7. L. L. Clements, "Problems in Testing iVramid/Epoxy Composites," Lawrence Livermore Laboratory, Rept. 3. L. L. Clements and T. T. Chiao, UCRL-79450 (November 1977), presen- "Engineering Design Data for an Organic ted at the AIMS Failure Modes in Fiber/Epoxy Composite," Composites 8, Composites IV, Chicago, IL. 87-92 (1977). 8. L. L. Clements and R. L. Moore, 4. L. L. Clements, R. L. Moore, and T. Comparative Engineering Properties T. Chiao, "Elongated-Ring Speciman for of Fiber Composites for Flywheels, Tensile Properties of Filament-Wound Lawrence Livennore laboratory, Rept. Composites," in Materials Review '75, UCRL-79575 (October 1977), presented Proc. 7th Natl. SAMPE Tech. Conf., at the 1977 Flywheel Technology Azusa, CA, 1975, p. 188. Symposium, San Francisco, CA.
64 9. L. L. Clements and R. L. Moore, 13. J. A. Rinde, B. T. Hones, and H. A. Composite Properties for S2-Glass Newey, Filament Winding Bpoxy in a Room-Temperature Curable Bpoxy Resins for Elevated Temperature Matrix, Lawrence Livermore Service, Lawrence Livermore Laboratory, Rept. UCRL-8I517 Laboratory, Rept. UCRL-52577 (August 1978), to be published in (October 1978). SAMPB Quart.
10. S. Kulkarni, Interlaminar and Stacking Sequence Considerations for Composite Flywheels, Lawrence 14. J. A. Rinde, E. T. Hones, R. L. Livermore Laboratory, Rept. Moore, and H. A. Newey, An Bpoxy UCRL-13888 (July 1978). Resin-Elastomer System for Filament Winding, Lawrence Livermore Labora- 11. L. S. Penn and T. T. Chiao, Bpoxy tory, Rept. UCRL-81245 (September Resins, Lawrence Livermore 1978), prepared for presentation at Laboratory, Rept. UCRL-79815 the 34th Annual Conference of SPI (October 1977), to be a chapter in Reinforced Plastics/Composites In- The Handbook on Fiberglass and stitute, New Orleans, LA (January 1979). Plastic Composites.
12. L. S. Penn, Physiochemical Characterization of Composites and Quality Control of Raw Materials, 15. E. M. Wu, Failure Analysis of Lawrence Livermore Laboratory, Rept. Composites with Stress Gradients, UCRL-81081 (Hay 1978), presented at Lawrence Livermore Laboratory, Rept. the 1978 ASTM 5th Conference on UCRL-80909 (August 1978), presented Composite Materials, New Orleans, at the US-USSR Seminar on Fracture LA. of Composite Materials, Riga, USSR.
NOTICE "This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Department of Energy, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or respon- sibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately-owned rights."
PLL 65 PROJECT SUMMARY
Project Tftle: The Laminated Dfsk Flywheel Program Principal Investigator: R. G. Stone Organization: Lawrence Livermore Laboratory P. 0. Box 808, L-123 Livermore, CA 94550 Telephone: (415) 422-8284 Project Goals: (1) Develop the technology of laminated disk flywheels. (2) Demonstrate a prototype of a reliable, economical, high-energy density flywheel based on the developed technology. (3) Evaluate and disseminate this technology. Project Status: This program was started very recently, July 1978. Initial analyses have been performed and are continuing. Manufac- turing development has been started. One flywheel has been built and tested.
Contract Number: W-7405-Eng-48 Contract Period: July 1978 - July 1980
Funding Level: $775,000 Funding Source: Department of Energy
S
67 THE LAMINATED DISK FLYWHEEL PROGRAM A ROTOR DEVELOPMENT PROJECT BY LLL AND G.E. CO.*
Richard G. Stone Lawrence Livermore Laboratory P. 0. Box 808, L-123 Livermore, CA ol»550
ABSTRACT
The Lawrence Livermore Laboratory and the General Electric Company** have initiated a joint program to develop the technology of fiber-composite, laminated disk flywheels for energy storage applications. The 2-year program was started in the simmer of 1978. LLL and G.E. will participate about equally in applying their complementary capabilities to the overall program. LLL is developing analytical methods applied to contoured disk wheels. LLL will also investigate bonded hub attachment methods and will select materials and define processing requirements. G.E. is developing an alpha cross-ply laminate disk, a rim overwrap, and a mechanical hub attachment. G.E. will also develop manufacturing processes and will test the developmental and prototype flywheels. Concepts originated in prior efforts and analyses are presented. Current efforts include the testing of a contoured, laminated disk flywheel.
INTRODUCTION recently started, this paper will sum- marize the concepts and analyses leading A number of studies have concluded to the program, describe the planned that energy storage flywheels have the program, and report on the program work capability to conserve energy in a variety accomplished thus far. of applications. Some of these studies have concluded that incorporating a fly- BACKGROUND wheel in a battery-powered vehicle will not only conserve energy but will also For several years the Lawrence provide the acceleration and hill—climbing Livermore Laboratory and the General performance demanded by users. Fiber- Electric Company have been involved in composite materials are most attractive programs related to the development of for energy storage flywheel construction, fiber-composite flywheels. LLL has been having the potential of high-energy developing fiber-composite materials data density, relative safety, and economy. and processing methods applicable to fly- However, the potentially high-energy wheels. We have also been engaged in the densities have not been realized thus far project to develop and evaluate mechanical in efforts to develop fiber-composite energy storage subsystem technology for flywheels. In particular, little atten- application to electric and hybrid ve- tion has been given to the development hicles, technical support to the Electric of laminated disk flywheels. and Hybrid Vehicle Demonstration project, and analysis of the role of energy storage The Lawrence Livermore Laboratory power systems in transportation. In the and the General Electric Company have course of this work, Christensen and Wu initiated a joint program to develop the performed a design analysis of fiber- technology of fiber-composite, laminated composite flywheels1 and Toland surveyed disk flywheels for energy storage applica- recent developments in the application of tion. Since this 2-year program was fiber-composite materials to flywheels.2
*This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore Laboratory under contract No. W-7^05-Eng-l(8. **Reference to a company or product name does not imply approval or recommendation of the product by the University of California or the Department of Energy to the exclusion of others that may be suitable.
68 G.E. has been active in flywheel density. A two-dimensional finite-element energy storage projects including develop- code that uses uniform in-plane properties ment of inductor/motor/alternator/flywheel and uniform, but different, axial proper- energy storage systems for automotive ties has been applied to our first model applications and development of various flywheel designs. Work is underway on fiber-composite flywheel concepts under more detailed analyses of critical areas G.E. company funding. An attractive con- taking into account the anisotropic mate- cept reported by Lustenader and Zorzi1* is rial properties and the combined inter- the "alpha cross-ply" pseudo-isotopic, laminar and continuum stress fields. laminated disk flywheel. G.E. has developed a two-dimensional As a consequence of these studies, orthotropic finite-element simulation of the two organizations independently the alpha-ply flywht-sls. They are currently arrived at the conclusion that a most analyzing the effect of hoop-wound over- promising rotor design is a disk type wrap on stress concentration at hub attach- composed of a number of laminas of fiber ment holes. They are studying overwrap composite materials oriented to maximize materials of higher modulus, such as strength. Disks of right-circular and Kevlar 4 9 and graphite, with various de- of tapered (contoured) cross sections grees of interference fit. are of interest. LLL and G.E. will compare and assess PROJECT DESCRIPTION their analytical models and will apply them cooperatively. GOALS 2. Manufacture and Test Model Flywheels. The objective of this 2-year program In this task about 30 model flywheels will is to develop the technology for high- be designed, built, tested, and evaluated. energy density, fiber-composite flywheels Sizes will range from 10 to 30 in. in based on the laminated disk concept and diam. by 1/U- to 1-in. thick. These tests to demonstrate a prototype, practical will verify the analytical models, the flywheel of this design. Our goal is success of manufacturing methods, and the a working capacity in prototype wheels effects of hub attachment perturbations. exceeding 50 Wh/kg of total storage capacity in the 1- to 2- Kwh range, al- All of the flywheel tests will be though we expect to demonstrate model conducted at the G.E. Corporate Research wheels failing at energy densities well and Development Center, Schenectady, New above 90 Wh/kg. Our goal with respect to York. the practicality of the prototype wheels is a design directed to optimize the LLL will produce about 10 model fly- overall flywheel energy system with wheels for testing. The first test wheels respect to low volume, low weight, manu- will be directed toward verifying the facturability and economy. analytical model. Subsequent wheels will be designed to evaluate anisotropy effects, TASKS optimized contouring, and manufacturing options. The program is planned in five tasks: (l) establish analytical models; (2) manu- G.E. will produce about 20 model facture and test model flywheels; (3) de- flywheels for testing. One group of velop hub attachment; {k) design, manufac- wheels will evaluate the effect of various ture, and demonstrate performance of alpha-ply angles. Another group will be prototype flywheels; and (5) assessment and used to study the effectiveness of outer recommendations for the developed tech- wrap rings of various radii and inter- nology. Tasks 1, 2, and 3 are closely ference fits. Hub attachment methods interrelated and thus will be conducted will be evaluated in a third group. concurrently. Finally, preferred alpha-ply, outer wrap contouring and hub attachment combinations 1. Analytical Models. We will use ana- will be tested and evaluated. lytical models capable of predicting stress and energy density of laminated Manufacturing development is a flywheels to make predictions of flywheel necessary and important part of this performance. LLL is concentrating on the task. G.E. will investigate four areas effects of contouring on stress and energy of manufacturing: (l) the lay-up of
69 fiber prepregs into composite slabs, 5. Assessment and Recommendations. This (2) outer ring winding and press fitting, task concludes the program with an evalua- (3) contoured wheel production, and tion of the technology developed and (It) attachment fabrication. During manu- recommendations for design and manufacture facturing development, G.E. will do the of these high-performance wheels, for following: applications of this flywheel technology to Department of Energy programs and for • Make flat slabs about 12-in. diam. and further flywheel development. 1-in. thick with the objective of assuring slab uniformity, low void content, and ORGANIZATION AND MANAGEMENT fiber parallelism. • Take micrographs to assure low void As stated previously, this will be content. a joint LLL/G.E. program under the co- • Evaluate resin characteristics and ordinating leadership of LLL. The analysis post cure heat treatments. and design efforts will be about equally- • Apply manufacturing processes to the divided between the two organizations. outer ring winding. Although concentrating on their specific • Study winding directly onto wheels problems, the approaches and results will and winding into rings for subsequent be shared by LLL and G.E. for maximum press-fitting onto the wheels. advancement. LuL will concentrate its • Develop contoured wheel manufacturing efforts on contoured wheels, bonded-hub methods. attachment, and materials selection and • Examine the effects of machining the processing. The G.E. effort will be contoured wheel. directed toward laminate disk manufacturing, rim overwrap development, mechanical hub 3. Develop Hub Attachment. The Lord attachment, and component and system de- Corporation and David Rabenhorst of the sign. G.E. will perform spin testing and Applied Physics Laboratory, Johns Hopkins will be responsible for manufacture of University, have developed an adhesive- the prototype flywheels. bonded elastomeric hub attachment method which has been used successfully in The program will be coordinated by testing laminated disk flywheels.5 LLL quarterly review and planning meetings will use this method for their early and ad hoc technical meetings. LLL will test wheels. This design, as well as be responsible for reports and program other potential bonded-hub designs, will reviews. be investigated for application to working flywheels. ACCOMPLISHMENTS
The G.E. overwrap concept alleviates, This program was approved and funded to some extent, the stress concentration by the Department of Energy in June 1978. at holes drilled in the wheel for hub The contract with General Electric Com- attachment. Other methods of mechanical pany was executed in August 1978. Thus, hub attachment designed to minimize this this 2-year program has just gotten under- stress concentration will be investigated. way. Installing slender axial rods without cutting fibers during lay-up of the disk G.E. had developed and analyzed the is one possible approach. alpha cross-ply concept previously and had conducted a few verification tests. h. Prototype Flywheels. We will apply Since August they have continued analyt- the information gained in the previous ical work, started developing manufacturing tasks to design, build, and test proto- methods and building model alpha cross-ply type, practical flywheels. The design flywheels for spin testing. will be directed to optimize the overall flywheel energy storage system with LLL had developed and analyzed their emphasis on the goals of low weight, low quasi-isotropic, contoured concept; ob- volume, reliability, manufaeturability, tained material; and was building a test and economy. We expect that we will wheel under their materials and processing arrive at two designs of 1- to 2- kWh program. Under this program, this wheel capacity an*3. will build and thoroughly was completed and tested; and the results test two wheels of each design. and material problems are under investigation.
70 GENERAL ELECTRIC CO. PROGRESS four samples taken from a 1-in. thick, 10-in. diam. disk made from S-2 glass The G.E. alpha-ply concept defines uniply-prepreg. Two alpha-9 wheels have the orientation of the fibers in adjacent been produced for spin testing. Bonded lamina as hubs, applied by Lord Corporation, will be used in testing these wheels. Angle between lamina = 90 .921 N Table 1. Test data of a laminated wheel where H is the alpha number. Thus, an by hydroclaving. alpha-3 lay-up has the fibers of adjacent lamina displaced 60° from each other, an Location alpha-9 is at 80°, etc. Their analysis Property 12 3 6 £ shows improved energy potential for alpha values above J as shown in Fig. 1. % Glass 65.56 65.50 65.1*1 65.5>t Very preliminary tests appeared to verify this trend. Measured 1.8001 1.8023 I.8O33 1.8017 density
2500 % Voids 2.062 1.9*123 1.781 1.974
With Kevlar wrap LLL PROGRESS
The LLL concept is a quasi-isotropie 2000 laminate fabricated into a rotor having a contoured shape similar to the idealized Without Kevlar wrap uniform stress rotor (Stodola) for iso- tropic materials. The "proof-of-prlnciple" flywheel was, in fact, manufactured very 1500 nearly to this idealized contour. Excep- 0 tions were that the central area was flat a-3 a-5 cn-7 a-9 to accommodate the bonded hub and that Alpha construction there was no attempt to modify the contour to compensate for a finite radius versus the idealized infinite radius. Figure 2 Fig. 1. 2-D finite-element analysis is a sketch of the wheel "as built." of E-glass flywheels, with center hole. The laminated slab from which the Included in the G.E. continuing wheel was machined was made of Celion analytical effort is a study of the 6000 carbon fiber prepregged with Harmco effect of a hoop-wound overwrap on the 5213. It consisted of 160 lamina oriented laminated disk. The winding has two at 9o/!t5°/90o/135°, etc. symmetrical about purposes. One is to reduce "shredding" the center plane of the wheel. This slab of peripheral glass fibers of the was fabricated by the Babcock and Wilcox laminate observed in some unwrapped fly- Company. The hub was manufactured and wheels. The other is to moderate the bonded to the slab by the Lord Corp., and effects of stress concentration at the machining and balancing were performed at hub mounting holes. Material properties LLL. As machined, the wheel was very of overwrap, thickness of overwrap, inter- nearly in balance. The minor correction ference fits, and materials costs are required was accomplished by "moving" being evaluated. the hole in the hub. There was no vibration problem during the spin test. Recent effort has concentrated on manufacturing process development directed Figure 3 shows the finished wheel. toward high-quality production of alpha The appearance of the numerous concentric cross-ply wheels for spin testing. The rings is the effect of contour machining procedures developed thus far include exposing in turn each lamina from the stamping out circular lamina, accurate central area out to the edge thickness. positioning in the lay-up, evacuating Nonuniformity of these rings, like con- and hydroclave pressing and curing. tour lines on a map, indicate nonuniformity Quality and uniformity are indicated by of the lay-up. Although noticeable, this the data in Table 1. The data are for was not believed to be serious.
71 -24 in. dia.-
Lord Corp. rubber bonded hub
Test wheel Material — Celion 6000 carbon fiber prepreg with Narmco 5213 Lay-up — 0°/±45°/90°; center plane symmetry; 160 Lamina Wheel weight - 11.5 Ib.; fiber volume 62%; density 0.056 Ib./cu. in.
Fig. 2. The Lawrence Livermore Laboratory "proof-of-prineiple" flywheel.
Fig. 3. The LLL flywheel for test No. 1.
The flywheel was tested at the and that fracture of the wheel probably Applied Physics Laboratory, Johns Hopkins initiated near the center. University. Failure occurred at 36,000 rpm and was sudden and catastrophic. A two-dimensional finite-element Energy density at this speed was 62.6 Wh/kg analysis of the wheel stresses and (28.h Wh/lb) for the entire wheel. Exami- strains had "been performed prior to the nation of the debris indicated that the test. This code uses uniform in-plane hub bonding had performed satisfactorily properties and uniform, but different,
72 axial direction properties. As-built lower stresses and strains at failure dimensions and "published data" properties than the 0° tests. We conclude the of the material were used. This analysis lack of strength isotropy contributed gave an expected maximum energy density significantly to the lower-than-expected of 120 Wh/kg (5iu 5 Wh/lb). We anticipated performance of the rotor. a lower energy density due to in-plane anisotropy of the material and manufac- Our conclusions on the results of turing defects in the rotor although this first test flywheel are: we did not observe appreciable defects during machining of the rotor. • The slab of material from which the wheel was machined had strength properties There was adequate excess material only about 2/3 of what one should expect from the lay-up slab to obtain "actual" of this material. material properties. However, these • The selected lay-up of O°/±lt5°/9O°was had not been tested prior to the wheel somewhat less isotropic in strength (in- spin test. These tests have since been plane) than we expected. conducted and we found that the "actual" • The design and analysis of the wheel material properties, stress and strain appears to be verified. LLL will publish at failure, were only about 2/3 of the a complete report on building and testing "published data" properties. Table 2 this flywheel. gives a comparison of published data, actual material data, and spin test data. Based on these rather limited data, we feel that the rotor performance was seriously reduced due to a material problem. Discussions with the material fabricator have not revealed the cause of the problem. It is speculated that lengthy or improper storage of the prepreg material or an inadequate cure cycle may have contributed to the problem.
To further evaluate the material problem, we obtained test panels of the same material from Celanese Company, the material supplier. Samples of a uni- directional fiber lay-up and a 0°/±l)5°/90o lay-up were tested and gave results essentially equal to the published data. In addition to testing the 0°/±2i5o/90o material at 0°, we tested samples at 22-1/2°. These 22-1/2° tests gave much
Table 2. Comparative properties of the O°/±k5°/9O° laminate and flywheel test data.
Failure Umaxial Biaxial stress strain, %& stress, kpsi estimate, a/(l-v)
Published data 1.20 77-8 112.7
Sample from flywheel 0.78 51.1 78. h blank
Calculated from spin o.6o — 61.7 test
Spin test as percent 78 78 of flywheel material aAt 0 deg.
73 REFERENCES 1. Christensen, R. M. and Wu, E. M., Optimal Design of Anisotropie (Fiber-Reinforced) Flywheels, Lawrence Livermore Laboratory, Rept7 UCRL-52169, November 1976. 2. Toland, R. H., Current Status of Composite Flywheel Development, Lawrence Livermore Laboratory* Rept. UCRL-806CA, January 1978. 3. G.E. Company, Volume I - Executive Summary — Demonstration of an Inductor Motor/ Alternator/Flywheel Energy Storage System, Phase I - Final Report, January 27, 1978. h. Lustenader, E. L. and Zorzi, E. S., "A Status of the a-ply Flywheel Concept Development," Proceedings of the Society for Advancement of Materials and Process Engineering, May 1978. 5. McGuire, D. P. and Rabenhorst, D. W., "Composite Flywheel Rotor/Hub Attachment through Elastomeric Interlayers," 1977 Flywheel Technology Symposium Proceedings, October 5-7, 1977-
NOTICE "This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Department of Energy, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or respon- sibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately-owned rights."
Reference to a company or product name does not imply approval or recommendation of the product by the University of California or the U.S. Department of Energy to the exclusion of others that may be suitable.
74 PROJECT SUMMARY
Project Title: Flywheel Technology for Automotive Use Principal Investioator: B. B. Smith/F. W. Jones Organization: Union Carbide Corporation Nuclear Div. Y-12 Plant P. 0. Box Y Oak Ridge, TN 37830 (615) 483-8611 Ext. 3-7385 Project Goals: Assist DOE in development of composite flywheel technology for automotive use; conduct testing operations to assess performance, determine failure modes, measure transient loads transmitted to vehicular-type containments; generate engineering design data for vehicular systems. Project Status: A prototype composite flywheel was designed, fabricated, and evaluated in FY 1O76-76T. An improved flywheel and an instrumented cc-^nment assembly were designed and fabricated in FY 1977. Testing of these components will be performed in me UCC-ND flywheel test facility built in FY 1978. This facility is approximately 90% completed. contract Number: W-7405-ENG-26 Contract Period: Oct. 1, 1977 - Sept. 30, 1978 Funding Level: $170,000 (FY 1978) Funding Source: DOE, Division of Energy Storage Systems, Advanced Physical Methods Branch
75 FLYWHEEL TEST FACILITY
B. B. Smith Union Carbide Nuclear Corporation P. 0. Box Y, Bldg. 9998, MS 001 Oak Ridge, Tennessee 37830
ABSTRACT A flywheel test facility was built at the Y-12 Oak Ridge Plant during Fiscal 1978. The facility is designed for 2 KWH energy level with minimum speeds to 60,000 RPM. The test machine is located below grade in a concrete pit. The objectives for constructing the facility are to provide a safe facility for pe forming dynamic testing while develop- ing a composite flywheel and to provide test data for containment and mounting studies. The facility is located tn an uncleared area to allow relatively easy access to industri- al representatives and other government agencies.
INTRODUCTION following specifications. The specifica- tions were also selected such that a nomi- A facility for testing and evaluation nal 0.5 KWH vehicular flywheel could be of high performance composite flywheels developed and failure tested at speeds and has been installed at the Union Carbide energy levels consistent with full utili- Corporation-Nuclear Division, Y-12 Plant. zation of the high-performance composite The flywheel test machine is designed to materials (i.e., 40 to 60 watt-hrs/lbs) withstand a torque of 400,000 ft. lbs., without being limited by test stand and is mounted on a 12-inch high sub- integrity. base .to allow for quartz windows through the floor of the machine for future 1. Maximum kinetic energy - 2 KWH dynamic studies using high speed photog- No combination of speed, weight, and raphy. The 4.5 inches of containment geometry can exceed a stored energy is comprised of two rotatable liners level of 2 KWH [i.e., KE = Funct. inside a 1.5 inch thick vacuum vessel. (speed, weight, geometry) <_ 2 KWH]. The inside diameter of the 2 inch thick inner liner is 37 inches. The vacuum 2. Maximum flywheel O.D. - 30 in. system is capable of maintaining 10 mm Hg. 3. Maximum RPM - 60,000 rpm. The test flywheel will be driven maximum 4. Maximum weight - 50 lbs. test speeds of 60,000 RPM with a 4-inch 5. Vacuum - <10 mm Hg. Barbour-Stockwell air turbine. The test facility is designed for in-place bal- 6. Adequate ports and space for mirrors ancing using distance detectors, oscillo- and associated equipment to permit scope, x-y plotter, and a tracking use of high-speed photography in analyzer. future activities. 7. Freely rotatable crash ring design, A precision balance machine is available for pre-balancing flywheels prior to test- 8. Constrained arbor design. ing. 9. 4" Barbour Stockwell turbine. 10. Capability to structurally attach DESIGN CRITERIA and effectively test the FY77 fly- wheel and containment assembly. The design specifications were to provide a spin test stand capable of i testing a flywheel in accordance with the
76 ~ DESCRIPTION The flywheels to be tested will be flfflkTFI TBT FACILITY driven with an air turbine mounted on the top flange of a test machine as shown In Figure 1. DttlCN EMMY LEVEL 2 MM
DHIM FACTO* » SAFETY H
DMIM MOUNT TOMM 400,004 rt-utf
EXPECTED Hourr TOMUC 150/000 FT-Up
DESIfN MOMENT TO HOUHT 287,000 rr-Up
EXPECTED HOHENT TO IfOUNT 253,000 FT-u,
DfSIM RADIAL iMD 100,000 FT-Up
CONTAINMENT THtCKMM - HWIAL 4.5 INCHES
Tor FLMK 2.5 INCHES
BUI IWt 2 INCHES
Figure 2. Qe|1gn for Flywheel Test Facility
It IIC FLYMHEEL
Figure 1. Cross Section of Test Machine 1977 DESIGN
The horizontal test machine provides for KINETIC EHERSY 1.12 KM quick access to the test flywheel through a 2-1/2 inch thick steel flange. The top VEI«HT 25 u,, flange is secured to the vacuum casing BURST ENEMY with 12 one inch SAE grade 5 bolts.which 44-H/U, will withstand an axial force of 188,000 OELAMINATION <35,Q00RPN Ibf with a factor of safety of 4. SPEED • 1.12 nm 38,350 RFM
The test machine has two rotatable EXPECTED TOMUE containment liners and a 1-1/2 inch 123,840 FT-Up thick casing wall. The machine is IAXIHUM TORQUE • BURST SPEED 10,209 x EQ6 FT-U. mounted to a concrete mass with 16 one inch SAE grade 5 bolts. A sub-base 1s Figure 3. 20 Inch Flywheel Design Criteria provided to accommodate optical viewing of the flywheel through the bottom W1«CH nvmn fiance for dynamic analysis and to study failure data of a composite flywheel. KINETIC ENEMY 2KHH A vacuum system will provide an atmosphere of about 10-' mm Hg. HEISHT 50 Up The mass of the concrete base is SELAHIKATION COLLAPSE LOAD 88,000 Up greater than the machine mass. The SPEED 8 2 KMH 22.458 RPN concrete mass is 43,000 1bsm. RADIAL LOAD F - NE»* 2,507,968 Up
CASINS HOVEHENT CAUSED >Y ORJITUM 0.078 INCHES FLYNHEEL Figure 4. 30 Inch Flywheel Design Criteria
77 v~;--'--i^tt::-J-^-:..-ii.--'
Figure 5. Flywheel Test Facility Figure 6. Flywheel Test Machine Mounted in a Concrete Pit
78 SESSION II; FLYWHEELS
79 PROJFXT SUMMARY
Project Title: Mechanical Energy Storage Technology Development Principal Investigator: R. 0. Woods Organization: Sandia Laboratories Albuquerque, NM 87185 (505) 264-7553 Project Goals: The selection of those mechanical components of flywheel energy storage systems that can be carried to their next stage of evaluation by efforts within the scope of our budget; the funding of developmental work treating such components. As a corollary, the documentation of the state-of-the-art in specific component technology where such information is scattered or fragmentary. Project Status: Efforts to date have treated the following components: Rolling Contact Bearings Magnetic Bearings - Active and Passive Composite Wheels - Hardware development, development of analytic and testing techniques. Composite Materials - Analysis, improved properties, vacuum properties, fabrication and thermal stress characteristics Ferrofluidic Seals - Power dissipation, lifetimes, permeation rates Funding has been provided to other agencies for the testing of overall energy storage systems,and the groundwork has been completed for a program to deal with the problens of vacuum technology. Contract Number: 5OL 78 Contract Period: FY79 Funding Level: $1,107 M (BO) * Funding Source: DOE, Division of Energy Storage Systems
*Included in this project are the following: E. D. Reedy, Jr., "Sandia Composite-Rim Flywheel Development," page 87; A. Keith Miller, "Structural Modeling of a Thick Rim Rotor," page 93; Melvin R. Baer, "Aerodynamic Heating of High-Speed Flywheels in Low-Density Environments," Page 99.
81 OVERVIEW OF COMPONENT DEVELOPMENT
Robert 0. Woods Sandia Laboratories Division 1+715 Albuquerque, KM 87185
ABSTBACT Sandia Laboratories is charged with advancing components technology for flywheel energy storage systems. Portions of this work are being done in-house, the balance on contract with outside agencies. At this writing, seventeen specific efforts can be identified. These comprise seven tasks under the heading of composite wheel develop- ment, four in bearings, three in vacuum seals, and three in general vacuum technology. Each of the major efforts will be represented at this conference by a paper presented by a representative of the organization actually performing the task. The intent of this presentation is to give an overview of the program's organization and objectives.
INTRODUCTION meet the criterion of not having already been evolved beyond a point of diminish- The functional components of a fly- ing return. The alternative to the com- wheel energy storage system are represen- posite wheel is the steel wheel; there ted in Fig. l. We have shown, by putting the technology is well understood and the labels in boxes, the subsystems which major advancements are unlikely at our have been identified as needing further funding level. In addition, composite development and that have not been worked- wheels may have attractive features such over to the point where prohibitively as greater energy density and safety large expenditures will be needed to when compared to steel wheels. achieve minor quantitative improvements. These items are the ones upon which we Sandia Laboratories, in addition to judge money will be spent to best advan- its own in-house efforts aimed at develop- tage . ing analytic techniques and at better understanding the behavior of composites Figure 2 indicates the time frame in flywheel applications, is currently projected for the program. All of the ef- funding the development of four new wheel forts spanning FY78/FY79 are in fact under concepts. Each of those will be described way. Negotiations for some (but not all) in detail by the contracting organization. of those starting in FY79 have been ini- Together, these concepts exhaust the range tiated. Note that late Spring 1979 repre- of geometries proposed to date. The tim- sents a well-defined transition. At that ing of these programs is such that preli- point the most fundamental components minary information regarding each of the studies will have been completed; although designs will be available when engineer- development of selected components will ing of the demonstration module begins. continue, it will at that time be possible to incorporate study results into the de- BEARINGS sign of a next-generation flywheel module intended for evaluation and demonstration. These programs will also be de- scribed by the contractors. In the case WHEEL DESIGN of rolling contact bearings, Sandia has funded a combined laboratory and theore- All efforts in wheel development have tical study which has explored the capa- thus far addressed the problems of wheels bilities of what may be the next-genera- made of composite materials. This is not tion design for ball bearings. At the only because it was Sandia's experience same time, this study provides inputs to with composites which first involved us a computer code that will, for trade-off in the flywheel effort. Composite wheels studies, predict bearing operating char- are undergoing development because they acteristics over a range of parameters
82 such as speed, size and loading. properties of the composites themselves, the effects of shock and vibration en- Two studies are also being funded to countered in a vehicle environment, and explore the characteristics of magnetic the close tolerances (hence high prices) bearings — both active (servoed) and found in most rotating vacuum equipment. passive. It is generally thought that The only hard datum is that a vacuum such bearings will find use only in sta- on the order of 10"* torr will be re- tionary applications. At some time in quired if the potential of a composite the future, probably mid-1979> we hope to wheel is to be realized. This informa- also fund work on hydrostatic bearings. tion in itself is not adequate to speci- The results of such work could find ap- fy the problem. What is also needed is plication in a later generation flywheel information regarding pumping rates at demonstration unit. that pressure. These will be controlled by the outgassing properties of the com- SHAFT SEALS posites and the permeation rates of the seals. No reliable information exists, We have concluded, and it appears although we have made a first step by that our judgement is supported by a funding a series of outgassing tests majority of the workers in the field, which investigated representative compo- that for high-performance (read "high sites . speed") flywheel systems, energy input and output will probably be electrical, The current status of the vacuum using motor/generators sealed within the technology program is that we are active- vacuum housing. If, however, it were ly trying to locate organizations work- possible to produce rotating shaft seals ing in the field and to interest them capable of the low permeation rates, high in undertaking all or any part of the surface speeds, and long wear lifetimes problem. needed for flywheel applications, a num- ber of options would be opened that CONCLUSION might make electric input/output look less attractive. With the current state- This is an open-ended program whose of-the-art in seals technology the pos- objective is to upgrade a wide range of sibility of using a shaft and directly- components for flywheel energy storage coupled mechanical transmission exists systems. There is no single touchstone only for a limited range of applications. by which the success of the program may In the interest of establishing (or be judged; rather, the objective is to demolishing) the credibility of mechani- produce a continuous progression of cal input/output concepts, we are funding technological improvements. By mid-1979 two seals studies and hope to fund a it should be possible to incorporate the third. The first two will establish the outputs of the program as of that date operational envelope for present-genera- into a demonstration unit. That unit tion shaft seals and perhaps produce should represent an improvement over what next-generation hardware. The later study, could now be produced. At the very that of the so-called "dynamic" seal is least, it will represent a permutation properly part of a vacuum technology of components which can credibly be program which is not yet under way. called "optimized". VACUUM TECHNOLOGY One of the companion papers to be delivered at this conference will report on a study of the aerodynamic heating of composite wheels. The unavoidable conclu- sion arising from this work is that the vacuum requirements of a high-performance composite wheel are stringent to the point of entailing a major effort if such wheels are ever to evolve beyond the state of a laboratory curiosity. The main pro- blems in vacuum technology are the dis- mally poor energy efficiencies of present- day pumping equipment, the outgassing
83 VARIABLE SPEED TRANSMISSIONS
SHAFT SEALS • FERROFLUIDIC • FACE
MOTORS/GENERATORS
FAILURE SENSORS POWER CONDITIONING CIRCUITRY
HEAT TRANSFER WHEEL DESIGN CO • AERO HEATING • CARBON SPOKE/RIM • INTERNAL • KEVLAR SPOKE/RIM TEMP. FIELD • RIMLESS • SPECIAL CONTOUR • TEST & ANALYSIS GsaSK CONTAINMENT
VACUUM BEARINGS • PUMPING • PASSIVE MAGNETIC • OUTGASSING • ACTIVE MAGNETIC • SEALS • ROLLING 'BACKGROUND TECHNOLOGY" SANDIA FLYWHEEL PROGRAM FY78 FY79 FY80 FY81 TASKS ONDJFKAMJJAS ONDJFMAMJJAS ONDJFMAMJJAS 0HDJFHAMJJA8 WHEEL DESIGN
• StA Design and
•SLA Testing « •• *• • Conrcercial "ho W-nr/lb" Wheel Carbon Spoke -
Kevlar Spoke - Rialess - Over- Rialesa - Contour Test and Analysis
BEARINGS
00 in •hydrostatic ro
SHAFT SEAIS
•Dynamic (cf. "Vacuum")
VACUUM TECHHOMGT
• Pmnping
PROCCCWE DEVEIOFMEBT
•Engineering •Fabrication •Teatiag *ai Dnonstratloa SANDIA COMPOSITE-RIM FLYWHEEL DEVELOPMENT*
E. D. Reedy, Jr. Sandia Laboratories^ Composite Materials Development Division 5844 Albuquerque, New Mexico 87185
ABSTRACT
This paper will briefly discuss a series of spin tests completed in March, 1978 (just prior to the last DOE Flywheel Contractors' Meeting) and report on the status of Sandia's current rotor development program. Two designs incorporating a 20 in. O.D. graphite/epoxy rim were tested. Each design utilizes a different method of attaching the rim to an aluminum hub with Kevlar 49/epoxy bands. Data on rotor performance, dynamics, and failure modes were gathered. Two rotors with a pin-wrapped hub design were spun to tip speeds in excess of 2500 ft/sec. This corresponds to an energy den- sity, based upon total rotor weight, of approximately 20 Wh/lb. These tests have helped identify areas requiring further research. In particular, there is a need for predic- tive dynamic analysis for this class of fJywheel.
INTRODUCTION attach the rim to an aluminum hub. Kev- lar 49/epoxy bands were chosen to match A hybrid propulsion system which the radial displacement and tip speed utilizes a flywheel for energy storage re- capability of the rotating rim.2 The quires a rotor which can both store energy Alleghany Ballistics Laboratory of Her- efficiently and fail in an easily con- cules, Inc. fabricated two rotors incor- tained manner. A flywheel design which porating their suggested pin-wrapped hub incorporates a circumferentially wound rim connection (Fig. 1). The hub region of a was identified as being likely to satisfy these design requirements as well as lending itself to existing fabrication techniques. An analytic study was per- formed to identify feasible design and material choices for a filament wound rim which is of a size appropriate for a hybrid vehicle.1 Based upon this analy- sis, two flywheels were designed to have (1) a total weight less than 25 lbs, (2) a rim swept volume of approximately 0.5 cubic feet, and (3) a storage capacity of 0.56 kWh at 31,000 rpm in order that 0.5 kwh could be delivered with a 3 to 1 speed reduction. Both designs incorporate a 20 in. O.D., high strength graphite/epoxy rim which has an inner-to-outer radius ratio of 0.7625. The rim has a semi-elliptical cross section with its 3 in. high flat Figure 1. Pin-Wrapped Flywheel edge facing the hub. The designs differ in the method in which overwrapped bands Pin-Wrapped flywheel is shown in Figure 2. Two rotors which utilize six radial over- wrapped bands to connect rim and hub This work was supported by the U.S. De- (Figs. 3,4) were fabricated by the De- partment of Energy (DOE) , under Contract fense Division of the Brunswick Corpora- AT(29-l)-789. tion. Due to its appearance, this rotor A U.S. DOE facility. design was designated the Wagon Wheel...
87 Figure 2. Pin-Wrapped Flywheel Hub Figure 3. Wagon Wheel Flywheel
The results of the spin tests of these rotors as well as the current program goals will be discussed below.
SPIN TESTS RESULTS
Four flywheels, two of each design, were tested at Sandia Laboratories, Liver- more. These tests are described and dis- cussed in detail in Ref. 3. The results of the tests are summarized in Table 1. The highest rotational speeds were ob- tained by the Pin-Wrapped rotors (29,000 and 30,100 RPM). The lower speeds reached by the Wagon Wheel rotors may be due, in part, to the presence of pre-existing rim flaws which apparently grew during spin-up. Also, the first Wagon Wheel rotor tested Figure 4. Wagon Wheel Flywheel Hub (Test #2) experienced a migration and loss of balancing weights during spin-up. The maximum tip speed was reached by a pin- or (3) a combination of the above. There wrapped rotor, 2625 ft/sec (Test #4). At is evidence of an increasing rotor im- failure this rotor stored 0.532 kWh, which balance in Test #2 due to migrating is within 5% of the design goal of 0.56 balance weights, in Test #3 due to fray- kWh. This corresponds to an energy den- ing bands, and in Test #2 and #5 due to sity of 21.3 Wh/lb based upon total rotor growing rim flaws (see Table 1). On the weight. other hand, calculations by C. W. Bert et. al.1* indicate the existence of criti- Rotor dynamics were monitored by cal whirl speeds of 23,000 EPM for the proximity gauges which measured the hori- Wagon Wheel design and 23,500 EPM for the zontal displacement of the adapter con- Pin-Wrapped design. These predictions necting the flywheel to the spinner shaft. may be altered to give better agreement In all tests the flywheels initially spun with experiment when more accurate esti- stably as their speed was increased. How- mates for rotor stiffnesses become avail- ever, each test (except Test #1 in which able. there was a facility failure) was termi- nated by excessive shaft rurout. The One rotor of each design was spun cause of this increasing 2r-»i runout has until its runout failed the breakaway not yet been determined. It appears that shaft and dropped it into the spin pit. it could have been caused by either (1) an Although the failures generated may not increasing rotor imbalance due to a struc- represent a spontaneous rotor burst, they tural failure or a migrating balance may be typical of those induced by a vreight, (2) approaching a critical speed, growing flaw which causes increasing
38 TABLE 1. COMPOSITE-RIM
TEST FLYWHEEL MAX SPEED (RPH) COMMENTS 1 Pin Wrapped 23,800 Terminated by facility failure #1 Bands abraided by instrumentation debris
2 Wagon Wheel 17,900 Pre-existing surface flaws (rim) #1 Terminated after balance weights thrown off Rim separations
3 Pin Wrapped 29,000 Same flywheel used in Test #1 #1 Terminated by excessive shaft runout Bands frayed where abraided in Test #1
4 Pin Wrapped 30,100 Best performance #2 Driven until shaft failed
5 Wagon Wheel 22,100 Pre-existing internal flaws (rim) #2 Driven until shaft failed Rim separations
runout to break the shaft. In both tests, 3-D finite element model is being used to the bands were ripped from the rim and study the interaction between the rim and shredded. The rim remained intact (Fig. bands. Such analysis could be used to 5). This failure mode is conducive to optimize the choice of band material and containment since there are no pieces with rim geometry. high radial momentum. Methods for predicting and control- CURRENT PROGRAM ling rim fabrication stresses are being pursued as this may provide a .•neans of Since, as discussed above, the origin extending a rim's operating speea. An of increasing shaft runout is not known, experimental program to measure rim re- its identification is a principle goal in sidual stresses is underway. In these the current program. An intensive effort tests the hoop strain at the rim I.D. is is underway to better understand rotor monitored as layers of the outer surface dynamics. Bert et_. al^.1*'5 are using lumped mass techniques which use calcu- lated estimates for distributed band, shaft, and support flexibilities to model the rotor. His calculations incorporate gyroscopic effects and centrifugal band stiffening. A modal analysis of the Wagon Wheel design is being pursued by Miller.6 His finite element calculations are being calibrated against the results of an experimental modal analysis. This experi- ment was performed on a non-rotating fly- wheel with traction-free boundaries. The next series of spin tests will attempt to identify the source of the excess shaft runout. One facet of these tests will be to try to observe experimentally the analytically predicted changes in the Wagon Wheel's dynamics when the band stiffness is increased or the hub's dia- metrical moment of inertia is decreased.
In addition to dynamic analysis, a more detailed structural analysis of the Figure 5. Failed Pin-Wrapped Flywheel flywheels is planned. In particular, a (Test #4)
89 are machined away. These data can be REFERENCES used to infer the initial radial stress distribution. Predictive analytic methods 1. E. D. Reedy, Jr. and F. P. Gerstle, which correlate with the experimental dis- Jr., "Design of Spoked-Rim Composite tributions are being sought. Flywheels," Proceedings of the 1977 Flywheel Technology Symposium, San Techniques for measuring and predic- Francisco, CA, Oct. 5-7, 1977, pp. ting a rotor's aerodynamic heating are 99-110. being developed. As pointed out by Woods et. al.,7 aerodynamic heating may be par- 2. E. D. Reedy, Jr., "A Composite-Rim ticularly severe for a resin matrix com- Flywheel Design," SAMPE Quarterly, posite flywheel. These composites can Vol. 9, No. 3, pp. 1-6. survive only within a limited temperature range. To further compound the problem, 3. E. D. Reedy, Jr., and H. K. Street, hoop wound composites have a poor radial "Composite-Rim Flywheels: Spin thermal conductivity which tends to allow Tests," to be published. the temperature to build up on the peri- phery. Analytic methods similar to those 4. C. W. Bert and T. L. C. Chen, "Lateral used to analyze re-entry vehicles have and Tilt Whirl Modes of Flexibly been developed by Baer8'9 to calculate the Mounted Flywheels: Analysis and Ex- effect of aerodynamic heating on a com- periment," Presented at the 49th Shock posite flywheel. Initial experimental and Vibration Symposium, Washington, efforts to measure rim temperature by on- DC, Oct. 17-19, 1978. board thermocouples during rotor spin tests have proven difficult.3 At high 5. c. W. Bert, T. L. C. Chen, and C. A. rotational speeds lead wires and gages are Kocay, "Critical Speeds and Natural ripped from the rim. Recently, an infra- Frequencies of Rim-rype Composite- red radiometer was added tc ^andia Liver- Material Flywheels," OU-AMNE-73-3, more 's spin test facility in order that September, 1978. rim temperatures can be monitored without on-board instrumentation. The data col- 6. A. K. Miller, "Structural Modeling of lected will be useful in verifying a Thick-Rim Rotor," Proceedings of the analytic predictions. First Annual Mechanical and Magnetic Energy Storage Contractors' Informa- A materials program is being con- tion-Exchange Conference, Luray, ducted by Allied, et. al.10'11 to both Virginia, October 24-26, 1978. characterize and improve through materials modifications the transverse strength of 7. R. 0. Woods and F. P. Gerstle, Jr., filament wound composites. Often the "Sandia Basic Flywheel Technology transverse strength of such materials Studies," Proceedings of the 1977 Fly- limit the energy storage capacity of a wheel Technology Symposium, San Fran- composite-rim flywheel design. The cisco, CA, Oct. 5-7, 1977, pp. 315- effects of (1) matrix toughening, (2) 322. resin/hardener chemistry, and (3) fila- ment surface treatment on transverse 8. M. R. Baer, "Aerodynamic Heating of strength are being studied. One aspect High-Speed Flywheels in Low-Density of this work was the successful develop- Environments," Sandia Report, SAND 78- ment of a self aligning specimen fixture 0957, October, 1978. to test hoop wound tubes in transverse tension. 9. M. R. Baer, "Aerodynamic Heating of High-Speed Flywheels in Low-Density SUMMARY Environments," Proceedings of the First Annual Mechanical and Magnetic The favorable rotor performance and Energy Storage Contractors' Infor- failure mode encourage the continued de- mation-Exchange Conference, Luray, velopment of composite-rim flywheels. Virginia, October 24-26, 1978. Future work will.concentrate on gaining a better understanding of (1) rotor 10. R. E. Allred, R. F. Foral, and W. E. dynamics, (2) rim-band interactions, (3) Dick, "Improved Performance for rim residual stresses, (4) flywheel aero- Hoop-Wound Composite Flywheel Rotors," dynamic heating, and (5) the factors which Proceedings of the 1977 Flywheel affect composite transverse strength.
90 Technology Symposium, San Francisco, CA, Oct. 5-7, 1977, pp. 377-392.
11. R. E. Allred and H. K. Street, "Im- provement of Transverse Composite Strength: Test Specimen and Matrix Development," Proceedings of the 24th National SAHPE Symposium, San Fran- cisco, CA, May 8-10, 1979.
91 STRUCTURAL MODELING OF A THICK-RIM ROTOR*
A. Keith Miller Applied Mechanics Division 5521 Sandia Laboratories Albuquerque, New Mexico 87185
ABSTRACT A NASTRAN structural model has been constructed of the Sandia designed, thick-rim rotor having six discrete overwrapped bands forming twelve pairs of spokes—the wagon-wheel rotor. The results of an experi- mental modal analysis of an actual rotor, using Past-Fourier transform techniques, are being used to aid in the definition and refinement of the numerical model. A description of the structural model is given in this work, and the resulting calculated normal mode shapes and frequencies are presented. These mode shapes and frequencies are compared to those obtained from the experimental modal analysis.
INTRODUCTION an outer diameter of approximately twenty inches and weighs approxi- The objective of the program is mately twenty three pounds. The to develop a flywheel rotor capable flywheel is designed to store 0.5 of being used in hybrid heat-engine kWh of energy when rotating at vehicles. One promising proposed 32,000 rpm. design of such a rotor involves a thick-rim, composite rotor having Because the success of any discrete composite spokes attaching rotor design is ultimately deter- the rim to an internal hub. The mined by the dynamic performance of proposed rotor, shown in Fig. 1, has a flywheel system, it becomes nec- essary to attempt to understand the dynamic characteristics of a rotor structure before it is incorporated into the larger system. Once the structural dynamic characteristics of the rotor are properly identi- fied, then one should be able to determine, with some confidence, the interaction between a flywheel system and a specific roto:.. One approach to quantifying tne struc- tural characteristics of a flywheel rotor, in a manner such that the information can be used in a series of flywheel system designs, is to construct a numerical, finite- element model of the rotor using existing general purpose computer Fig. 1. Thick-Rim Wagon-Wheel Rotor codes, such as NASTRAN. Experimental modal analysis of a wagon-wheel rotor has been used to verify the various assumptions This work was sponsored by the involved in the assembly of the United States Department of Energy, numerical structural model (espe- Dr. Robert 0. Woods, program cially items; involving connecting monitor. regions). When the numerical model
93 predicts mode shapes and frequencies - SPIN COLLAR of the rotor structure which are within reasonable agreement with the experimental results, it is believed that the structural model is suffi- ciently accurate to be used in future analyses of flywheel systems. ROTOR COMPONENTS A sectioned view of the wagon- wheel rotor is shown in Fig. 2. The wagon-wheel consists essentially of three components: the circular wound, carbon/epoxy rim; the uni- directional Kevlar/epoxy spokes; and the hub which includes the
aluminum cylindrical hub, the alumi- -WOUHTIHG I1LOCK num spin collar adaptor plate, the - CMrnrft r-tuG catcher plate and plug, and the steel spin collar. The mechanical properties of the graphite/epoxy, the Kevlar/epoxy, and aluminum used in the structural model are given Fig. 2. Cross-Sectional View of in Table 1. the Wagon-Wheel Rotor The rim, as shown in Fig. 2, which are machined into the ends of has a semi-elliptical cross section the hub. having a semi-major axis of 2.4 inches directed radially, and a The aluminum hub is a cylinder semi-minor axis of 1.5 inches. The having an inner diameter of 3.5 inner radius of the rim is approxi- inches and a wall thickness of 0.5 mately 7.6 inches. inches. The hub is 4.3 inches long but has slots of depths ranging The twelve overwrapped Kevlar/ from 0.15 to 0.65 inches machined epoxy spokes have a nominal width from the ends to allow the over- of 0.5 inches, and a nominal thick- wrapped spokes to pass through. ness of 0.1 inches. The spokes The aluminum catcher plate is 2.5 were measured to be approximately inches thick and 4.5 inches in 50% thicker than the nominal value diameter with a 0.6 inch diameter on the rotor used in the experimen- attached ball affixed to one side. tal model analysis. (The measured The aluminum spin collar adaptor thickness of the spokes was used in plate is 1.5 inches thick, with an the numerical model of the rotor.) inner diameter of 1.75 inches and The spokes are continuous, and pass an outer diameter of 4,5 inches, through the hub in a series of slots and is affixed to the opposite end
Table 1. Mechanical Properties of Materials Used in the Thick-Rim Wagon-Wheel Rotor
Property Graphite/Epoxy Kevlar/Epoxy Aluminum Longitudinal Modulus E. 18.00 11.00 10.50 (106 psi) Transverse Modulus E_ 1.23 0.71 10.50 (106 psi) Shear Modulus G 0.85 0.30 3.80 (106 psi) Density (lb/in3) 0.054 0.050 0.100
94 of the hub. To this is attached algorithms. The moment of inertia the steel spin collar which is essen- about a diametral axis for the hub tially a hollow truncated conical assembly is normally too large when section having a major outer diam- the computer algorithms distribute eter of 4.5 inches, a minor outer the mass to the node points. diameter of 1.75 inches, an inner diameter of 1 inch, and is 2 inches EXPERIMENTAL MODAL ANALYSIS long. All dimensions are approxi- mate. During the experimental modal analysis, a thick-rim, wagon-wheel More detailed dimensions and rotor was suspended from the catcher a description of the fabrication adaptor plate by surgical tubing to procedure of this wagon-wheel rotor simulate a free condition of the can be found in Reference 1. rotor structure. The natural fre- quency of the rotor on the tubing NASTRAN MODEL was roughly measured to be 1.2 Hz. A mounting block (See Fig. 2) was The finite element structural attached to the catcher plate model which is currently being used through which a circumferential to describe the wagon-wheel rotor impulse was input to the rotor hub consists of a series of beam ele- via a hard plastic hammer having a ments, plate elements, and solid force transducer mounted to the elements. The graphite/epoxy rim head. is modeled as a series of twelve beams each having a cross-sectional Triaxial accelerometer blocks area and area second moment equiv- were mounted at ninety degree incre- alent to the physical rim. Actual ments around the inside of the rim, orthotropic material properties of and at the top and the bottom of the the graphite/epoxy are used for hub assembly. these beam elements. As an impulse was supplied to Each of the twenty four Kevlar the structure, the time responses spokes were also modeled as a series of the accelerometers were measured of beam elements again having ortho- at a sampling rate of 5000 Hz (once tropic mechanical properties. The every 0.2 milliseconds). The res- hub was modeled as a series of plate ponses were averaged from ten im- elements consisting of four rings pulses, and experimental transfer with twelve elements defining each functions for each accelerometer ring. The catcher plate, the spin were obtained from a Hewlett-Packard collar adaptor plate, and the spin 5451B Fourier Analyzer, using a_ HP collar were all described in the software package. Fig. 3 is a typi- model by three-dimensional solid cal plot of an experimental transfer elements. 0.9 1.0 The modulus and mass properties 0.4 .8 were included as material properties for the beam elements of both the 0.3 .6 rim and the spokes. The computer 0.2 .4 code then distributes the mass of O.I 1 .2 i each element equally at each node 3 o k ] 0 1 i point at the ends of the bars. Only 2-o.i 1-.2 the modulus of the material used in 1 i r the solid elements was input as a -0.2 -.4 material parameter. Equivalent -0.3 -.6 masses were directly input to the -0.4 node points describing the hub -0.5 600 1000 1500 2000 B00 1000 1900 2000 assembly so that the moment of iner- (HZ) (HI) tia of this assembly could be REAL IMAGINARY adjusted to better model the struc- ture. The masses are not distri- Fig. 3. Experimental Transfer buted precisely enough for the hub Function on the Spin-Collar-End of assembly when done by the NASTRAN the Hub for the Wagon-Wheel Rotor
95 function; this particular one is modes, and descriptions of the mode for an axial accelerometer mounted shapes determined from the experi- to the spin collar side of the hub mental analysis are given in Table assembly. The plot on the left of 2. Fig. 3 represents the real portion of the transfer function, while the The fourth mode, which is a plot on the right represents the whirl mode of the hub with respect imaginary portion. Once the experi- to the rim, may be one of some con- mental transfer functions are cern. Although the frequency of obtained, a series of analytic trans- this mode (630 Hz) is beyond the fer functions are "fit," with the proposed operating range for the aid of the Fourier Analyzer, to rotor, it is important to consider match the experimental ones. The the fact that this mode was identi- "fitted" transfer functions are the fied while the rotor was in a static ones from which the model frequen- condition where centrifugal stiffen- cies and damping factors are actually ing of the spokes is not present read. (which may raise the frequency of this mode), and while the rotor is The peaks in the transfer func- in a free condition—not attached tions, such as are shown in Fig. 3, to the air turbine used in the spin locate the frequencies of the reso- testing. When the rotor is attached nant vibration modes in the struc- to the air turbine the frequency of ture. The amplitude of a particular this mode will be lowered because peak gives an indication of the the additional flexibility of the turbine shaft is in series with the relative strength of that mode. The rotor spokes for this mode shape. width of a peak is related to the amount of damping in the structure associated with that mode. It is presently unclear how strongly centrifugal stiffening of The modes for the wagon-wheel the spokes will affect the whirl rotor were, in general, found to be frequency. However, a preliminary very distinct and nonoverlapping, analysis using the present NASTRAN which indicates that the structure model of the rotor and a current was behaving essentially linearly estimate of the air turbine struc- for the input excitation used. ture indicates that the whirl fre- Although the normal modes up to quency may be decreased as much as 2000 Hz can be identified from the 30 percent when the rotor is pendu- transfer function shown in Fig. 3, lous ly attached to the turbine. If only those below 1000 Hz (60,000 centrifugal stiffening of the spokes rpm) are examined. Modes at fre- does not increase the whirl frequen- quencies above 1000 Hz are consi- cy, a 30 percent reduction of this dered to be sufficiently above the frequency places it near the 25,000 upper limit of the proposed oper- rpm range. As was noted by Reedy3, ating range that they will probably previous spin tests of a wagon-wheel not be excited during the rotor's rotor indicated possible dynamic operation. Therefore, only the instabilities near this angular frequencies of the first four normal speed.
Table 2. Frequencies and Normal Modes of the Thick-Rim Wagon-Wheel Rotor
Experimental Numerical
Mode Frequency Mode Frequency Description Number (Hz) Number (Hz)
1 102 1 90 Axial motion of hub w.r.t. rim 2 376 2 373 Torsional motion of hub w.r.t. rim 3 605 Hub oblonging (highly damped) 4 630 634 Whirl of hub w.r.t. rim
96 The whirl mode is also the one the transfer function shown in Fig. easily excited by gyroscopic moments 3 seems appropriate. which develop from mass eccentrici- ties as a rotor is spun. Therefore, In Fig. 3, both of the peaks in it is highly desirable to have the the real and imaginary portions of frequency of this mode as far as the transfer function associated with possible from the operating range. this questioned mode (at 605 Hz) are The relatively massive catcher plate of relatively small amplitude, indi- and plug were removed from the rotor cating the response of the rotor at to investigate the possibility of this frequency is not great. Also, raising the frequency of the whirl the peak in the imaginary trace is mode. The experimental analysis was wide, indicating a large amount of repeated and the whirl frequency was damping is associated with this mode. found to be raised 80 Hz (13%) while The mode at 605 Hz is obviously not the rotor was in the static, free strong, nor is it a clean mode. The test state. peak may be either the result of some nonlinear behavior of the rotor NASTRAN MODEL NORMAL structure when it responds to the mechanical impulse input or a small The frequencies of the first beating mode resulting from two three normal modes and descriptions higher frequency modes which are of the mode shapes calculated from excited by the impulse. In any the NASTRAN model are also given in case, this mode is not of serious Table 2. In general, the frequen- concern. If the mode is the result cies predicted by the model agree to of some nonlinear response, the be within 10 percent of the experi- structural damping associated with mental modal frequencies. it is large enough that th° response to possible excitations should remain The structural model does not small; or if the mode is the result predict a mode shape as observed in of two higher frequency modes, those the experimental analysis for mode higher frequency modes are well above number three (605 Hz). Instead, the the proposed operating speed of the shape of mode number three predicted rotor and will not be excited; there- by the numerical model corresponds fore, no beating will exist. to the shape of mode number four from the experimental analysis. It was discovered during the When a particular mode is predicted assembly of the structural model by the numerical model which is not that the distance from the plane observed in the experimental analy- formed by the rim (plane A-A in sis, an immediate question arises of Fig. 2) to where the spokes attach whether the mode was sufficiently to the hub strongly influenced the excited by the chosen mechanical whirl frequency predicted by the input to be detected by the trans- model. The further from this plane ducers. In this situation, an input (plane A-A) that the spokes were is selected which will sufficiently fastened to the hub, the lower was excite the mode, if it exists. the whirl frequency. The points Normally the mode does exist and is where the spokes connect to the hub detected when the alternate mechani- in the actual rotor vary in distance cal input is chosen. However, when from the plane of the rim. The a mode is detected by the experimen- inclination of the spokes vary in tal analysis which is not obtained uniform increments from being essen- from the numerical model (as in the tially parallel to the rim plane, current case), then both the model to vectors outwardly inclined appro- and the experimental data must be ximately four degrees when trans- re-examined. Because the predicted versing from the rim to the hub. modal frequencies from the numerical It is anticipated that the results model are in good agreement with the of this discovery will be further corresponding frequencies of the considered if it becomes necessury experimental analysis for the remain- to increase the whirl frequency for ing mode shapes, a re-examination or this design of rotor.
97 FUTURE EFFORTS The thrust of the effort to date has been to assemble a struc- tural model of the thick-rim, wagon- wheel rotor which is accurate and can be used reliably in future studies of flywheel systems. The immediate future effort will be to incorporate the current rotor model into a larger structural model con- taining the spin pit dynamic char- acteristics. The enlarged model will be used in an attempt to pre- dict possible dynamic instabilities as the rotor is spin tested. Items such as the effects of centrifugal stiffening of the spokes and the effects of removing the catcher plug on the dynamic characteristics of the flywheel system are expected to be addressed. Frequency response studies of various force inputs to the flywheel system are also anti- cipated.
The structural model can also be used to investigate the dynamic effects of possible design changes of components of the rotor in advance of fabrication of a new rotor. Dynamic control studies based on accepted structural models are also possible future efforts. ACKNOWLEDGEMENTS The author wishes to thank Mr. A. R. Nord and Mr. C. M. Grassham lor the excellent experimental modal analysis. The author also wishes to acknowledge the many useful discus- sions with Dr. E. D. Reedy concern- ing the structural modeling of the wagon-wheel rotor. REFERENCES JE. D. Reedy. SAMPE Quarterly, 9,3, ^(April 1978) 2E. D. Reedy, Proceedings of the First Annual Mechanical and Magne- tic Energy Storage Contractors' Information-Exchange Conference, Luray, Virginia (1978).
98 AERODYNAMIC HEATING OF HIGH-SPEED FLYWHEELS IN LOW-DENSITY ENVIRONMENTS*
Melvin R. Baer Fluid Mechanics and Heat Transfer Division 5512 Sandia Laboratories Albuquerque, New Mexico 87185
ABSTRACT
This study addresses the aerodynamic heating of high-speed flywheels in low- density environments. A computer code has been developed to predict temperature fields in flywheels of variable geometry and consisting of multiple composite materials with nonisotropic, temperature-varying thermal properties. Allowances have been incorporated for variable environmental conditions, time-varying spin rates, and the choice of slip-flow or free-molecular aeroheating. Major results from the code indi- cate that environmental pressures below 10~3 torr are necessary to avoid steady-state temperatures exceeding 50°F above ambient. Protruding surfaces have the major potential to cause thermal problems.
NOMENCLATURE a speed of sound flow angle of incidence Cp specific heat flywheel angular velocity G gap distance between flywheel and housing SUBSCRIPTS k thermal conductivity p gas pressure c convective Pr gas Prandtl number r radiation q heat flux H housing inner surface r radial coordinate W flywheel surface r1 recovery parameter CO ambient 5 reduced Mach number f fiber direction St Stanton number 1 normal-to-fiber direction T temperature z axial coordinate a. accommodation coefficient INTRODUCTION y ratio of specific heats 6 continuum boundary layer thickness The introduction of high-strength, e emissivity low-density materials such as Kevlar-49 H gas viscosity or other fiberglass/carbon-fiber com- \ gas mean-free path posites has greatly improved the energy (T Stefan-Boltzmann constant storage capabilities of high-speed fly- wheels. However, the requirements of a long system lifetime and a high power- Sandia Laboratories is a U. S. conversion efficiency constrain their use Department of Energy (DOE) facility. as a storage unit. At high rotational This work was supported by the USDOE speeds, aerodynamic considerations are under Contract AT(29-l)-789. important because of the induced parasitic
99 frictional losses and the associated ir- defined the laminar boundary layer thick- reversible conversion of kinetic energy ness for the rotating flow as to heat. It has been suggested that these undesirable effects can be virtually elim- inated by operating the system within a low vacuum environment (below P < 10"^ Note: This characteristic dimension is torr).' However, this vacuum require- independent of flywheel radius. ment may be impractical or unrealistic for some systems. For example, out- gassing characteristics of the organic composite flywheel may limit the achiev- able degree of vacuum.
Although drag losses can be re- COMPOSITE FLY WHEEL duced, heat-transfer effects may be significant since aerodynamic heating occurs over a long operating lifetime, and this integrated heating determines the temperature fields within the flywheel. The heating effect is of major concern because composite materials are heat- sensitive and experience severe degrada- tion of strength when the temperature exceeds 150°F.2 Also, the possibility of a pressure runaway condition exists because of an increased outgassing rate Fig. 1. The geometry of the flywheels at elevated temperatures. examined in this study. This study addresses the prediction As the pressure within the housing of temperature fields within high-speed is dropped (maintaining the ambient tem- flywheels spinning in low-density environ- perature), the mean-free path between ments. A computer code was developed consecutive molecular collisions becomes for this purpose that incorporates the larger. For an ideal gas, this dimension allowances of composite materials with is given as nonisotropic temperature-varying prop- erties, variable environmental conditions, X = 1.26N/yji/pa time-varying spin rates, and different axisymmetric geometries. Steady-state or transient time calculations can be Slip and/or transitional flow is en- determined for slip-flow or free- countered when the mean-free path ap- moiecular aeroheating. Results pre- proaches the boundary layer thickness. sented herein typify the predictions of At this condition the boundary layer is the heat-transfer code. diffuse, the flow appears to be "slipping" along the rotating surface, and the gas THEORY experiences a temperature discontinuity at the gas-surface interface. Figure 1 shows a typical flywheel geometry examined in this study. Under Free-molecular flow is reached standard conditions of pressure and tem- when the mean-free path is large. For perature, a> continuum boundary layer this rarefied flow, momentum and energy flow exists adjacent to the flywheel sur- exchange occur strictly from molecular- face as it is spun. Schlichting3 has surface interactions, and the incident
100 flow is undisturbed because of the pres- flux of energy is deduced from the ac- ence of the rotating surface. commodation property of the heated surface (determined experimentally and Figure 2 depicts the subdivision of related to the rate if the molecules were the flow regimes gauged according to the reemitted with a Maxwellian velocity ratio 6/X. Under conditions suggested distribution at the heated surface tem- for practical application (~10-2 to 10-5 perature). Typical values for the accom- torr), the noncontinuum flow regimes modation coefficient are a = 0.85 to 1.0. are of most importance and the heat The heat flux into the flywheel is then transfer associated with free-molecular represented in terms of a modified and slip/transitional flow follows in the thermal recovery factor, r\ and a next sections. Stanton number. Sj:
•20,000 rpm and
CONTINUUM FLOW
1 The expressions for r and St are given in the Appendix. s I SLIP/TRANSITIONAL HEAT TRANSFER Within the pressure range p = 10" * to 10"3 torr, the effects of intermolecu- lar collisions begin to be important. Contrary to free-molecular flow, the gap 1 rfr between the housing and the flywheel is important since a conduction path is pro- vided to the surrounding ambient tem- perature.
2 3 101 lO id io< 105 W6 To treat the heat transfer for slip GAS PRESSURE (lorrl flow, Couette flow between two parallel moving surfaces of different tempera- Fig. 2. Flow regimes at various ambient tures4 is examined (refer to Fig. 3). pressures and spin conditions. The following heat-transfer relationship incorporates the heat-conduction effect with the viscous dissipation in the flow: FREE-MOLECULAR AEROHEATING The convective heat transfer to a rotating flywheel in the free-molecular C 2 range is determined by the influx of (1+2X/G) G energy transmitted by the bombardment of molecules with the surface and is cor- Derivation of the above equation requires rected for the energy that is reemitted that the accommodation coefficient, a, into the incident stream. This reemitted and the momentum transfer coefficient, c, are unity.
101 COMPUTER MODELING
The previous heating relationships were incorporated in a con^-wter code that analyzes the steady-state or tran- sient temperature field within axisym- metric flywheel geometries. Options to the code include the allowances of varied geometry (including variable thickness), composite materials, non- steady spin rates, flow regime (pure Flywheel rotor slip flow or pure free-molecular flows) with variable gas properties, excess heating at the rotor tip, and nonisotropic temperature-dependent materials. The Fig. 3. Flywheel-housing geometry used results that follow in the next section in the slip-transitional flow aeroheating typify the predictions of the case. relationship. RESULTS
The first series of results are the THERMAL RADIATION CONSIDERATION equilibrium predictions of maximum temperature for a flywheel composed The radiant interchange between the of a material of low thermal conductivity flywheel and housing is of major impor- (an adiabatic surface results from a tance in the heat-transfer modeling be- zero-valued thermal conductivity). Sur- cause it equilibrates the system to tem- face temperatures are determined by peratures much lower than the recovery matching the convective aeroheating to temperatures of the flow. Without the the radiation cooling. radiation cooling, the steady-state tem- peratures would approach the recovery Displayed in Fig. 4 are these temperatures that are typically 1000°F. maximum temperatures at various angular speeds at the tip of a flywheel The radiant exchange between gray having a radius of 1 ft and subjected to surfaces5 is given as free-molecular aeroheating. The heating effects are small below a gas pressure of 10~5 torr. For an air environment at a pressure 10~3 torr, the free-molecular calculations predict a rise in temperature of 50 °F above ambient at a speed between W 40, 000 to 50, 000 rpm. For other radii, the tip-speed velocity should remain the where e w» £ H are respectively the same, with the angular speed adjusted wheel and housing emissivity and a is accordingly. the Stefan-Boltzmann constant. A modificftion of the above results Note: To maximize radiation cooling, a is necessary for aeroheating surfaces not high emissivity of the housing is re- parallel with the incident flow. Some quired. existing flywheel designs incorporate surfaces oriented 90° to the incident flow. For example, Fig. 5 depicts a portion of a flywheel that has an outer wrap as a supporting spoke. This effect of surface orientation in the aeroheating is displayed in Fig. 6.
102 700. 1000 9-90° (Air) 500 Air envlroment 600. Flywheel radius • 1 ' 'p-10'1 torr a "1.0; «"1.0 rr ADIABATIC SURFACE - HO. | »"0° (He)
§400. 50 Flywheel radius -1 ft • 40.000 rpm
§300. I 200. - io4 io3 AMBIENT PRESSURE (torr) 100. . Fig. 6. Maximum temperature at various ambient pressure and flow incidence angles for environments of 30.000 40,000 50.000 60,000 70.000 80,000 air or helium. ANGULAR SPEED (rpm) Also included in the same figure Fig. 4. Equilibrium surface tempera- is the effect of introducing an environ- tures at various angular speeds and ment of a low molecular weight; i. e., ambient air pressures. the inert gas helium. Hydrogen is per- haps the best environment from a molec- ular weight point of view; however, its highly reactive nature may cause prob- lems. The next section of results (Figs. 01 8 to 12) includes the effects of thermal i- conductivity in a flywheel geometry represented in Fig. 7. Thermal prop- \ erties of a Kevlar-49/epoxy composite spoke are used in the calculation and are given in the Appendix. \ Fiber orientation plays an impor- tant role in this modeling since the thermal conductivity changes by an order of magnitude with direction. For ex- Incident flow ample, circumferentially orienting the fibers places low thermal conductivity in the r-z directions, and the tempera- ture fields are well-represented by the Pig. 5. A flywheel design that incor- adiabatic predictions of the previous porates a spoke surface oriented 90° section. The higher thermal conductiv- to the incident flow. ity is in the fiber direction, and its effect to.relax the temperature fields is
103 convective heating load corresponding 30,000 rpm to a 90° angle of incidence to the flow. This modeling approximates the tem- perature field near a supporting spoke. The surface temperatures with and with- out excess heating are shown in Fig. 9.
220 - -
200 - tz. i Fig. 7. The flywheel geometry used in Excess heating at outer radius (k • this thermal analysis. E 180 - UT I insignificant since there are no thermal CO in / ^y gradients in the azimuthal direction due -u 160 " Excess heating at outer radius (k -1.0: k-0.11 to symmetry. However, a radial orien- a* tation of the composite fibers imposes a —— 1> /' higher radial thermal conductivity and 1 140 - the radial distributions of temperature y
become flattened. Figure 8 compares ; 120 - the surface distributions for various kr __.--*" NO excess heating (k r-l. 0; I^-0.1) and kz values. Linear profiles are pre- dicted, and the profiles cross at a com- 100 - mon location that divides the region at which the aeroheating dominates the 0.9 1.0 radiation cooling, and vice versa. NORMALIZED RADIUS r/r
Fig. 9. Radial temperature profiles for k
104 cross-sectional area exists to conduct Finally, a slip-flow calculation is the heat radially; consequently, the tem- shown in Fig. 12. Maximum surface perature fields are higher. temperatures at various housing/flywheel gaps are depicted for a Kevlar wheel spinning at 40, 000 rpm within an air en- vironment at p = 10~3 torr. Gap dis- tances have to be lower than the mean- free path to provide an effective con- duction path to the housing, assumed to
P-O.OI / / /, P- 0.001 TOHft be maintained at 70°F. Also, as the INOEXUSS HEATING -~V / • I™ (AUCMENIEO HEATING AT AT OUTER SURFACE! J / / OUTER SURFACE! distance becomes smaller, the heating contribution due to viscous dissipation is reduced. These calculations serve 0.001 IORR INOFXCESS HEATING AT OUTER SURFACE! as a lower bound to the temperature fields since a slip flow is not absolutely defined at this pressure.
Fig. 10. Transient temperatures at various pressures with or without excess /_.. -- heating at the rotor tip. AdWwttct Hftce P- in"' (orr wCOOO rvn / stipaow HlllimMMv
100
Fig. 12. Maximum steady-state tem- peratures for slip/transitional aero- heating demonstrating the effect of the housing flywheel gap spacing.
CONCLUSIONS
A computer program has been gen- erated for predicting flywheel tempera- ture subjected to free-molecular or slip-
>qccmvl iqrad Fig. 11. The effect of a change in The following conclusions have geometry on the radial surface been demonstrated by the code: distributions cf temperature. 105 1. For free-molecular aeroheat- ing, gas pressures from 10~3 s = exp (s sin e)2 to 10~5 are necessary to avoid t ijm j (- ) steady-state temperatures >50°F above ambient. + sfff (S sin 6) [l + erf (S sin 9» 2. A gas of lower density produces reduced heating effects. where 3. Protruding edges can cause S = rwvfy/2a temperature-related problems. THERMAL PROPERTIES OF KEVLAR 4. Increasing the thermal conduc- 49/E POXY COMPOSITE tivity in the radial direction relaxes the thermal field and The calculations of this study use lowers maximum temperatures. the following thermal properties: 5. Time to reach steady state is Fiber direction typical of ~1 hr. 6. For slip-ilow calculations, the k = 1. 39 + 1. 11 x 10"3 T + 0. 442 x 10~6T2 temperature field is decreased by reducing the housing/flywheel gap distance. (Btu/ft/hr/°R) 7. Since thermal radiation plays a Perpendicular to fibers major role in the heat transfer, a nigh housing emissivity is k = 0.074 + 0.22 x 10"3T + 0.023 x 10"6T2 desirable. APPENDIX (Btu/ft/hr/°F) (T is in °R) FREE-MOLECULAR HEATING RELATIONSHIPS p = 86. 4 lb/ft3; C = 0. 3 Btu/lb/°R; P The heat-transfer parameters r' e = 0. 8; a = 1.0 and St have been derived by A. K. Oppenheim7 for a flat plate inclined at REFERENCES an angle 8 to the incidence flow (negative angle yield back-surface heating). These 1. J. A. Kirk, "Flywheel Energy are given as: Storage - I and II. " Int. of J. Mech. Science, 19, pp. 223-245, 1977. 2S2 t - 2. J. A. Rinde, Fiber-Composite Fly- wheel Program-Quarterly Progress Report, January-March 1977, Lawrence Livermore Laboratory, UCRL-50033-77-1. 1 + 4n (S sin 9)( 1 + erf (S sin 9) exp (3 sin 9) ) 3. H. Schlichting, Boundary Layer Theory, 6th Ed., (McGraw-Hill Book Co., New York, 1968), pp. 213- 218. 106 REFERENCES (cont) 4. S. A. Schaaf and P. L. Chambre, Flow in Rarefied Gases, No. 8, Princeton University Press, Princeton, NJ, 1961. p. 38. 5. R. Stegel and J. R. Howell, Thermal Radiation in Heat Transfer, (McGraw-Hill Book Co., New York, 1972), p. 282. 6. M. R. Baer, Aerodynamic Heating of High-Speed Flywheels in Low- Density Environments, SAND78-0957, Sandia Laboratories, Albuquerque, NM., October 1978. 7. A. K. Oppenheim, "Generalized Theory of Convective Heat Transfer in a Free-Molecular Flow, " J. of Aero. Sci., 49, 49-58, January 1953. 107 SESSION III: FLYWHEELS 109 PROJECT SUMMARY Project Title: The Application of Flutd Film Bearings and a Passive Magnetic Suspension to Energy Storage Flywheels Prfncfpal Investigator: Martin W. Eusipi, Larry Martin, and Dr. Amit Ray Organization: Mechanical technology Incorporated 968 Albany-Shaker Road Latham, NY 12110 (518) 785-2211 Project Goals: The goal of this program was to establish realistic fluid film bearing performance parameters over a range of energy storage levels from 10 KW-hrs. to 100 KW-hrs. A passive magnetic lift was to be incorporated in the suspension to minimize the size of the fluid film thrust bearing. Project Status: A final report including detailed parts drawings of a bearing test model was included in the program goals. Contract Number: 076997 Contract Period: Jan. 1978 - Sept. 1978 Funding Level: $55,000 Funding Source: Sandia Laboratories, Albuquerque 111 THE APPLICATION OF FLUID FILM BEARINGS AND A PASSIVE MAGNETIC SUSPENSION TO ENERGY STORAGE FLYWHEELS by Martin W. Eusepi, Larry Martin, and Dr. Amit Ray of Mechanical Technology Incorporated 968 Albany-Shaker Road Latham, New York 12110 518/785-2211 ABSTRACT This paper presents the results of an investigation to establish realistic fluid film bearing parameters and a passive magnet suspension system for the support of energy storage flywheels in the range of 10 kW-hr to 100 kW-hr. A specific design for a three lobe preloaded journal bearing and a shrouded step thrust bearing was established for a 10 kW-hr flywheel which weighs 1600 lb and has a maximum rotation speed of 12,500 rpm. Equations are presented to assist in establishing the bearing design requirements for other flywheel sizes. The magnetic suspension system designed to support 90 percent of the flywheel weight operates in an attractive mode and incorporates a novel configuration which provides a positive stiffness slope for the attractive magnet assembly. INTRODUCTION The principal performance factors governing the flywheel suspension design were: power Point of consumption energy storage loss; magnet support capability; magnetic can produce worthwhile economies. The use support geometry; fluid film bearing geo- of flywheels as the storage means is at- metry; lubricant selection and distribution tractive but is limited by the continuous system. energy drain of frictional losses. This report will present the results of a study SUSPENSION CONCEPT SELECTION that established realistic fluid film and magnetic bearing performance parameters, FLYWHEEL SIZING over a range of sizes of energy storage flywheels from 10 kW-hr to 100 kW-hr. At Prior to establishing the final con- the 10 kW-hr size, a flywheel weight of figurations for both the magnetic suspen- 1600 lb (30 in. dia and 6 in. thick) and sion and fluid film bearing designs, it is a maximum speed of 12,500 rpm, in a verti- necessary to determine the range of support cal orientation, are representative condi- needed, as the storage capacity of a fly- tions. A vacuum environment of 10"3 Torr wheel is varied "over the range 10 kW-hr < and an external ambient temperature range E < 100 kW-hr. For any flywheel, the of 60°F to 100"F are assumed for design stored energy level is governed by the pro- purposes. portionality relationship: E % R^u^t; where R = radius, u> = rotation frequency Long life and reliability, coupled (sec"1), t = thickness. In addition, the with minimum power loss, are the design disk stress can be expressed as: a •*• R2&)2 goals. Power consumption is crucial, and the weight by: W i. R^t. since it represents a continuous drain in the stored energy. In order to minimize If the allowable stress level is main- power consumption, a passive magnetic lift, tained constant for all flywheels, then the supporting up to 90 percent of the flywheel product of rotational speed and diameter we ight and utilizing no direct electrical (DN) is a constant. From this assumption is incorporated in the design. and keeping the R/t ratio constant, it can 112- be shown that the following relationships At C/R - 1.0 x 10~3 and W/DL - 500 exist: or 250 psi, the load equation reduces to • Flywheel weight; H - W [|_] Wx - 1.62 @ 250 psi W2 - 3.24 @ 500 psi. ° o 1/3 • Flywheel Diameter; D - Do [f-] A total radial load of 60 lb, 14 lb o from a 1° tilt and 46 lb from a 12 x 10"6 1/3 • Flywheel Thickness; t • to l|-) in. unbalance has been chosen as a reason- able design value. The calculated bear- • Maximum Flywheel Rotational Speed; N • t) l-jrl1'3 Q ing data is presented in Table 1. Calculations of flywheel size and The extension of the performance pre- weight are based on a 10 kW-hr flywheel dictions listed in Table 1 to other fly- with the following characteristics: wheel sizes is performed as follows. W 1600 lb D = 30 o = o in. From the bearing's dimensionless load, 6 in. N 500 rpm power and flow rate equations and the fly- o o - 12, wheel extension equations, the bearing per- formance parameters can be reduced (for JOURNAL BEARING SIZING W/DL = constant) to: As a means for generating bearing Load Parameters W = W [^H loss data prior to the selection of a o o final flywheel bearing configuration, a clearance ratio (radial clearance of the journal bearing (C) to the bearing radius Power Loss H = HQ [-|-] [=-]; (R)) of C/R = 1.0 x 10"3 and a room temp- o o erature lubricant viscosity of Z = 5 centi- poise was selected. This viscosity is Flow Rate Q = Q [-^-] [-^]. o E -^ characteristic of commercially available o Q gyro-lubricants having a vapor pressure o of approximately 0.001 mm Hg at 181°F. The calculated performance of the The bearing calculations are based on a elliptical bearing extended over the fly- loading pressure of 250 psi and 500 psi wheel energy storage range 10 kW-hr < E < for the 10 kW-hr flywheel size. 100 kW-hr is based on the above equations. The quantities W, T, and Q may be different For an elliptical bearing , the di- for different bearing designs and cannot mensionless load parameter is given by be related directly to a change in flywheel size; however, for detailed design, a de- W = w (C/R)' Viscosity, signer must be aware that specific bearing MNLD lb-sec/in.2 geometries will produce somewhat different L Length, in. normalized results requiring less general D Diameter, in. calculations. The power loss is determined from THRUST BEARING SIZING the equation 3 2 The frictional power loss generated M — HP = 0.0042 [ . 3 T, and its flow by a thrust bearing is expressed as: c 2-2 rate from Q = 7 NDLCQ where f = fric- r where tion coefficient. ' V - viscosity; A Table 1. Journal Bearing Dimensional Performance Data; 10 kW-hr Flywheel. Bearing Length and Maximum Required Operating Design Pressure Diameter Safe Load Power Loss Lubricant Flow Film Thickness psi in. lb Watts in.3/sec 10"3 in. 3 250 .49 210 29 6.3 x 10r 0.25 500 .34 105 15 3.6 x 10' 0.17 113 active area; h = nominal film thickness, MAGNETIC SUSPENSION SIZING u = rotational frequency; r = mean radius. Thrust bearing area is established by Magnetic suspension design became a setting the nominal design average pressure significantly important study when it be- to: pave = 800 psi and by establishing came apparent that a magnetic lift was the criteria that the thrust bearing in- essential to minimize the size of the fluid ner area diameter will be one-half its film bearings. Both passive or active mag- outer diameter. In addition, since the netic suspension can be used for flywheel magnetic suspension is to provide a sig- load support. nificant portion of the gravity thrust load, the fluid film thrust bearing load The electrical techniques for servoing can be expressed as W^ = aWf. Therefore, magnetic bearings have reached the level the bearing load is reduced by a and the of sophistication that it is possible to 2 bearing diameter by [a]-*-/ . obtain magnetic suspension actively. The greater the ability to maintain passive The thrust bearing power loss, as a forces, the lesser the dependence on servo- function of energy level to be stored in ing and the associated electrical equipment the flywheels cannot be calculated until power requirements and propensity for in- the operating film thickness is establish- stability or non-reliability. Therefore, ed. The calculation of film thickness is the present study considers only the pas- based on Fuller2 where an approximate film sive magnetic lift where no servo technique thickness for a hydrodynamic thrust bear- will be utilized. ing is given by the relationship: 1/2 As a part of this study, several types of magnetic support bearings for the group , where: of flywheels in the energy storage range ave of 10 kW-hr < E < 100 kW-hr were consider- I = mean bearing length; ed; these were: conventional attractive n = pad size factor; taken as 0.44; design; repulsion design; new attractive k =hydrodynamic factor; taken as 0.04; axially stable design. u = mean velocity. Conventional attractive magnetic bear- The smallest thrust bearing (10 kW- ing designs all show axial instability or hr flywheel @ a = 0.1) with a mean radius, a negative spring rate. This bearing type, r = 0.69 in. at one-third its maximum therefore, was considered and subsequently speed of 12,500 rpm has a mean velocity discarded for this application. However, of u = 301 in./sec, a mean bearing the repulsion type magnetic suspension of I = 0.434 in., and a calculated minim; ich results from the installation of like film thickness of h = 1.2 x 10"4 in. magnetic poles across a bearing gap was analyzed and solved for high coercivity The bearing load, film thickness and rare earth permanent magnets. power loss equations can be used to cal- culate bearing power loss as a function of flywheel size. To calculate the effect of The following design equation, result- ing from this analysis is (in the MRS sys- speed variation on power loss, the effect tem) used to calculate The 10 kW-hr size flywheel with a weight of 1600 lb requires a magnet bear- The results of the parametric thrust ing support of 1440 lb. If a magnet width bearing power loss study show that unless of 1 in. is selected, then the total force fluid-film bearing loads are substantially equation reduces to: EFZ = 70 n and the reduced below the actual flywheel weight, total number of magnets required is n = 92. excessive bearing losses are experienced. The total magnet weight required to sup- As an example, at full speed a 1600 lb port 1400 lb is 54.28 lb. thrust bearing load would produce a 3500 watt loss while a 160 lb load produces Since the number of magnets is direct- only a 6 watt loss. ly proportional to the required support 114 M CURVE ES 1 00 2 aos 5 0.10 N S 4 0.20 S aso e 1.00 «; 7 2flO 1 1 (a) L F> S I 2 1.00 (b) Fig. 1. Axial separation force, Fig. 2. A simple axially stable repulsion magnet. attraction magnetic bearing. force, which in turn is directly propor- member which is fabricated from a soft mag- tional to the energy storage level, then netic material. When assembled, they form the relationship ,E. can be used a bearing configuration with each pole nl = no 115 system. The decrease in axial force, with Table 2. Magnetic Bearing Dimensions for the decrease in axial distance, can be a 10 kW-hr Flywheel. accentuated furthermore by saturating the moving members, thereby increasing the Parameter Ferrite Samarium-Cobalt Units axial stiffness. This concept is selected 3 • Flux Density 0.2000 0.4250 Telst o for optimization since it presents a posi- H " Magnetic Field Strength 150794 337302 At/n tive stiffness in the attractive mode. Mean aadius r 30.48 30.48 cm Because of the circular symmetry, the Lengch lg 0.508 0.508 fflffl bearings will be neutral in the circumfer- Gap i;g" for open position 1.524 1.524 am ential direction; in the radial direction, Gap "g" for closed position 0.508 0.508 tan it will also be neutral because the total 1.504 1.366 cm flux remains unchanged for any radial per- : turbation, due to saturation in the soft Lra 0.851 0.389 cm magnetic pieces. C 4.887 2.093 cm Vertical Force for Open Position 725.0 725.0 kg FINAL MAGNETIC SUSPENSION DESIGN Vertical Force for Closed Position 532.0 516.,; kg The permanent magnet assign problem Total Weight of Magnet 3.83 1.27 ke is, in general, to specify the permanent magnet materials in terms of defined unit properties and to arrive at a volume for a axial direction. Stationary magnetic given configuration which most efficiently pieces are made of 65 permalloy radial establishes the total field energy required stampings separated by insulating material in the given space. The bearing calcula- of equal thickness. The radial laminations tions were performed by assuming a flux with alternating nonconductive and non- path and then calculating the permeance in ferrous shims, serve the following func- order to determine the air-gap flux. The tions: Circumferential flux flow is pre- total permeance establishes the point of vented if the shaft is off vertical and operation on the demagnetization curve. the uniform gap is not maintained, thereby It is essential that the permeance be eval- reducing the non-uniform air gap flux uated very accurately in order to ensure densities and, hence, the unbalanced mag- that the magnet operates at the optimum netic pull; eddy currents, which might be point in the demagnetization curve for generated due to the non-homogeneity in miniaium weight. The optimum point of op- the flywheel and the insert materials, are eration is defined as the point in the de- prevented. The soft magnetic rotor fabri- magnetization curve where the product BmHm, cated front a solid piece of supermalloy. or available magnetic energy, is maximum. For the ferrite magnet, Bm = 0.2 Tesla and H,,, = 150794 AT/m at the optimum point of operation. CALCULATION RESULTS The magnetic bearing dimensions are computed for a 10 kW-hr energy storage fly- wheel. The total weight of the flywheel is about 725 kg (1600 lb) and the maximum speed is 12,500 rpm. A large percentage, but Jess than the total weight of the fly- wheel, has to be supported by the magnetic bearing. Computations were performed for both ferrite and Samarium-Cobalt magnets, and are presented in Table 2. Material used for soft magnetic pieces is 65 per- molloy in the stator and Supermalloy, in the rotating member. Z « .6 .0 1.0 *.2 1.4 1.6 l.« The variation of axial force with air gap is shown in Figure 3. If the magnetic Fig. 3 Axial force versus lift force is set at any value between the airgap "g" closed and open positions, the magnetic bear- ing will provide a stable system in the 116 FLUID FILM BEARING DESIGN The entire flywheel assembly will op- erate in a substantial vacuum, so that The design for the journal and thrust windage losses will be negligible. The bearings of a 10 kW-hr magnetically sup- rotor is modeled, as shown in Table 3 by ported flywheel includes the requirement eleven stations. This model is succes- for low power loss and excellent rotor- sively used for all critical speed, syn- dynamic performance. chronous response and stability (subsyn- chronous whirl) studies. Since the flywheel studied operates vertically, special considerations are im- The mathematical model given in Table posed on journal bearing selection, partic- 3 may be used directly in the critical ularly with respect to subsynchronous whirl. speed computer program. It is necessary, Figure 4 schematically illustrates the however, to compute an estimated stiffness vertical flywheel-bearing system. of the pintle supports so that their effects may be included on the critical The journal bearings are represented speed map. by short cylinders, or "pintles," which are designed with sufficient flexibility PINTLE FLEXIBILITY CALCULATION to provide necessary damping via a radial damping device. A fixed-pad, tapered land The pintles are treated as cantilever thrust bearing is provided at the lower pin- beams; the stiffness, due to bending and tle for support of axial load. shear deflection, can readily be estimated and combined to yield an "effective" stiff- ness. Assuming a simple cantilever beam, the bending and shear deflections for steel 5 are calculated to be K^ d _ 2.76 x 10 lb/in. t SCMWI 6 «' UMTHK) and Kshear = 2.084 x 10 lb/j.n., resulting J in an effective stiffness: Keff = 2.44 x 105 lb/in. 30'0U. Although the stiffness of the pintles have been used consistently in the rotor- dynamics calculations, the geometry assumed «"TWCX may change: for example, the length of the pintles may have to be increased due to space requirements for the thrust bearing NSSWASCN and damper. The stiffness, however, can MHOCT USEMU [FCH03CNI be fixed by providing the same effective lateral stiffness. This would entail, perhaps, boring out the pintles to a speci- THUSTBUKINS «N0 OUKR fied diameter or adjusting its active length. Fig. 4. Schematic of flywheel- bearing system Table 3. Mathematical Model of Flywheel-Bearing System. ROTOR DATA STAT MASS IP IT LENGTH STIFF. MASS INNER YOUNGS HOD. OENSITV SHEAR MOO. NO. (L8S> (LB-!N«2» (INI OIA. DIA. OIA. as/IN<"»2> ) form of a thin film of lubricant, similar to a journal bearing except that rotation K_ • *G.T4E3-T5J- does not occur. The damping forces in the f*mm * I2S00 WW pedestal will limit synchronous whirl of the flywheel-bearing system. From short bearing theory, and for eccentricity e =. 0, and L/D ~ 0, the damp- ing is calculated from: IT ,L. 3 . 5 "radial = 2 (C} (Full Film). The damping can be supplied, as an example, by a single ring. 2 9 Iff 2 8 10° Z 9 A damping value of 200 lb sec/in, BEMtO STIFFNESS, ft/In FUIT HOMES CONSTANT FED STIFFNESS OF 2 4 E SIb/h' obtained from a 2 in. dia x 2 in. wide x Fig. 5. Critical speed map; 10 kW-hr .001 thick oil film was used in the sta- flywheel; attractive magnetic bility and synchronous response calcula- tions as a first approximation for the bearing; flexible pedestals. required level of damping. JOURNAL BEARING SELECTION DYNAMIC BEARING LOADS Because of its excellent stability A vertical rotor can have bearing characteristics, the preloaded three lobe loads resulting from misaligning the plane bearing, as shown in Figure 6 was con- of the disk flywheel with respect to the sidered in place of the elliptical bearing spin axis. It can be shown that the dy- as the design for use on the flywheel namic load on each bearing (for a symmetri- rotor. Reference [3] provides most of the cal rotor) is given by: necessary information from which the steady-state and dynamic characteristics sinif> may be determined. R This design is appealing since no and for a thin disk and small angle ij> "tilting" is required to load the bearing; (radians), this equation reduces to the geometry provides the load. Thus, the I $ flywheel-bearing system may be operated R * such that the static rotor centerline is truly vertical. 118 10* 10* There is some uncertainty, however, that JOURNAL the geometry of the lowest loss case will RADIAL CLEARANCE yield a stable system and/or show a satis- C~ factory synchronous response. Thus, addi- tional parametric studies are required. DYNAMIC BEARING COEFFICIENTS Reference [3] provides tables of dimensionless stiffness and dampness coef- ficients for various preloaded three lobe bearings operating in vertical rotors. Also, we have K = K._. = Radial Stiffness; xx yy K - K = Cross-Coupling Stiffness; yx xy Radial Damping; B, - B xx yy yx xy = Cross-Coupling Damping. Actual (dimensional) values of stiffness and damping are computed from the following equations GROOVE (|)2 uNLD K typical Fig. 6. Schematic of preloaded three lobe bearing (|)2 pNLD typical fi«ference [3] a]30 provides dimen- when the quantities in {} are obtained from sionless friction force coefficients as a Reference [3]. function of bearing geometry and preload. The dimensional friction force can be com- STABILITY ANALYSIS puted from (|) uNLD (£){"} = pNLD Damped natural frequencies and associ- (•jb W> where {} is the dimensionless ated log decrements were determined for a quantity obtained from the appropriate number of cases of preloaded, three-lobe table in Reference [3]. bearings. Figure 7 indicates the sharp de- crease in log decrement when the radial For u = 0.75 x 10~6 reyns; N = 12500/60 = 208.3 rev/sec; L = D = 0.50 in. a The horsepower loss per bearing may now be « 1 4 N computed from ! 0 1 \ 1 1 V. = 1 V Bearing 63025 ' NJ 1MS1MLE : i N^ j where R = journal radius in inches; N = 1. rotational speed in rpm. For N = 12500 rpm and R = 0.25 in., the hp/bearing = 0.0496 Ff. 3 LCet SS. m. 1/2 vonicju. noroR Since the horsepower loss is of ut- PED. K-«Z5OIB/«l most importance, we will naturally examine in the configuration which yields the lowest M> 0.31b loss. 0005 0007 .001 Cm,aEWNG UDULCUAIUNCE (IN) This design study indicates that the lowest power losses occur when the set-up Fig. 7. Log decrement S versus clearances are the largest. This is to be clearance Cm, three lobe expected for the vertical rotor. journal bearing. 119 clearance became larger than 0.0005 In. It mils occurred at bearing 1. This level should also be noted that flexible pedestals of unbalance is clearly unacceptable, were included for all runs, with K as it exceeds the steady-state clearance ped of 0.250 mils. However, for an unbalance 81250 lb/in. and Bped 200 lb sec/in. of 0.1 oz-in., the semi-amplitude at Assuming that the synchronous response bearing 1 can be reduced to 0.0252 mils is satisfactory, the final bearing design which represents a very acceptable level for the 10 kW-hr flywheel is a 1/2" x 1/2" (10 percent) of dynamic whirl. preloaded (m = 1/2) three-lobe bearing hav- THRPST BEARING DESIGN AND PERFORMANCE ing a machined radial clearance of 0.000500 in., a set-up radial clearance of 0.000250 Hydrodynamic fixed-pad thrust bearing in., and a power loss of approximately 21.6 calculations were made for the flywheel watts. rotor-bearing system with reference to the procedure given in Reference [4]. SYNCHRONOUS RESPONSE The steady-state load was assumed to Synchronous response calculations cor- responding to both static and dynamic un- be 10 percent of the total flywheel weight, balance were made using the dynamic bearing 160 lb. Calculations were based upon a coefficients determined for the three-lobed maximum shaft speed of 12500 rpm, a coef- bearing. ficient of viscosity of 0.75 x 10~ reyns, and a six-pad bearing. Response calculations for these bear- ings assumed unbalancss of 1 oz-in. at Results of three cases are summarized all planes; results of this analysis are below shown in Figure 8 for both in-phase and (In.) h. •Hi) Alt tomr LOM tacta) out-of-phase unbalance. The major semi- O.D. (in.) I.D. ,C •n amplitudes are plotted as a function of 1 1. 0 0. 50 0 .203 21.6 36.6 rotor speed. 2 1. 25 0..50 0.295 13.7 60.9 3 1. 50 P. 50 0.416 10.2 87 .8 For 1 oz-in of dynamic unbalance, h. — • flla thlckaua the maximum semi-amplitude of 0.252 H z 1 1 PEDESTAL m • 0 3 Ib k- 812501b /in. J b • 2001b «c/in. ItfMOOE BRG2 ^— - • — K) / 2nd MOD!• BRGI •— 1 • • CM - ^—— ^- t ( 1 0 1,000 3,000 5,000 7,000 9,000 11,000 13,000 15,000 ROTOR SPEED,RPM Fig, 8. Synchronous regppns.e three lobe bearings; (m = 1/2), 120 In order to minimize power loss, it In conclusion, the successful comple- appears that as small an OD as possible tion of the flywheel's suspension system de- should be selected. Manufacturing sign would indicate that extension of that considerations may, however, dictate a lar- design to a finished test vehicle would pro- ger than optimum outer diameter. In any vide the necessary tools to evaluate both event, the thrust bearing power loss rep- the basic designs and the scaling factors. resents a substantial and probably a major It is recommended, therefore, that a rig de- portion of the total fluid-film fricticnal sign be implemented and that construction loss in the flywheel bearing system, and testing of the suspension design, both for the magnetic and fluid film systems be SUMMARY undertaken. A viable suspension has been estab- REFERENCES lished to support a 1600-lb, 10 kW-hr en- ergy storage flywheel. The flywheel design 1. Badgley, R.H., "Standardization Manual, consists of a flat disk, vertically ori- Sliding Surface Bearings," MTI 69TR61, ented, with fluid film journal bearings at December 1969 (Revised June 1970). opposite ends of its axle. This concept permitted the positioning of a small thrust 2. Fuller, D.D., "Theory and Practice of bearing at the lower axle and a partial mag- Lubrication for Engineers," New York: netic suspension system to act directly on John Wiley & Sons, Inc., 1956. the wheel's lateral surface. 3. Lund, J., Rotor-Bearing Dynamics Design Based on analysis, fluid film journal Technology, Part VII: The Three Lobe bearings incorporating a three^lobe design Bearing and Floating Ring Bearing; Me- will provide adequate load capacity. When chanical Technology Incorporated, combined with the flexible pedestal con- February, 1968. cept employing an extended pintle and dash- pot, the bearings maintain sufficient rotor 4. Raimondi, A.A., and Boyd, J., "Applying response control. An evaluation of rotor Bearing Theory to the Analysis and Design response has established that required of Pad-Type Bearings"; Part I; ASME paper levels of rotor balance and orthogonality No. 53-A-8A. Df the rotating inertia mass to its spin axis are reasonable and achievable by present manufacturing techniques. A new method of designing magnetic bearings has been presented. The method uses a concept which makes an attraction system with soft magnetic pieces stable in the axial direction. The conventional at- traction system is only stable in the trans- verse directions and requires servo tech- niques for stability in the axial direction. The new design develops a passive attraction system which is stable in the axial direc- tion and neutral in the transverse direction. The means for extending the 10 kw fluid film bearing design to flywheels of other energy storage levels is also presented. The extension of magnetic suspension design to other wheel sizes is also explained. Be- cause of its complexity, the totordynamic analysis does not lend itself to parametric curves for extending performance to other inertia systems. The procedure for the dy- namic analysis must be employed after pre- liminary sizing is accomplished. 121 PROJECT SUMMARY Project Title: Development of an Advanced Flywheel Bearing Performance Model Principal Investigator: David B. Eisenhaure Organization: The Charles Stark Draper Laboratory, Inc. 555 Technology Square Cambridge, MA 02139 (617) 258-1421 Project Goals: Develop a performance prediction model for ball bearings based on experimental data combined with a theoretical model. The objective is to have available a design tool to assist in the optimization of the flywheel bearing set when other system parameters become avail- able. Project Status: Completed Contract Number: 07-6995 Contract Period: Oct. 6, 1977 - Sept. 30, 1978 Funding Level: $99,000 Funding Source: Sandia Laboratories, Albuquerque 123 LOW-LOSS BALL BEARINGS FOR FLYWHEEL APPLICATIONS David B. Eisenhaure and Edward P. Kingsbury The Charles Stark Draper Laboratory, Inc. 555 Technology Square Cambridge, Massachusetts 02139 ABSTRACT This paper summarizes work performed by The Charles Stark Draper Laboratory, Inc. during a Sandia Laboratories sponsored contract (No. 07-6996) for the analysis and test of ball bearings suitable for vehicular flywheel applications. Particular emphasis was placed on the CSDL full complement retainerless concept which prior research indicated had extremely low losses and long life at high speeds in a vacuum environment. The primary accomplishments during the first six months of this 12- month program were the procurement and modification of test bearings, the design and construction of a test fixture, and the development of a theoretical representation of the bearing losses. The primary accomplishments during the second half of the pro- gram were the development of a good data base for the test bearings and a ball bearing performance prediction model. This performance prediction model was formulated by combining the experimental data base with the theoretically derived model and allows the optimization of future ball bearing configurations for vehicular flywheel systems. Three sizes of standard bearings were modified to the full complement retainerless configuration. These modified bearings and one set of standard retainer bearings were tested during the contract. Friction losses were obtained for a variety of radial and axial loads and the results integrated into the performance prediction model. INTRODUCTION The Charles Stark Draper Laboratory The three primary characteristics of Inc. (CSDL) is currently under contract flywheels which make them particularly to Sandia Laboratories for the develop- well suited for this application are high- ment of an "advanced flywheel bearing power density, reliability, and long life. performance model" (Contract No. 07- The last two lead potentially to low life 6996). This development deals with cycle cost. Despite these advantages, bearings for vehicular applications and flywheels have several unique technical places particular emphasis on the CSDL problems associated with them which retainerless bearing concept. This must be addressed. These problems paper summarizes work performed during revolve primarily around increasing the this contract and outlines possible future overall efficiency of the flywheel to an research. The following paragraphs sum- acceptable level by reducing the aerody- marize the background of this problem, namic, bearing, and possible seal losses. as well as some of the problems asso- The aerodynamic losses at high flywheel ciated with flywheel bearings and the speeds require a relatively high vacuum solution to these problems. which in turn demands either a low-loss bearing that will operate in a vacuum It has been generally recognized environment or rotating seals which that a high-efficiency energy-storage characteristically have substantial losses element will be necessary for the con- associated with them, especially at high struction of effective hybrid and electric- speeds. The bearing design problem is al vehicles. Flywheel energy storage complicated by the reaction torques of has many characteristics which make it the rotating wheel during vehicle dynam- particularly well suited for this applica- ics. These torques, without careful tion. This flywheel storage element system design, can create a hostile would be used as a storage node for re- bearing environment. generative braking energy and might be utilized for other power averaging func- The bearing requirements for a tions such as hill climbing, acceleration, vehicular energy-storage flywheel must or other high-powered modes of operation. include: stability, reliability, fail-safe 124 - I operation, low loss at high speed, temp- a) Four sets of angular contact ball erature insensitivity, and substantial bearings of various sizes were capacity for combined and shock loads at acquired. Three sets were modified zero speed, at high speed, and in vacuum. to the full complement retainerless Advanced retainerless ball bearings rep- configuration while one mid-size set resent an ideal candidate for this appli- was not modified. cation. These advanced bearings are characterized by long life and low power b) An appropriate test fixture was des- dissipation, and are well suited for oper- igned and constructed. ation in a vacuum environment. These c) An analytical model for the bearings bearings consist of standard commercially was formulated and a corresponding available ball bearings, suitably modified computer code generated. for operation without the conventional ball retainer (also called the cage or ball The following work was accomplished dur- separator). In conventional ball bearings, ing the second half of the program: the retainer is subject to violent instabil- ities of various types which are partic- a) A series of tests was conducted ularly troublesome in vacuum operation on each of the four bearing sets to and typically treble the driving torque obtain the bearing losses as a func- demand of the bearing. This problem is tion of speed under various axial and eliminated with retainerless operation as: radial loads. b) A detailed performance prediction a) There is no possibility of model for retainerless ball bearings retainer instability. was developed by combining the b) An optimum oil supply and scav- experimental results with the theo- enging system can be provided to retical model. A more limited per- to ensure adequate EHD lubricant formance prediction model for stand- film thickness in all the ball race ard ball bearings also was developed. contracts, without introducing c) A final report is currently being pre- excess oil and unnecessary vis- pared presenting the results of the cous drag. research and recommendations for The results are long reliable life and further research. minimum friction loss. There are no known disadvantages resulting from re- Test Bearing Construction tainerless operations. A prior CSDL program* has produced prototype retain- Angular contact bearings of standard erless bearings designed to replace the inertial geometry in three sizes, spanning spin axis bearings in the Skylab Control the range of interest, were procured for Moment Gyroscopes which had suffered this program, as shown in Table 1. failures in space triggered by retainer instability. These retainerless bearings have successfully passed every perfor- The outer race dams on these bearings mance test during their development, and were ground away to eliminate ball damage a pair is currently on life test at more in the assembly process. They are all than 8000 hours in vacuum, without diffi- thus "fall aparts" in the full complement culty. configuration. DEVELOPMENT EFFORT Since no long-term running is contem- plated in these tests (the driving torques to Summary of Work to Date be measured in the near-term steady state after run up), no provision for an oil sup- The goal of this research program is ply was required. to develop a performance prediction model for ball bearings based on experi- Oil in controlled amounts was applied to mental data combined with a theoretical the balls by evaporation from a dilute solu- model. The object is to have a design tion in Freon before assembly. This tool available to assist in the optimization method has been perfected in other pro- of the flywheel bearing set when other grams and allows sufficient running time system parameters become available. to make the torque measurements without The following summarize efforts during any damage to the bearings. the first half of the program: 125 Table 1. Angular Contact Bearings Geometry (inches) 1 Fill Size Bore Width OD Ball Diameter Complement 100H 0.3937 0.3166 1.0236 0.1875 11 105H 1.2910 0.4740 1.8504 O.2bOO 17 108H 1.9385 0.5922 2.5772 0.3125 21 Bearing Loss Analytical Model The quantities required at each contact on each ball are: Absence of the ball retainer and of uncontrolled bulk lubricant in the full com- Normal load N lbf plement configuration eliminates four of Maximum hertz lbf/in max the energy sinks discussed for convention- pressure al ball bearings. There remain three Hertz major semi- a in. sinks to be considered in a theoretical formulation of a retainerless energy de- axis mand model: Hertz minor semi- b in. axis a) Shear losses in the EHD films due to linear velocity differences (slips) Contact angle B deg between the load-carrying surfaces. Ball orbit rate S rad/s b) Shear losses in the EHD films due Ball spin rate rad/s to angular velocity differences s (pivoting) at these contacts. Ball spin orientation deg angle c) Hysteresis losses in the elastic deformations of the metal parts as Individual contact losses are then calcula- they pass through the loaded con- ted according to the equations developed in tacts. the following sections and summed for the One additional sink, which is peculiar to whole bearing to get the theoretical energy retainerless operation, involves losses at demand. the ball-ball contacts. Hysteresis The initial version of the loss model involves only items b) and c). The ball- The hysteresis losses are estimated ball losses are expected to be small based on the work of Drutowski^, who because the ball-ball forces and contact found, for a hard steel ball rolling areas are known to be small compared to on a similar flat, that the rolling those at the ball-race contacts. The slip resistance force could be represented losses are not accounted for initially be- by a power law as a function of normal cause slips are known to be small in the load. The loss equation is pure axial load cases. It is hoped to use a comparison of the axial and combined load results (both test and analytical) to estimate the overall importance of slips for the outer (o) and inner (i) contacts on for later iterations of the model. the jth ball. The hysteresis and pivoting losses Pivoting depend on specific conditions at each con- tact, which are, in general, different at The pivoting losses are estimated based each ball and also different at the inner on the work of Johnson5, who found that his and outer race contacts on the same ball. (and others) experimental traction results An existing computer program due to at high pressure could be explained as fol- Jones^ is used to calculate these individual lows: "at high pressure the (EDH) film contact conditions for the bearing sizes and exhibits a critical shear stress which is operating conditions of these tests. approximately proportional to the pressure 126 and decreases slightly with temperature, and determination of mounting details. The somewhat similar to a granular solid. development of a specially tailored lubri- This hypothesis leads to an approximately cation supply and scavenging system would constant 'coefficient of friction1 independ- also be required. Although the lubrication ent of film thickness, as would be observed problem has been successfully handled on in dry sliding or boundary friction. " This prior CSDL programs, it was not addressed is a great simplification for the purposes during the current program in order to of the present work, as it eliminates cal- direct the resources toward more extensive culation of the EHD film thicknesses. On testing. The technology generated during this basis, the pivoting torque at any con- this contract will allow an optimized des- tact is ign to be obtained in a very cost-effective manner. After the bearing is developed, (2) it should be tested extensively under antici- pated conditions of vehicle static and dy- where ri is the Johnson coefficient and the namic load. integral is taken over the contact area. The loss equation for the jth outer (inner) To determine the specifications of the contact is flywheel module bearings, one must start L 2 P fi with the vehicle and its operating environ- j p(o,i) = (3.17xlO~ ) j max(o,i) j (3) ment. The flywheel module would be (o,i) (watts) designed for use in some class of automo- where fi is the pivoting at that contact. bile (forinstance,a 3000-pound family car) in which the flywheel would be able to For the first iteration of the loss store the energy of a couple of stop/start model, it is assumed that the ball spin cycles. Past government studies on a vector is always normal to the line of con- similarly sized vehicle have used fly- tacts ( 127 the computer program would be the fly- wheel operating parameters and the bear- ing design specifications. The retainerless configuration showed considerably higher reliability during the test period together with somewhat lower losses when compared to retainer type bearings. The original premise of the pro- gram that this type bearing was ideally suited for high speed flywheel applications was confirmed. REFERENCES 1. Kingsbury, E., Evaluation of Alternate Bearing Designs for the Skylab GMG Final Report, Charles Stark Draper Laboratory Report R-1026, December 1976. 2. Townsend, D. P., C. W. Allen, and E. F. Zaretsky, "Study of Ball Bearing Torque Under Elastohydrodynamic Lubrication, " Trans ASME, JOLT, 1974. 3. Jones, A. B., "General Theory for Elastically Constrained Ball and Radial Roller Bearings Under Arbitrary Load and Speed Conditions, " Trans ASME, J. Basic Engineering, 1960. 4. Drutowski, R. C, "Energy Losses of Balls Rolling on Plates, " Trans ASMS, J. of Basic Engineering, 1959. 5. Townsend, A. and E. F. Zaretsky, "Elastohydrodynamic Lubrication of a Spinning Ball in a Non-Conforming Groove, " (Comments by R. L. Johnson), Trans ASME, JOLT, 1970. 128 PROJECT SUMMARY Project Title: Seals Evaluation for Advanced Flywheel Energy Storage Systems Principal Investigator: I. Anwar Organization: Franklin Research Center Division of the Franklin Institute 20th and Race Streets Philadelphia, PA 19103 (215) 448-1000 Project Goals: An evaluation of the problems associated with the use of rotating shaft seals in this application, along with a compilation of performance data for present state-of-the- art seals and suggestions for future developmental work. Project Status: The draft report has been approved by Sandia Laboratories and the final version of the publication is being produced. The report documents the performance of those seal concepts that are most promising for flywheel application. Particular attention is given to the so-called "dynamic" seals as being intrinsically well suited to the problem. It is concluded that a hybrid seal combining the screw-type and turbo-pump geometries offers the greatest prospect for future development. Contract Number: 07-7141 Contract Period: Jan. 1979 - Nov. 1978 Funding Level: $22,376 Funding Source: Sandia Laboratories, Albuquerque 129 SEAL STUDIES FOR ADVANCED FLYWHEEL SYSTEMS I. Anwar Frank)in Research Center Division of The Franklin Institute 20th and Race Streets Philadelphia, Pennsylvania 19103 ABSTRACT A seal application for an advanced flywheel system where housing pressure (vacuum) is equal to \0~l* torr is being investigated. Many conventional types of seals can be ruled out due to such extreme pressure conditions. For a dynamic sea! to work success- fully, it should develop a compression ratio of 106 magnitude during its normal opera- tion. In view of this, the principle of molecular vacuum pumps, where pressure ratios of this magnitude are obtainable, is being examined. It is believed that a fiywheel seal system based upon molecular vacuum pump principles on the vacuum side and upon viscous flow principles on the high-pressure side can be developed. For future work, a design study is recommended where the concepts of molecular drag and turbo pumps are combined with one of the conventional configurations of the viscous pumping seal. INTRODUCTION Since the seal is a part of the vacuum system, a definition of the overall system The present study involves a review of is required. At this stage one could vis- various types of seals with the purpose of ualize two basic systems. These are: identifying seal concepts that could be suc- cessfully applied to an advanced flywheel 1, A vacuum pump is employed to system. Two specific applications are con- lower the flywheel housing sidered. These are: pressure to \Q~h torr and then continues to operate at reduced capacity in order to maintain Automobile Peak Power the inside pressure within the Flywheel Flywheel prescribed limits. A system with this approach can tolerate Speed = 40,000(rpm) = 12,000(rpm) some leakage from the seal. Shaft dia. : 25.<» to = 76.2 to 2. A vacuum pump is used to lower 50.8(mm) 101.6(mm) the flywheel housing pressure 1 to 2(in.) 3 to Mln.) to 10"1* torr and is then shut off. In this case the seal Vacuum ^ 10 should have zero leakage at pre- scribed pressure levels and Wheel = 10(kW-h) should act as a pump to cope with capacity any small pressure changes in the housing. The basic criteria for the selection of The second system is, of course, more de- a seal are: small leakage, low energy con- sirable because of reduced energy require- sumption, and long useful service life. ments. The secondary criteria are: material, PERTINENT DEFINITIONS manufacturing cost, auxiliary requirement, failure mode, and experience. 1. Mean free path is defined as an average distance traveled by gas molecules be- tween collisions and is given by:1>2>3 130 - X = 2.331 x 10T20 L- (cm) where P = pressure in torr, T = abso- where y = absolute viscosity of gas; lute temp, in °K, and <5 = molecular to = angular velocity of moving surface; diameter in cm. Pa = ambient pressure; ft =_outside radius of bearing; c -nominal clear- 2. Gas flow is characterized by a para- ance of film gap. If A < 5 compressi- meter called Knudsen number de'ined as: bility effects may be neglected. In general, the compressibility effects N =i reduce the increase in pressure in gas K c bearings and seals. where A = free mean path, c = physical REVIEW OF SEALS chamber dimension Seals can be divided into two general N., < 0.01 Continuum Flow (viscous) K categories: static and dynamic. Static seals include 0 ring seals, metal dia- 0.01 < N,. < 1 Transition Flow (slip) phram type se:ils,and gaskets. For the present application, static seals are of no interest. Dynamic seals can be categor- NK>. Molecular Flow ized by their motions - rotary, oscillatory, and reciprocating. For flywheel applica- Since X is a function of pressure, tions, only rotary dynamic seals will be changes in pressure will cause changes considered. in flow regimes. Dynamic seals exist in many configura- tions and sizes and can be classified in 3. An accepted terminology for the degree numerous ways. However, their operation can of vacuum below atmospheric pressure is: be described in terms of a few fundamental principles. Zuk5 gives the following class- Low Vacuum 760 torr to 25 torr ification: Medium Vacuum 25 torr to 10-3 torr High Vacuum 10-3 torr to 10-6 torr 6 9 A. Positive (rubbing) contact: mech- Very High Vacuum 10- to 10- torr anical face, circumferential, lip, soft packing; k. Gas flow in vacuum technology is de- fined as: B. Close clearance: hydrodynatnic, hydrostatic, floating bushings; Q = SP C. Fixed geometry clearance seals: where S = speed (liter/sec), and P = fixed bushing, labyrinth; pressure (torr), Q= throughput (torr- 1iter/sec). D. Control of fluid properties: freeze, ferromagnetic; Conductance = ~ (liter/sec) E. Control of fluid forces: centri- fugal, screw pump, magnetic. 5. The ratio of velocity of flow to the velocity of sound is known as the Mach In view of the flywheel application, number. many types of dynamic seals can be elimin- ated. These include: positive rubbing types M = u/c hydrostatic fixed bushing, and labyrinth. Seals based on the principle of ferromagne- where u = velocity of flow, c = velo- tic, magnetic,and freezing phenomena are city of sound. not to be evaluated in the present study. If M < 1, the flow may be considered The seals that were found to be of in- incompressible. In theory of lubrica- terest for the flywheel application are the tion, compressibility or bearing number ones that have pumping capability. These is defined as:1* include: centrifugal, spiral groove, and viscoseal (screw). 131 Since the pressure is 10-1* torr on The theory of spiral groove bearings the vacuum side and 760 torr on the high is considerably developed,and the same re- pressure side, a seal in this application sults could be used for spiral face seals will operate in three different flow re- for preliminary analysis. Investigations gimes. These are: molecular, transition, cover viscous as well as transition regimes. and viscous. The review of seal litera- Since the compressibility number is small ture indicates that most work has been ( a 5 ), results of incompressible flow done in the viscous regime with the excep- can be used to estimate the performance. tion of a few studies which have been ex- tended in the transition regime. Therefore, Flat Thrust Bearing without Transverse the performance of these seals was evalu- (Radial) Flow for Viscous Conditions ated only for viscous flow conditions. Pressure Distribution: EVALUATION OF SEAL PERFORMANCE 3nurz The evaluation of seal performance of (I-X2)g ( Table 1. Performance of spiral groove bearing. (Without radial flow) outside dia. - 76.2 (mm) 3 (in.) inside dia. - 50.8 (mm) 2 (in.) axial clearance = 0.0025 (mm) 0.001 (in.) 5 2 li.M . = 1.l.88 xx 110"0 .((N sec/mc/m) air 9 2 ?.62 x 10- (lbf - sec/in ) Power Speed (increase Loss (rpm) in pressure) (watts) 12.4 x I03 (M/m2) k - number of grooves can be found in ient in the fluid annulus around a shaft 7 8 by means of a helical groove located either Ref. . . on the shaft or sleeve. Figure 2 shows the basic arrangement of the viscoseal. Sample calculations were carried out for the present application where a seal operates in viscous regime. Table 2 shows these results. Table 2. Performance of viscoseal (optimum geometry. 0 - 50.8 (mm) 2 (in.) L - 2.45 (cm) I (in.) c - 0.0025 (mm) 0.001 (in.) U . = 1.8 x 10"5 (N sec/m2) 31 r 2.62 x 10-9 (lbf - sec/in2) O'U-y)-r0tona Speed A for Power (rpm) Zero Flow Ap Loss (watts) a * Helix angle B « (h + c)/c 20,000 12 18.8x10*(N/m2) 5.95 Y - b/(a + b) 2.74 (psi) Fig. 2. Visco seal. 40,000 11.4 39 .2xlO 6UUL (4) For a seal to work successfully for the c2Ap flywheel application, it should develop com- pression ratios of the order of 106 during where JJ = viscosity, U = linear velocity, its normal operation. A preliminary review L - axial length, c - rad. clearance, and has indicated that compression ratios of Ap = increase in pressure that magnitude are achieved by molecular vacuum pumps. This led us to review the principles and constructions of various For viscous flow: types of molecular pumps. A"» f (a, B, Y» tan o, Rec) (5) MOLECULAR DRAG PUMP3.11 For transition flow* In 1912 Gaede introduced a type of A » f (a, B, Y» tan a, lO (6) mechanical pump which works on the princi- ple of imparting momentum to gas molecules 133 preferential)y in the direction of the Molecular Flow desired flow. In the molecular drag pump there is an open passage from the inlet to the outlet, between which a pressure dif- Pa _ S h ) (9) ferential is maintained by the high-velocity motion of one side of the passage relative 5 x IP'1* (rf to the housing of the pump in which the where K = inlet and outlet are located. 9.71 p - absolute viscosity of gas M = molecular weight of gas T = temperature For air at 20°C K = 1.62 x 10"5 Based on the simple design described above, Gaede built a multistage molecular vacuum pump where compression ratios of the order of I06 were achieved. The degree of vacuum produced is significantly affected (a) Principal of the molecular drag pump. by the speed of rotation and fore-pump pressure. Table 3 summarizes these effects for the Gaede pump.2 Table 3. Effect of speed of rotation on degree of vacuum obtained with Gaede molec- ular pump. a (b) Plane representation of the Speed of Rough-Pump Pressure on * molecular drag pump. Rotation Pressure, Fine Side, (rpm) P,,,m P Fig. 3. Gaede molecular pump. 12.000 0.05 0.0000003 In Fig. 3 the principle of the molecu- 12.000 1 0.000005 12.000 10 0.00003 lar-drag pump is illustrated. A cylindri- 12.000 20 0.0003 cal member rotates within a casing with 6.000 0.05 0.00002 a radial clearance h between them. At the 2.500 0.05 00003 top of the cylinder the clearance space is 8.200 0.1 Not measurable blocked by a projection of the cylinder 8.200 1 0.0COO2 wall which reduces the clearance locally 8.200 10 0.0005 to essentially zero. At either side of the 6.200 0.1 0.00001 projection the clearance passage opens into 6.200 1.0 0.00005 a closed volume. If there are no leaks 4,000 I.I 0.00003 in the system the total amount of gas in 4.000 1 0.0003 the system remains constant, but some gas is shifted by the motion of the rotor with mm Hg a reduction of the pressure Pj and increase in the pressure P2 . The equilibrium re- A number of alternative designs for lationship between P\ and P2 depends upon the molecular drag pumps have been devised the rotational velocity of the rotor and with two considerations In mind: upon the pressure regime in which the pump is operating, i.e., whether viscous 1. To ensure a low conductance leak- or molecular flow is involved in the pro- age path from outlet to inlet cess. through the running clearance of the pump. The following gives governing relations in two different regimes. 2. To vary the depth of the pumping channel in order to maintain the Viscous Flow molecular flow conditions over the compression range. P2 - Pj = (8) The following describes two such pumps. 134 Hoiweek Pump. Figure k shows the main ele- ment of the Holweck pump.12 The inner cylinder, A, is made of duralumin and is smooth. The housing, B, is made of bronze and has a spiral groove cut into its sur- face. The clearance between the cylinder and the housing is made not more than 0.05 mm (2 mils). Two spiral grooves are used, one right handed and the other left handed, and they meet at the inlet port C, which is at the center and connects with the high vacuum system. The.depth of the grooves increases from the ends where it is 0.5 mm (20 mils) to 5 mm (200 mils) at the center where thegrooves join and con- nect to the inlet port C. In operation the vacuum pump is first pumped by a suit- able forepump down to a pressure of a few microns. The molecular pump is then turned on, being backed by a forepump. The Rotating Housing Holweck pump operates at speeds of about disk 5000 rpm and produces pressure down to 10"6 torr. Fig. 5. Molecular drag pump of Siegbahn. Disk dia. = 5*»0 mm (21.26 in.) To vacuum system | To fore Vacuum Ball bearings All the pumps described above use air gaps of 0.02 to 0.05 mm maximum (1 or 2 Channel milt). Nonuniform thermal expansion, the presence of small foreign particles of the order of tenths of millimeter (a few mils) in diameter, or a sudden air shock can result in seizure of the rotor. A recent design intended to overcome this type of difficulty is described below. TURBO OR AXIAL FLOW MOLECULAR PUMP Induction motor Figure 6 shows the general arrangement Fig. k. Holweck molecular drag pump of a turbo pump which was first described 3 13 rotor dia. = 152 mm (6 in.) by Becker. ' Rotating disks all mounted on the central shaft are disposed alter- nately with stationary plates mounted in Siegbahn Pump.2'3 Figure 5 shows a cross- the housing. The disks and plates are cut section of the Siegbahn pump where pumping with slots set at an angle so that gas channels in the form of Archimedes' spirals molecules caught in the slots of the moving are cut in the two flat sides of the hous- disk are projected preferentially in the di- ing, within which a disk rotates at high rection of the slots in the stationary rotational velocity. The clearance between plates. The running clearances between the the disk surface and the flat section of rotating and stationary plates generally are the end plate between the adjacent spirals of the order of 1 mm (4o mils), which is an is made as small as practical for free order of magnitude greater than the permis- rotation. The inlet is at the periphery sible clearances in a conventional type of of the disk and the discharge at the hub. molecular pump. The rotational speed for a In the figure shown three spiral grooves pump having a rotor diameter of about 17 ~m (22 x 22 mm at inlet and 22 x 1 mm center) (6.5 in.) is 16,000 rpm, giving a peripheral are cut in parallel starting 120° apart, speed of 1.56 x 104 cm/sec (6.14 x 103 in/ providing three times the pumping speed of sec), about one-third average velocity for a single channel. air molecules at room temperature. 135 Vacuum System The observed dependence of the inlet Rotating \ stationary pressure Pj on the outlet pressure P2 is Disk \ f shown graphically in Fig. 7 for hydrogen, air, and the refrigerant Freon-12. A com- pression ratio P2/P1 of the order of 107 is obtained for air when P2 is equal to 0.1 torr, but the value drops off rapidly with increasing pressure to about 10 when P2 is equal to 1 torr. The compression ratio for hydrogen is significantly smaller for the To Fore same outlet pressure. Vacuum According to Becker, the turbomolecular pump is like a Gaede pump except that it operates with large clearance between the cooperating drag surfaces, and with these pumping drag surfaces fed by a series of in- clined slots in such a way as to greatly Gas improve the pumping speed over that of the Gaede pump. Kruger and Shapiro1**'15 produced a theory of turbomolecular pump performance based on the kinetic theory of gases. The basic concept of the theory, which is appli- cable in the free molecule range only, is the transmission probability; that is, the fraction of molecules entering a rotor or Rotor Disk Rotor Disk stator from one side which ultimately leave that same rotor or stator from the other X-Section of Disks side. The theory was validated by experi- ments on a single test rotor. The following Fig. 6. Turbomolecular pump, rotor summarizes some of their important conclu- dia. = 165 mm (6.5 in.) (Becker) sions. 1. Pressure ratio developed per stage is dependent on the ratio of blade speed/velocity of gas molecules. A pressure ratio 5:1 (per blade) can be achieved for air with blade speed = 1000 ft/sec. This would give a pressure ratio per stage of 25:1. 3. Performance begins to fall when the mean free path is about five times as large as the minimum blading dimensions. if. For a single rotor the geometric variables are the blade angle and ratio of blade spacing (circum- ferential ly)/blade chord length. 10 HYBRID MOLECULAR PUMP Alcatel16 has recently developed a new molecular pump called the "hybrid molecular Fig. 7. Observed dependence of inlet pump." The pump has two stages. The first pressure on outlet pressure for H2 stage uses the turbomolecular structure air, Freon-I2. (open blading), while the second stage 136 has a drag pump construction. The arm of Dynamic Seal. Alcatel17 developed a this concept is to combine the high com- screw-type seal to withstand a pressure pression ratio of a drag pump with the difference from 1 x 10"3 to 760 torr for high pumping speed of a turbomolecular 50 mm (2 in.) diameter shaft at 24,000 rpm. pump. Figure 8 shows the pump in detail. Figure 9 shows the seal. The seal is The turbo part has three stages on the specifically designed to operate in three rotor and four for the starter. The drag different flow regimes, i.e., molecular, pump is of multigroove type and the grooves transition, and viscous. For each of the are in the stator. There are five grooves flow conditions efforts were made to opti- on the low pressure side and 15 on the mize the geometry of the threads. high pressure side. The following gives some of the pump parameters. Turbomolecular Stage -THREADS Rotational speed = 24000 rpm Pump diameter = 200 mm (= 7.8 in.) Pumping speed = 460 H/s Compression ratio at zero flow (Nitrogen) = 150 Drag Stage Pump diameter = 138 mm (= 5-4 in.) Cylinder length = 80 mm (» 3.15 in.) Depth of groove varying Pumping speed - 45 H/s Compression ratio at zero flow (for Fig. 9. Dynamic Seal for Alcatel Hybrid Pump Nitrogen) = 109 Rotor Dia. = 50 mm (1.968 in.) Threaded Shaft Length - 50 mm (1.968 in.) Radial Clearance - .01 mm (.0004 in.) to .015 (.0006 in.) Figure 10 shows the pressure vs. flow rate. It is interesting to note that a seal designed to withstand a pressure difference between 760 and 1 x 10"3 torr wii? leak when it withstands a bigger pres- sure difference. On the contrary, if the pressure difference is smaller the seal will pump. This property is unique as the dynamic seal could be used as a backup pump. to SPIHDLE 40*.(.ei.lo.0. / H» 24,000 rpm 1 19 - 1/ 4 Fig. 8. Cutaway of Alcatel "Hybrid Pump" Rotor dia. = 200 mm (7.874 in.) turbo. 10 = 138 mm (5.433 in.) drag. 10-' PRESSURE ,»rr The pump uses a dynamic seal which has Fig. 10. Vacuum pressure vs. flow. unique design features. The following Fore pressure = 760 torr describes the seal in detail. (Alcatel dynamic seal) 137 ratios for a given length and runs at CONCLUSIONS small radial clearance (0.05 mm = 2 mils), whereas turbo is a multistage 1. Many conventional types of seals may be pump with high pumping speed and eliminated because of high rotational operates at much larger radial clear- speed and extreme pressure conditions. ance (0.5 mm = 20 mils). These include face seals, lip seals, and packings. 5. The performance of molecular pumps, drag and turbo, is significantly 2. Close clearance seals such as capillary, affected by the fore or downstream face, were evaluated. These were found pressure. Typically a molecular pump not to be suitable due to high leakage with fore pressure less than 0.01 rates. torr, can develop a compression ratio about I0G; but for higher values of 3. Seals of interest were found to be those fore pressure compression ratio falls which have the pumping capability. off drastically. These are: centrifugal, spiral groove, and viscoseal (screwj. Since the 6. To k^ep the gas flow in molecular flow pressure is lO""* torr on the vacuum regime, physical dimensions of the side and 760 torr on the high pressure pump are varied along the length of side, a seal in this application will the rotor. Generally one tries to operate in three different flow regimes, keep the values of minimum-pump di- i.e., molecular flow, transition, and mensions to about 1/4 to 1/5 of mean viscous. The review of literature in- free path of gas. dicated that most of the work has been done in viscous flow regime with the ex- ception of a few studies which have been 7. The review of fluid seals and molecu- extended in the transition flow regime. lar pumps Indicates the principle of However, evaluations were carried out at molecular pumps can be combined with viscous flow conditions to get some idea the principle of viscous pumping seals of their relative performance. The in order to develop a seal for an results indicate: advanced flywheel system. The analysis will require defining the three flow a. The visco seal has the lowest power zones along the length of the seal. losses-to-pressure sealed ratio. Once the flow zones are defined, theories of molecular pump, slip flow, b. The performance of all the seals de- and viscous flow can be used to pre- teriorates when the flow conditions dict the performance. change from viscous to transition. 8. Since molecular pumps are of two con- structions, two seal configurations c. It was also concluded that a seal de- are possible. The advantage of drag sign based on the continuum fluid type construction is that it is simple dynamics will show further deteriora- and develops high compression ratio tion in its performance when applied in molecular flow regime. as compared to the turbo type. However, Therefore, theories developed for the disadvantage of the construction these seals cannot be applied to de- is that it runs at fairly small radial sign a seal for the present applica- clearance (1 to 2 mils). tion. RECOMMENDATIONS To study the pressure building ability 1. Investigate the screw types of seal of gases in molecular flow conditions, concepts where the design on the vacuum theory of molecular pumps was reviewed. side is based on the principle of Basically there are two types of molecu- molecular drag pump and continuum fluid lar pumps; drag and turbo, both working dynamics on the high pressure side. essentially on the same principle, but For analysis, existmg works could be their construction is entirely different. used, and efforts should include ex- Both types of pump can achieve compres- tensive parametric studies prior to sion ratio of the order of 107 provided selecting the final configuration. the gas is in molecular flow condition. A drag pump achieves higher compression 2. Investigate the seal concept which uses the construction of a molecular turbo 138 pump. For analysis, existing works Vacuum Technology (Peramon Press, 1961), could be used, and efforts should include pp. 6-12. optimizing the rotor and stator design 15G.E. Osterstrom and A.H. Shapiro, Improved at each stage. Turbomoleeular Pump, J. Vac. Sci. Tech., Vol. 9, No. 1 (1972). 3. For both seal concepts (1 and 2 above), 16L. Maurice, A New Molecular Pump, Proc. investigate a design for control of 6th International Vacuum Congr. 1974, leakage at start and shutdown. Japan, J. Appl. Phys. Suppl. 2, Part 1 (1974). 4. Evaluate both designs and select one for 17L. Maurice, Dynamic Seals, Proc. 6th manufacturing, testing, and final evalu- International Vacuum Congr. 1974, Japan, ation. J. Appl. Phys. Suppl. 2, Part 2 (1974). REFERENCES ^.J. Santeler, D.W. Jones, O.H. Holkeboer, and F. Pagano, Vacuum Technology and Spaae Simulation (NASA SP-1O5, 1966). 2S. Dushman and J.M. Lafferty, Scientific Foundations of Vacuum Technique, 2nd ed. (John Wiley 6 Sons, 1962). 3C.M. Van Atta, Vacuum Science and Engi- neering (McGraw-Hill, New York, 1965). "*E.A. Muijderman, Analysis and Design of Spiral-Groove Bearings, J. Lub. Tech. (July 1967), pp. 291-306. 5J. Zuk, Dynamic Sealing Principles, (NASA TMX-71851, April 1976). 6F.C. Hsing, S.B. Malanoski, Mean Free Path Effect in Spiral-Grooved Thrust Bear- ing, Paper No. 68-Lubs-17, presented at Lubrication Symposium, Las Vegas ASME (1968). 7E.F. Boon and S.E. Tal, Uydrodynamisahe Dichtung fur Rotiersnde Wallen, Chemie- Ing-Technik Vol. 31, No. 3 (Jan. 1959), p. 202. 8W.K. Stair and R.H. Hale, The Turbulent Visco Seal-Theory and Experiment, Third International Conference on Fluid Sealing, Cambridge, England, Paper H2 (April 1967). 9M.W. Milligan and H.J. Wilkerson, Theo- retical Performance of Rarefied-Gas Vis- coseals, ASLE Transactions 13 (1970), pp. 296-303. 10M.W. Milligan and H.J. Wilkerson, Visoo- seal Performance for Rarefied Gas Sealant, 5th International Conference on Fluid Sealing, Warwick, England, Paper Bl (1971). 1]W. Gaede, The Molecular Air Pump, Ann. Physik, 41 (1913), pp. 337-380. 12A. Guthrie, Vacuum Technology (John Wiley 6 Sons, 1963). 13W. Becker, Zurtheorie der Turbo-Molekular- Pumpe, Vakuum-Technik 7 (1961), pp. 199- 204. llfC.H. Kruger and A.H. Shapiro, Vacuum Pumping with Bladed Axial Flow Turbo- machine, Trans, of 7th Symposium on 139 PROJECT SUMMARY Project Title: A Composite Flywheel for Vehicle Use Principal Investigator: F. C. Younger Organization: William M. Brobeck & Assoc. 1235 Tenth Street Berkeley, CA 94710 415/524-8664 Project Goals: Design and fabrication of fiber-composite flywheels with a usable energy in the range of 1 to 5 kWh with a total energy density in excess of 80 Wh/kg at the maximum operating speed. The design power output rate is 37 kW. The maximum diameter is 0.6 m and the maximum thickness is 0.2 m. Project Status: A preliminary design of the flywheel has been completed. This design uses a biannulate rim of S2 fiberglass/ epoxy overwrapped with Kevlar 49/epoxy. This rim is supported from an aluminum hub by polar catenary spokes filament wound with Kevlar 29/epoxy. The com- bination of materials and shapes yields a compatible set of stress/strain relationships for all of the fiber-composite components to insure that centrifugally generated stresses are principally in the direction of the various fiber orientations with minimum stresses transverse to the fibers. The geometrical parameters have been selected to provide that the various fibers are all stressed to approximately the same percent of their ultimate strength to assure that the maximum energy density will be achieved. Design calculations show that project energy and energy density goals will be satisfied. Design drawings and specifications for the biannulate rim have been prepared and sent to fabricators for quotations. It is expected that orders will be placed by the end of October 1978. Design work is presently concentrated for optimum spoke and hub parameters. Adequate rigidity of the spokes is required to insure dynanic stability. However, excessive rigidity will give undesirable transverse stresses and a reduction in maximum energy density. Stability criteria have been established, and designs which satisfy these criteria are being subjected to stress analysis. Winding forms for the spokes are being designed, and spoke samples for testing will soon be fabricated. The design of assembly fixture has not yet started, although preliminary concepts and methods have been briefly studied to identify the problem areas and indicate probable solution methods. 141 Contract Number: 13-0291 Contract Period: The contract covers only the first phase of a two-phase program. The first phase contract period is nine months and will be completed by the end of April 19.79. Funding Level: $99,000 Funding Source: Sandia Laboratories, Albuquerque 142 A COMPOSITE FLYWHEEL FOR VEHICLE USE Francis C. Younger William M. Brobeck & Associates 1235 Tenth Street Berkeley, California 94710 ABSTRACT A program for the design and fabrication of a fiber-composite flywheel is described. The objective of the program is a usable energy in the range of 1 to 5 kWh with a total energy density in excess of 80 Wh/kg at the maximum operating speed. The design output power is 37 kw and the maximum dimensions are 0.6 m diameter and 0.2 m thickness. The design uses a biannulate rim of S2 fiberglass/epoxy overwrapped with Kevlar 49/epoxy. This rim is supported from an aluminum hub by polar catenary spokes filament wound with Kevlar 29/epoxy. The Kevlar 29/epoxy spokes have a lower tensile modulus than the rim materials so compatible strain levels can be achieved without excessive stress in the spokes. The stress analysis for the biannulate rim shows nearly optimum tangential stress levels in each of the two materials and a very reasonable distribution of radial stresses. Compressive radial stress occurs at the interface between the S2 fiberglass/ epoxy and the Kevlar 49/epoxy. The spoke design concept permits a careful balance of spoke flexibility to assure that the spoke stiffness is adequate to maintain concentri- city of the rim and hub to satisfy dynamic stability requirements while at the same time being flexible enough to allow dilation of the rim due to centrifugal loading without imposing excessive radial loads at the rim-to-spoke attachment points. INTRODUCTION The program is to be divided into two phases: A design and fabrication effort is underway to (1) produce an automotive type Phasa I: Design and fabricate flywheel having suitable mechanical char- specimen wheels to acteristics of energy density, volume, and given specifications. cycle life; and (2) to establish the prob- able cost of production of such flywheels. Phase II: Testing and character- The approach to the development of a fly- ization of specimen wheel to satisfy specific performance wheels. specifications is outlined. A preliminary design of a flywheel to meet the perform- This paper covers the first phase ance specifications is presented. only. Upon completion of this first phase, at least two specimen flywheels will During the course of the work, this have been fabricated and the required preliminary design is being completely documentation will have been delivered. analyzed and reviewed as required to assure a flywheel which is economically and tech- BACKGROUND nically feasible. Material and geometric trade-offs are being examined to maximize The effort required for the design, the probability of achieving a flywheel fabrication, and documentation of at least which satisfies the program objectives. two specimen flywheels must satisfy the Our primary effort is directed toward the following specifications: objective of achieving a maximum total energy density (Wh/kg). However, we also WHEEL SPECIFICATIONS v recognize that the hub-wheel interface is a critical problem area and that a success- (1) The usable energy, defined as the ful solution of hub-to-wheel attachment difference in kinetic energy between may require some compromise in maximum that stored in the wheel at its energy density as may the requirement to maximum design operating speed and at provide a high energy within a limited 33% of that speed, will be in the volume (Wh/m3). range of 1 to 5 kWh. 143 — (2) As i design objective, the total energy the effects of material density will be at least 80 Wh/kg at creep upon balance. the maximum operating speed. For the purposes of this calculation, the (4) An estimate of the costs to weight of the wheel will be taken as produce similar wheels in the weight of the flywheel itself, quantities of 100, 1000, including its hub and whatever por- and 10,000. Note will be tion of a shaft is necessary to couple taken of trends in the cost to it. The weight of additional of any exotic materials. shafting, mechanical drive components, etc., are not to be charged to the PROJECT DESCRIPTION wheel nor is their kinetic energy to be credited to it. High strength-to-weight ratio fiber composite materials seem to have the (3) The overall diameter of the rotating greatest promise for satisfying the ob- assembly will not exceed 0.6 m, and its jectives of an automotive energy-storage thickness must be less than 0.2 m. flywheel. However, the orthotropic nature of these materials presents some serious (4) The wheel shall be designed to run for design problems. Their low transverse . 1000 total hours at its maximum design strengths make it difficult to transfer operating speed without failure. It loads from one radius to another. A simple is to be designed to be cycled from thin rim spinning about its axis can store 33% to 100% of that speed, 10,000 a great amount of energy per unit weight. times without failure. During test- However, finding a suitable support for the ing it will be required to demonstrate ring is difficult because the support must 100 such cycles without failure. not impose large radial loads. Increasing During tests, the wheel will be spun the rim thickness can beneficially lower down at an average power output rate the attachment loading by reducing the of 37 kW. radius at which the rim support is to be attached; however, increasing the thick- (5) Choice of materials is at the discre- ness imposes difficult radial stresses in tion of the contractor, although the the rim. objective of the program is to produce a wheel which is economically as well A design concept to avoid these prob- as technically feasible. lems employs a rim supported by a system of tension-balanced spokes which minimize the The required documentation will include radial forces upon the rim. The design but not be limited to the following: features a biannulate rim composed of two concentric components of different material (1) Complete construction details tightly bonded at their interface. This as adequate to allow the wheel biannulate arrangement permits efficient to be duplicated by others utilization of the available space, allows skilled in the necessary use of material with differing specific arts. These details will moduli of elasticity and reduces the include such things as inform- radial span of the tension-balanced spokes. ation regarding special Figure i shows the design. materials and any unusual fabrication techniques such An energy density in excess of 84 as the use of specialized Wh/kg appears to be possible with such a winding facilities. design using available technology. This arrangement is dynamically stable and will (2) An analysis of the stresses permit more than adequate power extraction. induced in the wheel during fabrication and operation, RIM DESIGN with an evaluation of fatigue effects and of the scaling The design objective for very high laws involved in changing total energy density requires that much of the size of the wheel. the material be highly stressed by centri- fugal loading. A radially thin ring spin- (3) An analysis of the dynamics ning about its axis is a well-known example of the wheel with particular of a geometric arrangement where all the emphasis on critical speeds and material is uniformly stressed in pure 144 tension (in the limit of infinitesimal thickness). For such a geometry, it may be shown that the energy density is related to the strength-to-weight ratio of the material by the equation: and the dilation is found from: w " 2 p u; where E = the energy in Joules w = the weight in kilograms 0= the strength in Pascals The maximum radial stress is found as: p = the density in kilograms/metre3 The factor of one half can be considered as a shape factor for a very thin ring. As the radial thickness of the ring is in- wher•_ e a = ra- creased, this value drops off slowly. For example, bringing the radial thickness to »• 11 percent of the outer radius would drop a = inner radius the shape factor from .5 to .465. In- b = outer radius creasing the radial thickness of a ring u = angular velocity increases the total energy stored in the r = radius at which stress is available space, while lowering the total calculated energy density (Wh/kg). R = outer radius w = weight density The requirement for usable energy in E = modulus of elasticity the range of 1 to 5 kWh from a flywheel g = gravitational constant not to exceed 0.6 metres in diameter with a = unit stress a thickness less than 0.2 meters indicates u = Poisson's ratio that the fiber composite must occupy a fairly large fraction of the available Thus, increasing the radial thickness space. Economic feasibility and vehicle causes a large increase in radial tensile integration also demand an efficient stress which could lead to failure because utilization of space. From our analysis, filament-wound fiber-composite materials it appears that a thin ring would not pro- have very low transverse tensile strength. vide the most efficient utilization of There are several ways to circumvent the available space and that a thick ring or difficulty. The one we will use is to combination of rings is necessary. fabricate a biannul ate ring with two concentric elements having materials with Of course, there are also disk-type differing densities and tensile moduli.* flywheels utilizing radial fibers and cross- These two rings can be wrapped one on top ply composite. We recognize that these of another so that they will bond on curing. other type flywheels also may provide an Under the action of centrifugal loading, efficient utilization of space (Refs. 1, the dilation of the innermost ring will be 2, 3, 4). However, we prefer a design resisted by the outer ring and a pressure similar to those which we have tested that at the interface will be produced. show good promise of success. The interfacial pressure can be com- The stresses in a thick ring may be puted from the deflection equations and calculated by equations in Ref. 5. The the requirements for continuity. Reference equations in Ref. 5 are for isotropic 5 gives the stresses and deformation of rings rather than for orthotropic rings. isotropic thick rings subjected to exter- However, comparisons of calculations of nal and internal pressure. the stresses for isotropic and orthotropic materials (Ref. 6) shows that for rings with radial thicknesses less than 20 per- *This design is quite similar to one iden- cent, the simpler isotropic equations are tified by Reedy and Gerstle (Ref. 7). quite adequate. For centrifugal loading, This build-up gives a combined radial the radial and tangential stresses are thickness great enough to give excellent found by: space utilization and to reduce the span of the spokes to an acceptable value. 145 For two concentric rings in intimate design of the rim itself. We expect this contact, the external pressure of one equals to be a difficult engineering problem for the internal pressure of the other and con- any developer of flywheel technology. tinuity requires that the total deformation Allowing rim dilation while maintaining of each at the interface must be equal—The concentricity is a difficult task. We deformation being that due to the combined believe that tension-balanced polar-cate- action of the interface pressure and the nary spokes provide an acceptable solution centrifugal loading. Superposition is to this problem. They will provide ade- used for finding the total deformation; thus, quate rigidity and torque-carrying capa- the deformation of the interface edge of the city. inner rings can be represented by the follow- ing equations: The rim-to-hub attachment method employs a system of tension-balanced polar- catenary spokes as shown in Fig. 1. A Pal,2 = \ (6) polar catenary is the naturally-assumed curved shape of a flexible element under Pa2,2 = h (7) tension in a centrifugal force field. These spokes will be pre-formed on a where P = interface pressure special template to obtain the desired <*•. i = dilation of outer edge of inner polar-catenary shape to insure pure ten- ' ring due to centrifugal force at sion loading with minimum radial loading unit angular velocity of the rims at the points of attachment. a2 , = dilation of inner edge of outer Continuity requires that the average ' ring due to centrifugal force at stress level in the spoke be proportional unit angular velocity .' to the stress in the rim at the point of a, « = dilation of inner ring due to attachment, with the proportionality ' unit pressure factor being equal to the ratio of tensile a dilation of outer ring due to 2 2 unit pressure modulus of the spoke material to that of = dilation at the interface the rim. It is desirable to use a spoke 1 material with a modulus lower than that of si = dilation at the interface the inner portion of the rim in order to allow spoke stresses to be lower than rim stresses. By setting 6j = 62, the value of P can be found as a function of u> as below. Figure 3 shows the polar-catenary shape and the force balance on an element. P = (8) The derivation of the shape is given in Ref. 8. The amount of curvature of the spoke is determined by its tension and den- By using these relationships, a two- sity. Increasing the tension or lowering part rim using an inner ring of S-2 fiber- the spoke density gives less curvature. glass and an outer ring of Kevlar 49 has The amount of curvature limits the radial been designed. This arrangement allows an span which can be bridged by a practical overall radial thickness of slightly over spoke. For the proposed design, the hub- 25 percent. Thus, slightly more than 45 to-rim radius ratio is almost 3 to 1. This percent of the available volume is used for ratio appears to be too great for an S-2 the rim. The centrifugal force field at fiberglass spoke but appears to be accept- the spoke-to-rim attachment point is about able for a Kevlar 29 composite spoke. The half that at the rim outer edge. Kevlar 29 is an ideal candidate material for the spokes because of its low density Figure 2 shows the stresses in the and because its modulus of elasticity is proposed two-part rim at 32,725 rpm, in just slightly less than that for S-2 fiber- the absence of creep, thermal stresses and glass. This modulus is low enough to fabrication stress. All the values in insure reasonable stresses while at the Fig. 2 vary with the square of rotational same time is high enough to provide ade- speed. Thus, all stresses are zero at quate rigidity for the hub-to-rim connec- zero speed. tion . Inadequate rigidity would lead to dynamic instability of the ring support. TENSION-BALANCED POLAR CATENARY SPOKES The means of connecting the rim to The spokes will provide a sufficiently the hub is at least as important as the accurate and rigid hub-to-rim connection 146 to insure balance and dynamic stability. where F = the force on a spoke element The concentricity of the hub and rim will L = the length of spoke not be absolutely perfect. Some small E = modulus of elasticity initial or residual eccentricity will exist. A = cross-sectional area Thus, at high speed a centrifugal force pro- portional to the square of the angular ve- The displacement associated with locity and the first power of the offset straightening the spoke is found from distance between the center of mass of the (Ref. 8): rim and its center of rotation will be generated. (13) The combined stiffness of all the spokes oppose this force but some increase in the offset will occur. This increase is inversely proportional to the combined where w - density of spoke material spring constant of the spoke system and R = hub radius may be expressed as: u> = angular velocity L = curved length of spoke Ax = F/K (9) x = length of chord of curved spoke where "K" is the spring constant of the Since the component 6% is dependent upon u2 spoke system. and the density of the spoke, the centri- fugal force on the spoke acts to increase Even though tha initial eccentricity its stiffness. of the rim may be very small, there will be a force which will cause the eccentricity The requirement for the critical to increase which will in turn permit the speed to be in excess of the operating force to further increase. In order to speed sets a minimum value of the overall have elastic stability, the force associated spring constant, which in turn sets a with the spoke stiffness at a given offset minimum cross sectional area for the spokes. must be greater than the centrifugal force This minimum cross sectional area can be associated with that offset. This requires found using Eq. (12) and (13). that: In addition to the stability of K > Moo2 (10) radial offset (eccentricity), there will be a requirement that the spoke system pro- Thus, there will be a critical speed u vide adequate rigidity to prevent instability of out-of-plane motion. What is important (ID here is that the principal axis of the rim and the hub will not be absolutely parallel The curved spokes appear to be quite with the axis of rotation; consequently, flexible. A force applied at the end of there will be a gyroscopic torque acting the curve spoke tends to straighten the between the rim and hub tending to change spoke and also acts to stretch the spoke the alignment of their axes. This change material. Thus, there are two components in alignment will be resisted by the stiff- of the elasticity of the spoke. The ness of the spoke system. stretching of the spoke material is direct- The out-of-plane whirl modes are de- ly related to the modulus of elasticity. pendent upon the rotational frequency and The straightening of the spoke is more the ratio of the polar moments of inertia complicated and is found to be dependent to the diametral moments for both the hub upon the centrifugal loading on the spoke and rim. The moment of inertia of the rim (Ref. 8). The overall spring constant, is so much greater than that of the hub that Kg can be expressed in terms of the com- the frequency of this mode may be approxi- bination of these two components. mately calculated from the moments of the hub alone. The approximate relationship The stretch of the spoke is found at low speed is: from: * _ £k (12) (14) 6 -VF 1 EA 14 7 where K is the spring constant defined as Laboratory are used as a guide for fiber- the restoring torque due to a unit of angular composite properties. Where additional rotation of the hub about its diameter. data are required, specimens will be pre- pared and tested. At high rotational speed, the gyros- copic moment of the hub becomes signifi- The main features of the design will cant and the frequency for this mode be retained as it is believed that this increases to approach: concept provides promise for solving the problems faced by filament-wound fiber- composite flywheels. = An (15) The stresses associated with acceler- where a is the rotational speed of the fly- ation and deceleration are very small at wheel and A is the ratio of polar moment high speed. A preliminary examination of inertia of the hub to its diametral shows that the added spoke tension due to moment. deceleration is a few pounds compared to about 20,000 pounds for centrifugal load- For a thin disk A equals 2 but for the ing. thick hub of the flywheel: REFERENCES A = 2r2/(r2 + V) (16) 1. Hatch, B. D., "Alpha Cross-Ply Com- where r = radius posite Flywheel Development," 1977 a = axial height Flywheel Technology Symposium, CONF- 771053, March 1978. At very high speeds, the modal fre- quency will be less than the rotational 2. Lewis, A. F. and Gupta, P. W., "Opti- frequency. Figure 4 shows the shape of the mization of Hoop/Disk Composite curve of whirl mode frequency as a function Flywheel Rotor Designs," 1977 Flywheel of rotational speed. The point where the Technology Symposium, CONF-771053, modal frequency equals the rotational fre- March 1978. quency is shown at point P. This critical speed for self-excitation of whirl is al- 3. Rabenhorst, D. W., McGuire, D. P., ways greater than the value from Eq. (14). and Lewis, A. F., "Composite Flywheel Operating above the critical speed is to Disk/Hub Attachment Through Elasto- be avoided as dynamic stability cannot meric Interlayers," 1977 Flywheel be assumed. By raising the stiffness of Technology Symposium, CONF-771053, the spoke system, the critical speed can March 1978. be kept above the maximum operating speed. 4. Garber, A. M., "Polar Weave Composite The relationship for dynamic stability Flywheels," Proceedings of the 1975 for the whirl or wobble motion is similar Flywheel Technology Symposium, ERDA to that for the radial offset so that a 76-85. critical stiffness must be exceeded. This will require axial spacing of an array of 5. Timoshenko, S., "Strength of Materials, spokes. Part II," Third Edition, D. Van Nostrand Company, Inc., 1956. PRELIMINARY DESIGN 6. Morganthaler, G. F., and Bonk, S. P., The preliminary design for the fly- "Composite Flywheel Stress Analysis wheel is shown in Fig. 1. Its character- and Material Study," Society of istic? are summarized in Table 1. The Aerospace Material and Process available energy of 4.60 kWh is within Engineers, Vol. 12, 1967. the desired range and its energy density of 84.3 Wh/kg is above the desired minimum. 7. Reedy, F D., Jr. and Gerstle, F. P., In order to finalize this preliminary design, Jr., "Dfc-ign of Spoked-Rim Composite trade-off studies, additional stress cal- Flywheels," 1977 Flywheel Technology culations, and further dynamic analysis Symposium, CONF-771053, March 1978. are required. The effects of dimensional tolerances and material variations are 8. Younger, F. C, "Tension-Balanced being considered. Preliminary engineering Spokes for Fiber-Composite Flywheel design data from the Lawrence Livermore Rims," 1977 Flywheel Technology Symposium, CONF-771053, March 1978. 148 FIBER - COMPOSITE Fig. 1. Tension balanced catenary spoke flywheel. 149 091 Tangential Stress x 10 psi 1—* I—» ro o in o in O o o o o Radial Stress x 10 psi Interface H + AH V + AV T + AT Fig. 3. Polar catenary force diagram. Fig. 4. Whirl mode frequency as a function of rotational speed. 152 Table 1. Flywheel Characteristics Outer Ring Inner Ring Spjikes Outside Dia. 23.52 in 19.70 in Type Polar-catenary Inside Dia. 19.70 in 16.50 in Loadi ng Tension-balanced Height 7.85 in 7.85 in w/ added weights Material Kevlar 49/epoxy S-2 Glass/epoxy Material Kevlar 29/epoxy Fiber Fraction 60% , 60% , Fiber Fraction 65% , Density .050 lbs/inJ .075 lbs/inJ Density .05 ibs/iV Elastic Modulus 11.76 x 106 8.15 x 106 Modulus 5.85 x 106 Poisson's Ratio .31 .282 No. of Spokes 16 x 4 Weight 50.9 lbs „ 53.6 lbs 9 Total Weight 4.64 lbs Moment of Inertia 15.4 Ib-in-sec 11.4 lb-in-sec^ Total Moment of Inertia .41 lb-in-sec^ Hub Summary of Characteristics Diameter 6 in Total Weight 135.3 lbs , Height 7.85 in Total Moment of Inertia 28.09 lb-in-sec Material Alum.alloy 7075 Rotational Speed 32,725 rpm Density .1 lbs/in3 u 3,427 rad/sec Weight 22.2 lbs , Stored Energy 5.18 kWh Moment of Inertia .26 Ib-in-sec Energy Density 84.3 Wh/kg Available Energy 4.60 kWh Max. Hoop Stress in Kevlar/epoxy 216,736 psi Loading Weights Max Radial Stress in Kevlar/epoxy 3,013 psi compression Max Radial Tension in Kevlar/epoxy 1,144 psi Material Aluminum Max Hoop Stress in S-2/epoxy 189,349 psi No. of Weights 8 Max Radial Stress in S-2/epoxy 3,013 psi compression Weight Each .50 lbs Max Radial Tension in S-2/epoxy 1,004 psi Total Weight 3.98 lbs Max Tension in Spokes Kevlar 29/ Moment of Inertia .63 lb-in-sec epoxy 163,057 psi Critical Speed for Radial 64,000 rpm Stability 153 PROJECT SUMMARY Project Title: Prototype Development - Composite Flywheel Having Nominally 40 Watt-hrs/lb Energy Density Principal Investigator: P. W. Hill Organization: Hercules Aerospace Division Hercules Incorporated Allegany Ballistics Laboratory P. 0. Box 210 Cumberland, MD 21502 (304) 726-4500 Project Goals: The development of hardware that will demonstrate the very latest technology relating to composite wheels. This project is one of four that explore different approaches to producing a wheel having the highest practical energy density. Project Status: This concept is unique in that it incorporates a solid circumferentially-wrapped composite wheel with an elongated hour glass cross-section. Such a profile places most of the mass near the periphery of the wheel ~ as distinguished from the more common modified Stodole disk shape which concentrates the mass at a relatively snail radius. Considerable attention is being given to the disk-hub interface, where the use of an elastomer as a grading medium is being explored. Sandia's work in aerodynamic heating of composite wheels is being incorporated at the design stage in the interest of managing thermal stresses. At present, design and optimization of the disk contour are complete. Aerodynamic heating analysis is under way, and dynamic analyses by a subcontractor (Rockwell) are beginning. Contract Number: 13-0292 Contract Period: July 1978 - June 1979 Funding Level: $86,000 Funding Source: Sandia Laboratories, Albuquerque 155 PROGRESS IN COMPOSITE FLYWHEEL DEVELOPMENT P. W. Hill T. C. White D. G. Drewry W. B. Stewart Hercules Aerospace Division Hercules Incorporated Allegany Ballistics Laboratory P. 0. Box 210 Cumberland, Maryland 21502 ABSTRACT Development of filament reinforced polymer (FRP) flywheels for energy storage in vehicles has been continuing at ABL with emphasis on material selection and design syn- thesis based on the filament winding process. Three species of designs have been studied: 1. Helically wound shells; 2. Circumferentially wound rims with radial spokes; 3. Circumferentially wound contoured disks. The objective has been to evaluate the potential of design concepts with commercial merit. A new approach to the type 3 flywheel offers competitive performance in a wheel configuration that is moire compact than other types and is also attractive from a mass production viewpoint. Combining optimum contour with programmed winding tension in a process that fixes the winding tension as the wheel is fabricated results in rated per- formance ar~rnr...'u.La& 30 W-hr/lb and 1 W-hr/cu in. for a rotor storing 1270 W-hr at the design operating speed. Design and fabrication are in progress under contract from DOE/ Sandia. A description of the design concept, the basic theory of the design, and a summary of material properties are provided. INTRODUCTION 6. Applying a beneficial (compres- sive) prestress field by press fit or Circumferentially wound composite disk shrink fit assembly of individual rings flywheels offer the advantage of a low cost or rings and hub (loses cost advantage). manufacturing process. However, in many instances their performance is unsatisfac- 7. Applying a beneficial prestress tory as a result of radial tensile stresses by successive winding and curing of concen- which exceed material strengths. Some tric rings (loses cost advantage). mechanisms to reduce the radial stresses include: 8. Applying a beneficial prestress by modifying the winding tension program 1. Increasing radial compliance of and cure process during fabrication. the material. All of these have been studied and reported 2. Introducing bands or rings of with differing degrees of optimism, but another more compliant material. with little experimental success, unless the depth (Ro-Ri) or radius ratio (Ri/Ro) 3. Using concentric but separate was limited so that the "disk" was more rings with mechanical coupling. nearly a "rim." A. Using concentric bands of different Hercules has studied combinations of fibers which provide a modulus gradient the above mechanisms to identify benefits increasing toward the rim and/or a mass that might be gained thereby and which gradient decreasing toward the rim. might offer acceptable performance. Em- phasis was placed on methods 4 through 8. 5. Contouring the faces of the disk. Results have indicated that combining 156 - methods 5 and 8 is most beneficial and can temperature cure cycle. This latter prop- achieve useful performance without further erty permits all of the beneficial (compres- complication. The following discussion sive) radial prestress to be applied to will outline the basis for this observation. counteract the spin stresses, thereby im- proving energy storage rather than merely BACKGROUND combatting manufacturing residual thermal stresses. Toland1 and Hill2 have described the use of a transfer matrix approach for RADIAL STRESS. °r computing the stresses in a rotating orthotropic disk for the purpose of find- ing the optimum disk contour. An extension 15-j J / X/ 60 of that approach was prepared to calculate the stress distribution in a stationary disk resulting from an arbitrary program a 10 II \ / 40 of winding tension. The two solutions are V HOOP STRESS. *£ \/ superimposed in a computer program that seeks the optimum contour for the prescribed 20 u. winding tension program. Reuter3 has shown that the residual stresses associated with conventional fila- ment winding and cure processes are detri- -20 mental. Fig. 1 illustrates the effect for FLA1• DISK an example case with uniform winding ten- R = 11.8 IN sion. These residual stresses can cause o Ri = 2.0 IN -40 failures by and of themselves. The maximum 6 permissible AT of 163°F in the example of E 157 contours, stress distributions, and Table 4. Design allowables. geometrical correlations. The first step (150°F, 104 cycles, 103 hours) was to establish the residual stress distribution resulting from "freezing" Type AS/ Type AS/ Kevlar-49/ the winding tension as a function of p opertv Epon-82B/T403 3004 (Ri/Ro) and the tension program. (Note: LongIcud!n 1 Tensile Str ngth. 155.0 142.0 101.0 In all cases the O.D. was fixed at 23.6 Longitudlna 1 Conpresslve 94.0 60.0 26.3 inches.) Strength, i Hc (Wai) th, 2.7 0.50 Table 1. Properties of candidate Transverse Tensile Stren Z.9 fibers and resins. Transverse^ Compressive 11.5 12.0 8.9 Strength, F Fibers: In-Plane Shear Strength, 3.5 7.3 1.0 Property Type AS-6 Carbon Kevlar-49 5.3 6 Tensile Modulus <10 psl) 34.0 19.0 F23 Table 2. Nominal room temperature composite material property data. 7i) en Type AS/3l)U<. Type AS/Epoxy Kevlar-49/Epoxy Property (Vf - 57iQ (VF - 603? (VF • 60S) Longitudinal Tensile Hodu- 16.5 18.0 10.5 lus. En HO6 psl) Transverso Tensile Modulus, 1.2 1.0 0.50 "?_ -0.5 6 E22 (10 pal) b Major Poisson's Ratio. Vl7 O.34 0.2? 0.34 In-Plane Shear Modulus, 0.56 0.B5 0.30 Gi2 (106 psl) Longitudinal Tensile 190.0 210.0 225.0 r Strencth, F'n F12 Cksl) Short Beam Shear Strength 11.6 B.O 5.0 (test) -3.0 CoeffUienr of Thermal Expirsion 6 U-nf-itudinal (10" -0.006 0.00 -^.G TRANSVERSE in./in./°F) COMPRESSIVE Transverse (10"6 27.0 15.0 20.0 STRENGTH in./in./°F) LIMIT- -4.0 (l)Typical for an epoxy reain such as Epon 828/JcfFaalne T-403. Fig. 2. Stresses in an orthotropic disk flywheel. 158 Table 3. Material degradation to design environment. 150°F Temperature Degradation Due Degradation Due Material Variation Degradation to Static Fatigue to Dynamic Fatigue Material -3X) (%) (% of Ultimate) (% of Ultimate) (% of Ultimate) Lone. Trans. Shear Long. Trans. Shear Long. Trans. Shear Long. Trans. Shear Type AS/Epon 826/T403 (2) Tension 15 3<2> 5 10") 2O(2) Compression 25* ' io"> (3) 3) Shear 30(1) 15(1) 5 20< Kevlar-49/Epon 826/ T403 (3) (5) 20(i) (4) 35 15 20(3) Tension (1) 30 Compression 35<1) 25* ' io»> Shear 25(D 2O(3) 3O<3) 20(3) Type AS/3004 ,5(6) (6) (6 (6) 5(6) Tension 303°(6<6) 0 o ' 3 10(6) 20(6) Compression (6) (3) (3 Shear 30 10(3) 5 20 > (l)Hercules Data for Carbon & Kevlar-49 Composites. (2)Based on Data from Hofer, K. E., et al., Reference 6. (3)Estimated. (4)Based on Data from Chiao, T. T., et al., Reference 7. (5)Based on Data from DuPont, Reference 8. (6)Based on Data from Reference 5. In. Fig. 3, the effect of the combined and admits simplification of hub design prestress and spin stress on the failure and high volumetric efficiency as well as condition throughout the disk is shown. weight efficiency. Although the critical condition is still the radial stress, the hoop stress is now elevated to 65% of the allowable fiber r- SAFE OESIGN LIMIT strength. Further improvement of the flat disk, is possible by adjusting the winding tension program so that the peak negative prestress coincides with the peak positive spin stress and by better definition of the maximum prestress allowable. To balance the design so that radial (transverse) and hoop (fiber) failure condi- tions occur simultaneously, the sides of the disk are contoured. Fig. 4 illustrates the relative benefits from prestress and contouring for carbon fiber/polysulfone matrix composite flywheels as computed by the above methods of contour optimization. Substantial deviations from these plots might be expected for other materials and winding tension programs. However, the significant trends are valid: «/»„ 1. Prestress has useful value for Fig. 3. Failure condition in spinning disk relatively thin rims (a between 0.5 and 0.8). after prestress by winding tension. 2. The effect of contouring increases for disks with small holes (a below 0.4) 159 Table 6. Expected performance of the CONTOOtEf preliminary design. WltHltKTKS! Bunt 5pccd (Aftar Service Life Knockdown) 31,560 rp« Dealgn Operating Speed (DOS) 28,230 rp» Stored Energy at DOS 1,270 Uh Delivered Energy at DOS 1,131 Wh Weight, Dl»k 41.3 a Weight, Hub 4.0 lb Weight, Total 45.3 lb Stored Energy Denalty 28.0 Wh/lb Maxima Hoop Streit at DOS 113,600 pal Matimm Mdlal Streaa at DOS 2,320 pal REFERENCES Poland, R. H., and Alper, J., "Transfer Matrix for Analysis of Composite Fly- OJ 0.4 O.« O.» 1.0 wheels," Journal of Composite Materials, RAOftJS tATIO. •••* Vol. 10, July 1976, page 258. Fig. 4. Effect of contour and prestress 2Hill, P. W., et al., "Advanced Flywheel on the performance of orthotroplc disk Development," 1978 Flywheel Technology flywheels. Conference, San Francisco, CA, October 1978. PROJECT DESCRIPTION 3 Hercules Aerospace Division/Allegany Reuter, R. C., Jr., "Fabrication and Ther- Ballistics Laboratory Is under contract to mal Stress in Composite Flywheels," Pro- DOA/Sandia to design and fabricate two test ceedings of the 1975 Flywheel Technology wheels based on the above approach. Energy Symposium, Lawrence Ball of Science, density targets were established as shown Berkeley, CA, November 10-12, 1975. In Table 5. The expected performance of the preliminary design Is given In Table 6. '•Laakso, J. H., "Potential Merits of The final design will be supported by com- Thermoplastic Composite Materials for plete quasi-static stress analysis and Modular Rim Flywheels," Proceedings of normal modes dynamic analysis. The the 1975 Flywheel Technology Symposium, dynamics analysis and planning for all spin Lawrence Hall of Science, Berkeley, CA, testing will be performed by the Rpcketdyne November 10-12, 1975, page 164. Division of Rockwell International under 5 subcontract. The project is just under Anon., "Mechanical Property Data for AS/ way and in the design phase. 3004 Graphite/Polysulfone Composite," Issued by AFML and Prepared by University Table 5. Practical energy density targets of Dayton Research Institute under Contract for composite flywheels in vehicles. F33615-75-C-5085, November 1976. 6 Material Hoefer, K. E., et al., "Development of Kevler-49/ AS Carbon/ AS Carbon/ Engineering Data on the Mechanical Cbaraetarletlc Prosertv fnoTt tooicr PolTaulfona T Properties of Advanced Composite Mate- Teoalle strength, PU (HI) 223 210 190 Denalty (lb/ls.3) 0.050 0.054 0.054 rials," IIT Research Institute, Technical O.S ffu^g), (tfc/lb)1 70.6 63.9 55.2 Report AFML-RT-74-266, February 1975. 2 birlroraental Xnclidoira Factor, (I) SO 26 25 Oltlaete Energy Denalty, Holt OHi/lb) 35.3 47.3 41.4 7Chlao, T. T., et al., "Lifetimes of Fiber General Safety Factor 1.25 1.2S 1.25 Composites Under Sustained Tensile Load- Ideal Operating Energy Denaity, no 28.3 37.> 33.1 (Ub/lb) ing," Lawrence Livermore Laboratory, Faraaitlc Valgbt AUo>ance (I) 13 15 *3 University of California, Contract No. Deelgn Inergy Deaelty Target, tw 24.0 32.1 2S.2 WWlb) B W-7405-Eng.-48, 1976. Ideal ahape factor for coapoalta Mtarlala - 0.5. Sae 8 taferance 9" (Cerat- -U •" Ugg""Ia . 1»7S). Anon., "Kevlar-49 Data Manual". 2. lee Tabu 3. 160 9Gerstle, F. P., and Biggs, F., "On Optimal Shapes for Anisotropic Rotating Disks." Proceedings 12th Annual Meeting, Society of Engineering Science, Inc., University of Texas, Austin, TX, October 20-22, 1975, see also Gerstle, F. P., and Biggs, F., "On Effec- tive Use of Filamentary Composites in Fly- wheels," Proceedings of the 1975 Flywheel Technology Symposium, Lawrence Hall of Science, Berkeley, CA, November 10-12, 1975, page 146. 161 PROJECT SUMMARY Project Title: Prototype Development - Composite Flywheel Having Nominally 40 Watt-hrs/lb Energy Density Principal Investigator: Dr. D. E. Davis Organization: Rocketdyne Division Rockwell International 6633 Canoga Avenue Canoga Park, CA 91304 (213) 884-3075 Project Goals: The development of hardware that will demonstrate the very latest technology relating to composite wheels. This project is one of four that explore different approaches to producing a wheel having the highest practical energy density. Project Status: The concept here is a radially-overwrapped circutoferentially- wound wheel. This design is a follow-on to a wheel that had been developed for attitude control and energy storage in spacecraft. The earlier design utilized an elastomeric buffer at the interface between the windings; this has been identified as a source of mechanical problems and will be omitted in future designs. At present, the first iteration of stress calculations has been completed. Materials characterization tests are now being performed. When completed, these and the stress computations will be incorporated into a detailed design. Contract Number: 07-6955 Contract Period: July 1978 - May 1979 Funding Level: $149,562 Funding Source: Sandia Laboratories, Albuquerque 163 ADVANCED COMPOSITE FLYWHEEL FOR VEHICLE APPLICATION Dr. D. E. Davis Rocketdyne Division Rockwell International 6633 Canoga Ave. Canoga Park, California 91304 ABSTRACT Rockwell's high energy density composite wheel concept is based upon work done as part of a program in spacecraft power and attitude control. Experience and analytic capabilities which evolved during the early program will be utilized in the course of producing next-generation hardware. This paper will describe the design and present typical results of computer analyses treating operation and fabrication stresses. The advantages of the concept (viz., great axial stiffness and high volume efficiency) will be discussed. Earlier design problems will be briefly touched upon. INTRODUCTION design utilized a metallic shaft upon which a composite core was circumferen- Energy wheel development is a high tially wound. An elastomer matrix was priority Rockwell International techno- used for the core. A transition liner of logy effort consisting of both contract elastomer was used between the core and work and related company research.*>2»3 A the radial overwrap of Kevlar/epoxy. The recently completed research program, par- important features of this design concept ticularly applicable to the DOE composite are: flywheel programs, is the Integrated Power • High energy density, and Attitude Control System (IPACS). The IPACS program resulted in the design, fab- • Good volumetric efficiency, rication and test of a Rockwell proprie- • Uniformly distributed loads, tary Kevlar/epoxy tape-wound composite flywheel. The IPACS data is being uti- • Rugged configuration, lized in the current Sandia Laboratories • Conventional mfg. technology. Contract No. 07-6955 issued under DOE Prime Contract No. AF(29-l)-789. BACKGROUND ROCKWELL RESEARCH PROGRAM Rockwell International, in previous contract and research efforts, has comple- KEVLAR/EPOXY OVER WRAP 9 IN. ted composite materials and wheel confi- RADIUS guration evaluation studies which will be utilized in the existing flywheel pro- gram. More importantly, these studies RUBBER LAYER enable the effort to be focused on a relatively narrow set of alternative CIRCUMFERENTIALLY materials and a specific wheel concept. WOUND Rather than studying alternative wheel E GLASS CORE configurations, the effort can be directed toward the evaluation of design TITANIUM refinements that will optimize the SHAFT selected design concept. (^ ROTATION The original Rockwell-evolved concept Fig. 1. Original Rockwell research is shown in the sketch of Fig. 1. This composite energy wheel t> Rockwell carried out a research pro- of tooling concepts, winding, wrapping, gram in 1975 to fabricate and test a model curing techniques, and machining pro- of the rotor concept shown in Fig. 1. The cesses. Application of the elastomeric rotor was designed and fabricated and is interface material was demonstrated on shown in Fig. 2. Testing was completed this research program. Following fabri- utilizing the test setup shown in Fig. 3. cation, the model wheel was tested to evaluate actual performance in relation to design predictions. A test to des- truction provided significant insight into failure mechanisms and the various forms of energy dissipation (Fig. 4). This research activity was completed during 1975 at the Rockwell International Space Division Facility. Fig. 2. Rockwell research rotor Fig. 4. Rockwell research rotor - posttest The Kevlar/epoxy tape-wound research composite test wheel was designed to have a storage capacity of from 30 to 32 watt- hours per pound. The first test wheel measured approximately 18 inches in dia- meter and had a nominal design operating speed of 35,000 rpm. Burst speed was calculated to be approximately 44,000 rpm. Studies indicated that wrapped cores made of S-glass/elastomer and Kevlar/elastomer would give higher speci- fic energies (15 to 20 percent and 35 to 40 percent, respectively) than the E- Fig. 3. Rockwell research rotor being glass/elastomer core. The working installed in spin test allowables for the composite material facility used in the study were taken at approxi- mately 65 percent of the ultimate This research program contributed signi- strengths. Significant test data con- ficantly to the present composite wheel cerning rotordynamics, resonant frequen- development project. Fabrication methods cies and failure modes and effects ana- were demonstrated, including evaluation lysis were obtained although premature 165 failure of the first test rotor occurred titanium alloy, (2) a Kevlar/nitrile core early in the test program. wrapped circumferentially around the titanium hub, (3) a nitrile rubber inter- 'IPACS RESEARCH COMPOSITE ROTOR face liner, and (4) an overwrap applied radially over the interface liner and Under a NASA Contract NAS1-13008, a core. five phase Integrated Power and Attitude Control System (IPACS) flywheel program The Kevlar/nitrile core consisted of was completed. Under Phases IV and V of a series of 6 rings circumferentially the program (Phase V having been com- applied over the shaft hub. For applica- pleted in February 1978) composite test tion of the circumferential plies, a spe- flywheels were designed, fabricated and cial tensioning device and special side tested. tooling plates were used. The tensioning device allowed incremental tape tension A cross-section drawing of the IPACS reduction as the winding of each ring rotating assembly Is shown in Fig* 5, proceeded. Consolidation of the core for The rotor was fabricated from composites each ring was accomplished, using six in- built upon a titanium hub. The rotor dividual, concentric stepped rings, by core was Kevlar in an elastomer matrix applying lateral pressure to both annular and the core was overwrapped with Kevlar side plates. As the material was cured, in an epoxy matrix. the inward movement of the side plates applied the required hydraulic pressure to the prepreg tape to give the desired wrinkle-free and minimum void structure. VACUUM ENCLOSURE In addition, the pressure was controlled COMPOSITE ROTOR to give a 58 percent to 60 percent fiber MO STATOR HOUSING volume composite. • M G 5TATOR END COVER FIXED BEARING Machining of the core to the design HOUSING contour required special tooling and pro- .TACH PICKUP SENSOR cedures as determined from previous com- HOUSING posite wheel fabrication experience. The OILER key to the machining operation is to PBELI SENSOR SCREW chill the wheel to a low enough tempera- PRELOAD HOUSING ture for efficient cutting and yet not to ' PRELOAD PISTON ANGULAR a point where the nitrile matrix would PRELOAD SPRING CONTACT BALL BEARING craze. ILSLINGER WASTE OIL RETAINER MOTOR GENERATOR ROTOR The final stage in the wheel fabri- cation (after the rubber liner was ap- plied) was the application of the over- wrap. This task was subcontracted to the Brunswick Corporation and consisted of radially wrapping the core/liner assembly Fig. 5. IPACS composite rotating with a Kevlar/epoxy composite system. assembly cross-section The outerwrap was applied in a seven pointed star pattern using a numerically Two identical motor-generator (M-G) controlled winding machine. units were designed to be incorporated at both ends of the rotor and together to The IPACS wheel was machine wrapped provide 2500-watt motor or generator with 9 layers of 0.0203 cm (0.008 inch) capacity. Losses were minimized under roving (Kevlar 49) impregnated with a normal operating conditions with magnetic Brunswick formulation of DER 332 matrix efficiency projections greater than 96 material. A tension of 13.3N (3 lbs) per percent. roving was applied and held constant during the winding operations. Fabrication Techniques. The IPACS compo- site flywheel consisted of: (1) an iso- Figure 6 is an edge view of the tropic shaft/hub fabricated from 6A1-4V wheel showing the tapered sides and the 166 shaft stubs with balance discs and the bearings mounted on the journals. Figure 7 shows a side view of the wheel to illus- trate the radial wrap pattern. Fig. 8. IPACS test wheel installed in spin pit using dual bearing support housing This housing which was attached to Fig. 6. Complete IPACS test wheel - the bottom side of the spin pit lid edge view. Rotor weight is allowed versatility in the test setup 59.9 pounds and supported the wheel in case the drive quill shaft failed. The wheel could be (1) suspended freely on the turbine quill with nylon burnout blocks at the upper and lower shaft journals to help contain the wheel in case of failure, or (2) the wheel could be supported in the rig with bearings on the upper and lower shaft journals. In either case, the turbine quill could be attached to the wheel directly or through a flexible coupling. A series of 36 tests were completed investigating dynamic stability, thermal and radial growth, wheel balance shifts and axial stiffness. Although the design goal was 61.7 watt hours/Kgm (28 watt hours/pound) the design goal was never reached due to dynamic instability. Following the test program, the wheel was inspected thoroughly from visual observations and through the use of NDT techniques. No crazing, delamina- Fig. 7. Complete IPACS test wheel - tion, or other failures were noted on side view close visual examination of the wheel. NDT inspections including X-ray, ultra- Test Program. A series of tests were run sonic and neutron radiography, revealed on the IPACS composite wheel. These no gross defects in the wheel such as tests were performed at the Rocketdyne voids, core ply separations, fiber buck- spin test facility using a specially ling or fractures. There were some small designed dual bearing support housing areas shown in the X-rays which could not fixture as shown in Fig. 8. be explained. These areas appear to be 167 located at the bond surfaces between the • APSA - Finite element axisym- core, rubber interlayer and overwrap. metric and planar struc- Development work is required to adapt the tural analysis with or- X-ray and ultrasonic techniques to thick thotropic temperature- section structures such as the composite dependent material pro- wheel. perties; Posttest Analysis. The primary problem • APSAC - Finite element axisym- with the earlier IPACS composite rotor metric and planar struc- design is dynamic out-of-balance. Review tural analysis with of the stress analysis, detail fabrication loading and creep duty procedures, and test results has led to cycles; the conclusion that imbalance resulted from eccentric shifts in core material • APSAPLOT - Plotting program for that varied with rotational speed. It is APSA and APSAC; known that minor amounts of air were trapped on the core during wrapping and • APSADUMP - Data recovery program were undoubtedly retained after cure even for APSA and APSAC. though the core was evacuated and exter- nally pressurized under heat during cure. The IPACS analysis utilized the APSA Further, repeated cure cycles on inner and APSAPLOT programs. wraps would tend to ca .se variation in material properties. These programs allow for orthotropic, temperature-dependent material properties An even more significant problem under thermal and mechanical loads. The with the core was the compression of mechanical loads can be surface pressures, inner wraps by the buildup of outer mate- surface shears, and nodal point forces as rial which was tensioned during wrapping. well as an axial acceleration and angular This effect, in conjunction with external velocity. The continuous solid is re- pressure during cure, produced wrinkles placed by a system of ring or planar ele- and local buckling of inner wraps. The ments with quadrilateral cross-section. test program also indicated that permanent Accordingly, the method is valid for shifts in balance occurred. These shifts solids which are composed of many differ- could be attributed to creep which was ent materials and which have complex geo- noted in the structure. Another reason metry. could be circumferential delamination of core material from defects, insufficient The APSA program analyzes elastic or wetting, overstress, etc., which would elastic-plastic problems with a single set thus allow local expansion and produce of loads. The APSAC program analyzes pro- imbalance. Upon reduction of speed these blems where plasticity and creep must be local cracks would close, preventing de- considered during a duty cycle represented tection by NDT methods. by a series of load increments. Either isotropic or kinematic strain hardening IMPROVED ROTOR SYSTEM can be specified. The APSAPLOT program is used to furnish a visual display of the Utilizing the technology that has finite element results. Contour plots, been obtained through Rockwell's previous grid plots, and graphical plots of radial wrap rotor system programs, seve- stresses and strains can be obtained. The ral improved stress analysis, design and APSADUMP program is used to recover infor- fabrication approaches are being imple- mation stored on the data tape generated mented for the present DOE flywheel pro- by the APSA or APSAC program. gram. As an example of the APSA program, ANALYTICAL ANALYSIS the undeformed mathematical model of the wheel used in the APSA program is shown in The finite element method is used to Fig. 9, and the deformed model is shown in determine the displacement, stresses, and Fig. 10. An axisymmetric rotating elastic strains in axisymmetric and planar solids. body is assumed. The deformed model (Fig. The programs available for this method 10) shows a large displacement of the are: radial rubber liner between the outer radial wrap and the hub. The cause of the large displacement is assumed to be the 168 program analysis, which did not represent and detail wheel dimensions are being the wrap accurately in this area, but suc- traded off with properties to optimize cessive iterations can accurately repre- energy density while providing adequate sent this area. stress/strain margins. 32 p In the outerwrap, the basic convex contour and design is being retained for ease of fabrication and structural loading. 10.000 An improved matrix is being used having more elongation. This matrix material will result from the materials and pro- 20 cesses studies now being carried out in s z order to identify and define properties 16 - oc for stress analysis and factors in fabri- 1 cation. IT 5.000 Of 12 The elastomer liner between the metal 8 hub and outerwrap which serves to distri- bute loads between the harder composite 4 matrix and metal will be retained. How- ever, the rubber liner ovar the remainder oL of the wheel will probably be deleted to 0 5.0000 provide better strain compatibility Z. DIMENSION (INI between the core and outerwrap. 40 4 8 12 Z, DIMENSION (CM) The circumferentiall}' wrapped core design will be retained in concept, but Fig. 9. Composite energy wheel - otherwise will be substantially changed to undeformed structure provide a stiffer structure, free from creep and with an improved strain match between core and outerwrap. For increased stiffness, the core matrix will be changed; TYPICAL ELEMENTS FOR CORE SHEAR polyurethane and blended epoxies will be STRESSES AND STRAINS considered. The wrap from inner to outer 10.000 - layers will be modified to vary modulus and core density to provide improved strain compatibility and increased energy density by increasing stresses of the z inner core. re 5.000 - MODIFIED FABRICATION TECHNIQUES Major fabrication technique improve- SPEED = 32,500 RPM ments are possible in the core which MAXIMUM RADIAL GROWTH = should substantially improve the perfor- 0.396 CM mance of the radial wrap wheel. Of para- (0.156 IN.) I mount importance, the circumferentially 0 5.0000 wrapped fibers will be maintained straight Z, DIMENSION UN.) and concentric during layup by two possi- 1 I I I 4 0 4 8 12 ble means: (1) computer control of wind- Z. DIMENSION (CM) ing to automatically maintain proper *-<:n- sion and concentricity and (2) single Fig. 10. Composite energy wheel - filament wrapping with wet layup. Layers deformed structure will be cured during the wind process or in intervals of 0.63 cm (0.25 inch) or MODIFIED DESIGN less to preclude compressing inner fibers. These techniques will hopefully avoid en- An improved design wheel will be trapping significant bubbles in the matrix based on many factors of materials, pre- and prevent buckling. At prescribed in- tervals, the core will be machined as dictable stress versus allowable stress, required and balanced. This will be ac- and fabrication processing. The APSA complished by wrapping variable density stress analysis program is being updated 169 tape around the circumference to obtain balance and not change concentricity. The outerwrap may be applied in stages to allow dynamic balancing of the wheel by applying patches as above with entrapment by subsequent layers of fibers. In either case, it will be the purpose of this procedure to arrive at a finished fabricated wheel that is substantially in dynamic balance. REFERENCES 1. SD77-AP-OO38, Fabrication and Test of a Composite Rotor for Integration With an IPACS Rotating Assembly, Final Report, Rockwell International Space Division, July 1977. 2. SD76-SA-OO14, Design Report for the Fabrication of a Composite Material . Rotor and Integration With an IPACS Rotating Assembly, Rockwell Inter- national Space Division, September 1976. 3. Davis, D. E., C. Heise and D. Hodson, Rocketdyne's High-Energy-Storage Flywheel Module for the U. S. Army, Rockwell International Rocketdyne Division, October 1977 170 PROJECT SUMMARY Project Title: Prototype Development - Composite Flywheel Having Nominally 40 Watt-hrs/lb Energy Density Principal Investigator: David L. Satchwell Organization: Garrett-AiResearch 2525 W. 190th Street Torrance, CA 90509 (213) 323-9500 Project Goals: The development of hardware that will demonstrate the very latest technology relating to composite wheels. This project is one of four that explore different approaches to producing a wheel having the highest practical energy density. Project Status: This is a rim-and-spoke design based on a wheel produced for the Army Mobility Equipment Research and Development Command ("MERADCOM"). That wheel, which incorporated a brute force hub and spoke con- figuration intended only as an expedient for rim testing, has already been successfully spun to an energy density of 26 watt- hrs/lb. The new version will use a similar rim, but incorporates a sophisticated carbon-fiber hub and spoke design. At present, fatigue and ultimate strength tests have been performed upon samples of the new design spokes. Hub fabrication has begun. Contract Number: 13-0293 Contract Period: July 1978 - June 1979. Funding Level: $267,759 Funding Source: Sandia Laboratories, Albuquerque 171 HIGH-ENERGY-DENSITY FLYWHEEL David L. Satchwell Garrett-AiResearch Manufacturing Company of California 2525 W. 190th St., Torrance, California 90509 ABSTRACT This paper describes the design and fabrication of a flywheel rotor with an energy density of 80 w-hr/kg. the design features a multiring, S-glass and Kevlar composite material rim that is mounted on a graphite composite spoked hub. Composite material flywheels have been constructed with identical rim design and with aluminum spoked hubs. These flywheels have been successfully tested to energy densities of 53 w-hr/kg. The use of graphite composite material in place of the aluminum material significantly reduces the weight of the rotor assembly and thereby increases the energy density. This paper will present the progress to date on this program and describe the program planned to complete the fabrication of a test specimen. INTRODUCTION INNER S-2 GLASS EPOXY COMPOSITE KEVLAR 29 EPOXY COMPOSITE A flywheel is being developed by KEVLAR kS EPOXY COMPOSITE Garrett-AiResearch to demonstrate a high energy density of 80 w-hr/kg in the rotor, RINGS which represents a significant increase in flywheel energy density. Because rotors with high energy density are smaller and lighter, this 80 w-hr/kg rotor promises to substantially reduce cost and weight of the installed unit. The rotor design consists of a multi- ring, subcircular, S-glass and Kevlar composite material rim that is mounted on a star hub of graphite composite material (see Fig. 1). The lightweight Kevlar com- posite is very strong for its weight and therefore has a high energy storage capa- city. SUBCIRCULAR -GRAPHITE COMPOSITE RIM Because of its multiring design, the CRUCIFORM HUt S-3M50 rim can accommodate the radial stresses Fig. ?. Flywheel rotor caused by centrifugal forces. Also, the rim was given a subcircular shape so that when placed on an oversized star hub, the SCOPE OF WORK resulting rim forces hold it onto the hub. This eliminates the stress associated with Two rotors will be designed and fabri- attaching the rim to the hub by ordinary cated for evaluation—an effort that com- attachment methods. prises six principal tasks. The graphite/epoxy composite material Task 1, Materials Selection and Hub hub has a small cross section and is very Fabrication Process Proof. A new rim epoxy liqht, yet can withstand the forces imposed is to be selected because the epoxy used on on it. The combination of high-energy- the previous AiResearch flywheels is no density storage in the rim and a very longer commercially available. A replace- light, strong hub yields an exceptionally ment epoxy has been recommended by T.T. high-energy-density flywheel rotor. Chiao of Lawrence Livermore Laboratories on the basis of composite material and will be utiIIzed as needed to confirm neat epoxy sample tests. (The epoxy is these tests. Dow DER 332, the diluent is Ciba Glegy RD-2, and the hardener Is UnlRoyal Tonox Task 2. Design Analysis. Design analysis 60-40.J The contract provides for the Includes an evaluation of the stresses fabrication and test of flywheel rings Induced In the flywheel rotor during fab- made from a composite of S-2 fiberglass, rication and operation, with an evaluation Kevlar, and the new epoxy. These tests of fatigue effects and scaling laws will be run to confirm the acceptability Involved In changing the size of the fly- of the new epoxy. The test ring Is shown wheel. The analysis treats the following: in Fig. 2. • Stress in the wheel caused by I—T = APPROXIMATELY winding, curing, thermal effects 0.080 IN. and assembly. • Stress during operation, both steady-state and transient. • Comparison of trlaxial material properties to the stress state, with particular emphasis upon source and reliability of mater- ial properties data. • Special stress conditions such as those at rim-hub interface. • Critical speeds and the effects of possible material creep upon operation and balance. 0.938 IN. Stresses caused by fabrication and assembly can be accurately defined, and the S30448 results of such an analysis will be used Fig. 2. Test ring for final assembly design refinements. The calculation of the triaxial stress The hub, to be made of a graphite- state within the hub will be accomplished epoxy composite, requires proof of fabri- as a guide to the tests of Task 1. The cation process and material properties results of Task 1 wilI be used to refine acceptability. The hub Is being fabri- calculations for the full axial length of cated by Structural Composites Industries the hub. Inc. (SCI) of Azusa, California. Graphite composite is chosen for this application Task 3, Hub and Rim Fabrication. Two because of its modulus of elasticity of assembled rotors, shown in Fig. 3, will be 18 x 106 psi In composite, because of its comp I eted. light weight of 0.05 Ib/cu in., and its ultimate flexural strength of .210,000 psi. The combination of these properties allows the reduction of hub weight from 20.5 Ib for an aluminum hub to 6 Ib for the graph- ite epoxy composite hub. Thus, the aluminum-hub rotor yields 53 w-hr/kg In the rotor and the graphite-hub rotor yields 80 w-hr/kg in the rotor. Task 1 will Include mechanical testing of hub samples. Hub testing will be conducted on three hub sections of 1-ln. axial length and will consist of compression testing across one opposite set of spokes to learn the location, mode, and load at failure. One hub will be cycled 100 cycles under load to a stress that wilI give an accel- erated, full-life test. The third hub Flywheel rotor assembly 173 The hub will be fabricated by SCI, Task 5, ProducibiIity and Cost Analysis. using the methods developed and proved The costs and processes involved in the during Task 1. The hub consists of multi- full-scale production of flywheels will ple slats, alternately bonded one upon be analyzed to provide data for accurate another to form a four-spoke hub with a cost tradeoffs of potential flywheel appli- rigid cruciform shape. (See Fig. 4). The cations. Cost-effectiveness is affected two-module rim will be similar to the one mainly by the design configuration, by qualified by AiResearch on the DOE'S Near- seal ing for economic use of materials, and Term Electric Vehicle contract. by effective tool ing and production tech- niques. Depending on the application, whether vehicular or stationary, or whether large or small, material tradeoffs will produce the greatest economy. Rotor speed, and therefore rotor size, not only affect the cost of the rotor, but very strongly affect the cost of housing and installa- tion. These relationships between size, speed, and cost will be presented in this task. ProducibiIity and tooling of uni- axial composite products are areas that require a great deal of development. Certain cost-effective processes are being appIi ed to the fIywheeI rotor i n Phase I to allow closer evaluation of producibiIity. Task 6, Final Report and Program Review. All work accomplished will be described in a final report which will include results 19 of (1) material evaluations, (2) material CRUCIFORM HUB MADE tests, (3) stress analyses, (4) flywheel OF 69 SLATS dynamic studies (critical speeds, creep vs Dalance, etc.), (5) producibiIity studies, and (6) Phase II test plan. Fabrication techniques and equipment will be described and flywheel outline drawings will be included. Photos will supplement the SLAT descriptions. Fig. 4. Composite Hub ACCOMPLISHMENTS TO DATE Task 4, Test Planning. A spin test of the The epoxy matrix material has been the rotor will be planned for Phase II. In qualified as a satisfactory replacement for this test, the flywheel will be suspended the previously used epoxy by hydrostatic on a quill shaft in a spin pit evacuated testing of flywheel rings. Fig. 5 shows to approximately 1 micron of Hg. The the hydrostatic test fixture used to deter- flywheel will be cycled 100 times between mine modulus of elasticity from tests of maximum and 33-percent speed. Deceleration ring stretch vs internal pressure. Ulti- will be accomplished at a shaft power out- mate strength also was obtained on the put rate of 37 kw by means of a reverse- test rings by bursting them in the fixture thrust nozzle on the air turbine. The and recording the burst pressure. Place- wheel will be examined after the test for ment of the test ring in the fixture is signs of fatigue fracture, material pro- shown in Fig. 6. Table 1 compares the perty changes, and any deterioration or results of the current and previous epoxy damage. matrixes in Kevlar 49. 174 Table 1. Comparison of Current and Previous Epoxy Matrixes in Kevlar 49. Parameter Current Previous Modul us of 14.23 x 106 13.37 x 106 elasticity, psi Ultimate 157,000 163,000 strength, psi Graphite hub slats and the hub assembly mold have been tested and evalu- ated for their ability to produce parts with Iimited stress concentrations, smooth faying surfaces, and good dimensional uni- formity. Inspection of preliminary samples produced by that tooling is under way. The design analysis has been com- pleted, and the results of that analysis have been applied to the hub tool ing. Specifically, the spoke/rim interface design has been analyzed to ensure that radial forces result in manageable trans- verse stresses in the rim. The analysis of mechanical and thermal stresses associ- Fig. 5. Hydrostatic test fixture ated with fabrication and assembly will a I low the composite hub to be mated with the composite rim to compensate for the TEST SPECIMEN operational deformations. Detailed analyses of stress concen- trations at the hub root have been made as guides to spin testing where actual stresses must be empirically determined. FUTURE WORK Upon the completion of the three 1-in. hub test specimens, they will be tested statically and cyclically to determine the basic slat quality and the quality of the slat-to-slat bond integrity. These tests results will be used to validate the analy- sis. Final adjustments, if needed, will be made to tool ing and processes in pre- paration for the fabrication of the 4-in.- Fig. 6. Hydrostatic test section wide hubs. 175 SESSION IV: SOLAR MECHANICAL 177 PROJECT SUMMARY Project Title: Solar Mechanical Energy Storage Project Principal Investigators: H. M. Dodd and B. C. Caskey Organization: Sandia Laboratories, Division 4716 Albuquerque, NM 87185 (505) 264-8835 Project Goals: The goal of this project is to identify and develop mechanical energy storage technologies for eventual commercialization, with the emphasis on solar (sun and wind) sources. The project utilizes results of systems analysis, special studies, and detailed design studies to define areas for hardware demonstration projects. The resulting cost and performance data will be available for considera- tion by industry. The project involves both Sandia in-house analysis and contracts with universities and industries. Project Status: System analysis has identified flywheel energy storage as the most promising of the mechanical energy storage technologies for resi- dential applications. Therefore, ongoing special study and de- tailed design study contracts (three with universities, two with industries) address various aspects of stationary flywheel energy storage. All will be completed late in FY 1979 when a decision on construction of a demonstration flywheel system will be made. Additionally, pneumatic (compressed air) industrial energy storage systems are being studied, along with some advanced concepts. Systems analysis, in addition to evaluation of proposed systems, is being used to investigate the relationship of various storage usage strategies with factors such as time-of-use pricing and energy sell-back. Contract Number: AT (29-D-789 Contract Period: Oct. 1977 - Sept. 1982 Funding Level: FY 1978-$285,000; FY 1979-$553,000 Funding Source: DOE, Division of Energy Storage, Advanced Physical Methods Branch 179 SOLAR MECHANICAL ENERGY STORAGE PROJECT I B. C. Caskey Sandia Laboratories Systems Analysis Division 4716 Albuquerque/ New Mexico 87185 ABSTRACT Plans for the Solar Mechanical Energy Storage Project are presented, and its current status is described. Systems analyses have identified flywheel energy storage systems as the most promising technology for small to intermediate load applications. Detailed design studies and special studies are under way to investigate this storage mode. Addi- tional analyses and pneumatic system studies are planned for FY79 as well as a decision on whether to construct a prototype flywheel energy storage system. INTRODUCTION ization, scope, and management philosophy of the Solar Mechanical In FY77 Sandia Laboratories Energy Project, followed by an ac- was assigned primary responsibility count of the status and plans of for the Solar Mechanical Energy each of the three project tasks. Storage Project by the US Depart- ment of Energy (DOE, formerly ERDA). ORGANIZATION Activities for that year included development of an optimizing sys- The DOE's Advanced Physical tems analysis computer program Methods Branch funds this project. called ENERA and subsequent analy- This Branch, directed by Dr. G. C. ses utilizing ENERA. H. M. Dodd, Chang, is in the Energy Storage et al presented a paper in Septem- Systems Division of Energy Tech- ber 1977 titled "An Assessment of nology. Figure 1 shows the Sandia Mechanical Energy Storage for Solar Laboratories project organization Systems."1 This paper and later and identifies the principals with results showed that flywheel sys- their functions. Sandia will de- tems are the most promising mechan- vote -2.5 man-years to this project ical energy storage devices for in FY79. residential applications. There- J. H. SCOTT fore, the primary emphasis has DIRECTOR OF been on development of low-cost ENERGY PROGRAMS stationary flywheel technology. X G.E. BRANDVOLD Activities for FY79 will include a SOLAR ENERGY PROJECTS more detailed investigation of | DEPARTMENT pneumatic energy storage systems MANAGEMENT H. H. DODD for industrial and residential SYSTEMS ANALYSIS applications, storage for agricul- DIVISION tural use, and an evaluation of advanced concepts. C. CASKEY [H7 E. SCK1LDKNECHT EM^ ANAl YM TECHNICAL CONTRACT STAFF MANAGEMENT PROJECT DESCRIPTION Fig. 1. Solar Mechanical Energy Stor- Described below are the organ- age Project Organization at Sandia Laboratories, Albuquerque, NM. Sandia Laboratories is a US Department of Energy (DOE) facility. This work was supported by the Division of Energy Technology, US DOE, under Contract AT(29-1)-789. 180 SCOPE This project is concerned with | PROJECT PHILOSOPHY | mechanical energy storage modes that may be used in conjunction with solar (including wind) inputs DETAILED MOUSING 'N DESIGN supplying small to intermediate WUCWONS J STUDIES loads. Since, for central station utility applications, energy stor- age would be tied to the grid rather than to a solar energy DEMONSTRATION > „ source, this project does not ad- PROJECTS J dress central station applications. Figure 2 lists the sources, storage modes, and applications to be con- sidered. [PROJECT SCOPE I TECHNOLOGY TRANSFER TO INDUSTRY FLYWHEELS . Fig. 3. Solar Mechanical Energy Storage Project Management Philos- SOLAR RESIDENTIAL ophy Leading to Mechanical Energy Storage Commercialization. This information flow cycle is WIND INTERMEDIATE expected to continue for other stor- age modes and applications until they have been either demonstrated or rejected. Additional advanced mechanical ADVANCED' energy storage concepts are ex- pected to be proposed, studied and I SOURCE| |MODE| | APPLICATION! entered into the development cycle as appropriate. Fig. 2. Solar Mechanical Energy Storage Project Scope. Figure 4, which presents the project milestone chart for the MANAGEMENT PHILOSOPHY next three fiscal years, reflects this philosophy. Positive deci- The goal of this project is to sions are assumed on both flywheel identi fy and develop appropriate and pneumatic system demonstration mechanical energy storage technolo- projects. Development of an inter- gies for eventual industrial com- face between a vertical-axis wind mercial uses. Figure 3 reveals the turbine (VAWT) and a flywheel energy underlying relationship among tasks. storage system (FESS) is also shown Based on existing technology, fly- as a likely candidate. wheel energy storage has been identified as the most promising TASK I. RESEARCH ON ADVANCED residential application; detailed CONCEPTS design studies are under way to generate credible cost and per- Systems Analysis. During FY78, the formance data* During the fourth systems analysis code ENERA was quarter of FY79, additional systems used to develop representative sup- analyses will use these data to ply and demand data for use in the form the basis for a decision on detailed design studies for fly- constructing and testing a demon- wheel energy storage systems. Both stration flywheel energy storage photovoltaic2 and wind3 energy system. These detailed design sources were used. Improvements to studies, in conjunction with sys- the code are planned in three areas: tems analysis, may identify collec- tor-to-storage interfaces that re- 1. Time-of-Day Pricing—utility quire detailed development. power costs will be a function 181 FY79 FY80 FY81 ONDJFMAMJJAS ONDJFMAMJJAS ONDJFMAMJJAS Task I. Advanced Concepts A. Systems Analysis (SLA) B. Special Studies (Contracts) 1. Advanced Pneumatic Storage 2. Cellulosic Flywheel 3. Variable Inertia Fly- wheel 4. Flexible Flywheel 5. Industrial Compressed Air Storage 6. Advanced Concepts Task II. Prototype Develop- ment 00 ro A. Detailed Design Studies 1. Contract Management 2. Residential PV/FESS 3. Residential Wind/FESS -A .RFP. 4. Intermediate-size FESS 5. Advanced Concepts .RFP B. Storage/Collector Inter- face RFP, 1. VAWT-FESS Interface C. Construct and Test 1. FESS (10 kWh) -A 2. Pneumatic ESS .RFP. 3. FESS (100 kWh) Task III. Headquarters Suppt Note: V Start A Finish Fig. 4. Solar Mechanical Energy Storage Project Milestone Chart of the time of day. Many utili- cepts are periodically forth- ties are proposing or adopting coming from individuals, uni- this technique to reflect their versities, and industry. Each production costs. The effect of concept must be examined to time-of-day prices on storage assure that no good idea is desirability will be studied. overlooked. All special studies funded under this pro- 2. Utility Sellback—energy sell- ject resulted from unsolicited back to the utility at some proposals judged to have fraction of the utility's sell- potential payoff. ing price may affect the use- fulness of storage, since the Special Studies. The following grid essentially provides special studies will be described storage capability. in detail in subsequent presenta- tions at this conference: 3. Storage Logic—with the im- provements listed above, the 1. Variable-Inertia Flywheel— disposition of energy from the David G. Ullman from Union source as a function of time, College in Sch'enectady, NY. state of storage, anticipated use, and predicted supply, 2. Cellulosic Flywheels—Arthur poses a significant modeling G. Erdman from the University challenge. Several storage of Minnesota in Minneapolis, strategies will be tried to de- MN. termine the appropriate level of sophistication. 3. Flexible Flywheels—John M. < Vance from Texas A& M in In addition to refinements in College Station, TX. the flywheel system analyses, the following areas will be explored There are two additional special subject to manpower constraints: studies: 1. Pumped Hydro/Wells—underground 1. Industrial Compressed Air— aquifers have been proposed for Jalar Associates is investi- storage by utilizing existing gating the feasibility of wells and adding a downhole utilizing solar sources to pump/generator and an above- generate compressed air for ground pond. storage and use by industries presently consuming significant 2. Industrial Compressed Air—Jalar quantities of compressed air Associates is studying the in- in their operations. The dustrial use of compressed air 6-month study began in Septem- as a power source as an appro- ber and will be reported in priate use of solar energy detail at the next contractor's (wind turbine) and storage. conference. 3. Agricultural Applications—the 2. Pneumatic Energy Storage Sys- BDM Corporation, under contract tem—Sandia has been negotia- to Sandia's photovoltaic (PV) ting for several months to project, is identifying agri- arrive at satisfactory terms cultural applications for PV for placement of a contract energy. One application that for a 12-month study of pneu- would require daily storage is matic storage systems and com- a dairy farm with two daily ponents. This study will also peak demands, just before and be reported in detail at the after PV energy is collected, next conference. which requires that essentially all the collected energy flow TASK II. PROTOTYPE DEVELOPMENT through storage. Detailed Design Studies. Two de- 4. Other Advanced Concepts—new sign studies are under way for storage or applications con- residential flywheel energy-storage 183 systems and will be described at and pneumatic systems are subjected this conference. They are: to detailed scrutiny. Advanced concepts and unique applications 1. Residential Flywheel with are constantly being evaluated for Photovoltaic Array Supply— further study. Francis C. Younger from Wil- liam M. Brobeck & Associates A major decision milestone in Berkeley, CA. late in FY79 will provide for con- structing and testing one or more 2. Residential Flywheel with Wind energy storage systems. Turbine Supply—Theodore W. Place from Garrett-AiResearch REFERENCES in Torrance, CA. 1. H. M. Dodd, R. E. D. Stewart, Depending on the results of et al, "An Assessment of Mechan- these contracts and additional ical Energy Storage for Solar anlyses, a request for proposal Systems," 12th Intersociety (RFP) may be issued late in FY79. Energy Conversion Engineering This RFP will request study of an Conference Proceedings, Wash- intermediate-sized system, perhaps ington, DC, August 28-September for an agricultural application. 2, 1977, Vol II, pp 1174-1180. Storage/Collector Interface. Based 2. B. C. Caskey, Supply and Demand on ongoing studies and analyses, Data for the Conceptual Design it is anticipated that an RFP will of a Residential Flywheel be issued during the fourth quarter Energy Storage System, Photo- of FY79 to develop the special voltaic Array Supply, Sandia hardware required to interface Laboratories, June 1978. storage to the collector, A VAWT interface to a flywheel energy 3. B. C. Caskey, Supply and Demand storage system is a likely candi- Data for the Conceptual Design date. Another possibility is a of a Residential Flywheel wind-turbine interface to an air Energy Storage System, Wind compressor. Turbine Supply, Sandia Labora- tories, July 1978. Construct and Test. Assuming fav- orable results from either or both the flywheel and pneumatic energy storage system studies now under way, RFPs will be issued to con- struct and test demonstration systems. This decision milestone will occur during the fourth quarter of FY79. TASK III. HEADQUARTERS SUPPORT Project personnel support DOE Headquarters by acting as techni- cal monitors for DOE contracts and by attending meetings and con- ferences. SUMMARY The Solar Mechanical Energy Storage Project is committed to developing appropriate mechanical- energy storage technologies for eventual commercialization by in- dustry. In the initial phase of this project, small-scale flywheel 184 PROJECT SUMMARY Project Title: Investigation of the Potential of a Variable Inertia Flywheel Energy Storage System Principal Investigator: D. G. Ullman Organization: Union College Mechanical Engineering Dept. Schenectady, NY 12308 (518) 370-6264 Project Goals: Study of dynamics of a Band Type Variable Inertia Flywheel (BVIF) with fixed-ratio power recirculation for rotational rate control. From these studies, determine if the BVIF concept offers potential advantages in the areas of economy, reliability, and efficiency. Project Status: The dynamic modeling of the BVIF less its fixed-ratio power re- circulation system for controlled inertia change was completed and documented. Computer simulation studies were initiated to analyze the dynamic characteristics of this truncated configu- ration. Expansion of the dynamic model to include the power recirculation system was started. Parts fabrication/procurement for a small scale proof-of-concept BVIF was completed and assembly work initiated. Contract Number: 07-3662 Contract Period: June 1978 - June 1979 Funding Level: $25,325 Funding Source: Sandia Laboratories, Albuquerque 185 THE BAND TYPE VARIABLE INERTIA FLYWHEEL AND FIXED RATIO POWER RECIRCULATION APPLIED TO IT C. G. Ullman Mechanical Engineering Department Union College Schenectady, New York 12308 ABSTRACT A flywheel with variable moment of inertia created by thin bands winding about a central hub in a hollow casing is introduced. This mechanism combines the functions of energy storage and power control. Configurations with the moment of inertia change internally controlled through fixed gearing are introduced. INTRODUCTION Due to its very simplicity, the If the equation for stored energy is concept of storing energy in the rotating reconsidered, another approach to the mass of a flywheel seems very attractive. problem becomes evident. It may be A closer look, however, reveals subtleties possible to alter the moment of inertia of which significantly complicate the imple- the flywheel to gain added control of the mentation of an operational system. release of stored energy. The resulting Primary among these complicating factors mechanism is called a Variable Inertia is the difficulty of transferring the Flywheel (VIF). This mechanism is an energy to and from the energy storage fly- energy storage device whose output can be wheel . controlled to meet the load requirements. The governing energy relation for a THE BVIF flywheel is Potential Variable Inertia Flywheel E = %Iw2 (1) Designs fall into two classes, fluid and mechanical1'2. The configuration which where E is the energy content of the fly- seems to have the qualities necessary for wheel, I is the moment of inertia of the a successful system is the coiled band flywheel about the axis of rotation and u) variable inertia flywheel BVIF which is is its angular rate. Thus in order to shown in Fig. 1. This flywheel is retrieve stored energy from a flywheel the composed of a hollow outer casing and a rotational rate must decrease. This separate central hub. Connecting these is decrease in rotational rate is not desir- a band of some flexible material mounted able as systems being powered by flywheels as the mainspring in a watch. Centrifugal usually require a rotational rate which is force due to rotation pushes the band to constant or even increasing. This mismatch the outer edge of the hollow casing. As in rotational rate is usually compensated the central hub is rotated relative to the outer casing the band is wound onto the for by the use of a variable ratio hub lowering the moment of inertia of the transmission such as a traction drive, a system. What controls the state of the hydraulic pump and motor, or an electric band is a balance between the centrifugal generator and motor. Each of these force on the band, the angular rate dif- systems has drawbacks either in terms of ference between the central hub and the cost, reliability, or most importantly, casing and the torque flow through the efficiency. Specifically, the electric mechanism supplemented by the torque and hydraulic transmissions (the most variation caused by the change in state of technically developed) may be inefficient the band itself. Work on this concept has and costly enough to hamper the practical appeared since an initial patent in 19653. development of commercial flywheel energy Much of this work has been in the USSR storage systems. 1S6 -1-53 with two patents and one published A VIF with the power for inertia article in English6. However the authors change coming from an outside source is have been unable to find any dynamic much like a mechanical power amplifier. analysis of such a mechanism in the open This configuration is the primary consid- literature. The dynamic balance of eration of the Russian patents*5 and is torques, forces, and momentum is surely not discussed further in this paper. complex. Understanding of the dynamics is required to fully understand the potential When one unit of power is fed back to of the device. change the moment of inertia, the power is said to recirculate within the VIF. It would be most desirable to have this recirculating power flow for inertia change to be completely controllable allowing for differing inertia change requirements. Unfortunately, complete controllability of the recirculating power can only be had with a variable ratio transmission such as that which the VIF is proposed to eliminate. Also, it was shown that with constant rotational rate loading the amount of power to be recirculated through the variable ratio transmission would be the same as that output. Thus it appears that the VIF with a variable ratio transmission taking power from the output shaft to change moment of inertia would have no better efficiency than a fixed inertia flywheel with the variable FIGURE 1 ratio transmission on the output shaft. BAND TWE VARIABLE INERTIA FLYWKEEL However, complete control of the recirculating power may not be required The major appeal of this configuration for most tasks as the nature of the load- over others presented in References 1 and ing is usually reasonably well known and 2 is in its high energy density compared slight deviations from the desired are to other designs and that the moment of usually tolerated; thus it may be possible inertia change is accomplished in a rota- to create a fixed ratio power recircula- tional manner. This latter aspect of the tion path to meet the requirements of a BVIF is very important as all other mech- specific load. Thus use of the term anical configurations considered1 require "fixed ratio power recirculation" implies linear motion to change moment of inertia that the mechanism carrying power to such as moving a mass on a radial spoke. create the inertia change is of fixed The transmission of powe~ from a high rpm geometry or at a maximum countably few rotating source through a mechanism to variations in geometry (eg: a fixed ratio cause a small distance translation motion transmission). The result of this VIF is very inefficient. with a fixed path, is a flywheel with fixed properties which differ from those FIXED RATIO POWER RECIRCULATION of a standard fixed inertia flywheel. Another way of stating this is that now In References 1 and 2 it was shown that the moment of inertia is a function of for any variable inertia flywheel rotating angular rate so that the torque is now at constant rpm it takes one unit of power totally a function of angular rate but to change the moment of inertia for each different from that of a standard fixed two units of power exchanged between the inertia flywheel. VIF and an external system. Thus if a VIF gives up two units of power, one must go PROJECT DESCRIPTION to change the moment of inertia while one to power the load or in an alternate case The two concepts of the Band Type one unit of power can come from an outside Variable Inertia Flywheel and Fixed Rat*? source to change moment of inertia and Power Recirculation are not independent then two units can go to the load. as, the potential results of Fixed Ratio Power Recirculation are dependent on the 187 characteristics of the BVIF considered. Thus the project is organized to first BAND MATERIAL determine the dynamics of the BVIF and then apply fixed ratio power recirculation to it. The work on this project is being approached by first modeling the dynamics of the system and then building and test- ing a proof-of-concept model to check the adequacy of the mathematics. Thus the four resulting tasks are: 1. Dynamic Modeling" of the Band Type VIF (BVIF). 2. Construction and Testing of a Proof-of-Concept Band Type VIF. 3. Dynamic Modeling of Fixed Ratio Power Recirculation as applied to the BVIF. 4. Construction and Testing of a Proof-of-Coneept BVIF with Fixed Ratio Power Recirculation. These tasks constitute a program which FIGURE 2 is,at this writing, less than one half BVIF PLAN VIEW complete. The first task is finished and Each of the n bands is of thickness its results are reported in this paper. h, width b, and density p. The other tasks are scheduled for completion by May 1979. The moments of inertia of this system are considered in two parts, that BAND TYPE VIF DYNAMIC EQUATIONS for the outer casing and the band material rotating at u , hereafter called I , and Consider the Band Type VIF composed that for the inner hub and material rotat- of a band(s) in an inner casing connected ing at angular rate oij, hereafter called to a central hub as shown in Fig. 2. I.. Included in these moments of inertia mist be the non-variable, fixed moments of The inner radius of the casing is r,. inertia of the outer casing I and of the The inner radius of the band material hub I. °f pressed by centrifugal forces against the wall of the casing is r,. The outer radius of the material wound about the Thus inner hub is and the radius of the h = inner hub is r... It is assumed that only and a small amount of material is floating between the inner and outer portions. There is experimental evidence to support this assumption. The dependent variables, the moments of inertia, are functions of the radii r. The outer casing is rotating at and r.. These two radii are not indepen- dent with their relation fixed by the angular rate (i)o and the inner casing at rate U). in the same direction. The total amount of material in the casing. difference between w. and to causes the With n bands of length L and thickness h, material to be moved from the inner to the total volume of material in the VIF the outer portion or vice-versa thus is changing radii r. and r,. Radii r, and r^ are constant, fixed by the solid V = nLbh. (A) geometry of the hub and casing (except for expansion due to the forces of rota- This volume is distributed as two tion). hollow cylinders, one against the outer casing and one around the inner hub or 188 (5) BVIF. The second of these utilizes Fixed Ratio Power Recirculation and is one of The small amount of material between many fixed ratio systems being studied. radii r2 and r3 which is not accounted for is insignificant when compared to the The VIF as modeled has effectively total volume of material. Equating (4) two shafts leaving the system. One shaft and (5) and writing r, as a function of is connected to the outer race, is rotat- then ing u>Q and has an external torque on it of T (positive in the direction of rotation). (6) Tne other shaft is connected to the inner hub, is rotating at angular rate u>. and Thus with Eq. (2), (3) and (6) the has an external torque on it of T. moments of inertia of the system can be (positive in the direction of rotation). written as a function of radius r^> But, The loads on these two shafts are indepen- r_ is a function of the time history of dent. tne angular rates u and w.. The simplest loading (called a Type I Although the mass of the material system) which will be considered is one between the inner and outer band masses where the band begins at time zero in the has been ignored it does act to transfer natural state of being pressed entirely power between the material wound around in the outer casing which is rotating at the hub and that against the outer casing. angular rate w and has no load torque on The torque which transmits the power, T , it (T = 0) an8 the inner hub is, at time tries to accelerate the inner hub and zero clutched to a load. In other words retard the outer casing. This torque is the outer casing is free and the shaft given by,* from the inner hub is connected tc the load. The load torque is represented by With the external torque on the inner hub (10) as T. and that on the outer casing is T the equations of motion for the BVIF are: where T = 2— (8) = -T±, (11) t dt and Ty is dry friction, I is the moment of inertia of the load and T^ is an aero- T - T = ^^ (9) dynamic drag coefficient. This config- o t dt uration is shown in Fig. 1. To complete the description of the A type of BVIF with Fixed Ratio Power BVIF system, the characteristics of the Recirculation has the load again attached external systems which relate T , T., w , to the shaft from the inner hub, but now 0 and to. must be derived. These external also has the outer casing connected to systems determine the exact configuration the inner hub by the band and by a gear of the BVIF as it is those items outside set. Thus in this configuration, here- the BVIF itself which cause the BVIF as after referred to as Type II, the angular a power amplifier or a flywheel with rate of the inner hub shaft and the outer Fixed Ratio Power Recirculation or in some casing are related in some fixed manner. other manner. A sketch of this configuration is shown in Fig. 3 where the gear set is an There are many potential ways of epicyclic. This form of gearing has been connecting the two shafts to an external chosen as the ratio between w. and 0) system (1,3,5). Presented here are 2 needs to be close to unity. The exact types of loading techniques. The first value of the gear ratio depends on the of these configurations is what is con- nature of the loading and the BVIF config- sidered the most basic loading of the uration as a gear ratio close to unity will cause the band to transfer slowly and a ratio far from unity will cause a Derivation to appear in a forthcoming rapid transfer creating a higher torque paper. at a given rotational rate. For this 189 configuration with the gear ratio, g, and These equations along with those describing the load rotational rate equal to the the dependence of I , I , and T on the inner shaft rate then band state and angular rate history constitute a set of nonlinear differential (12) equations which are stepwise solved on the computer. -T.-T /g Also T.L o (13) Input for each run of the simulation where T is as defined in Equation 10. LJ must be: Material Density, Material Width, EPICYCLIC GEAR Material Thickness, SET Number of Bands, Length of Each Band, Inner Hub Radius, Outer Casing Inner Radius, Fixed Inner Moment of Inertia, Fixed Outer Moment of Inertia, Initial Inner Radius of Band Material, Initial Outer Angular Rate, Load Friction Torque, Load Aero Torque Coefficient, and Load Moment of Inertia. For each run to be discussed in this paper the band material is steel, density = .00073 lb sec2/in, it is 1 in wide, .005 in thick, and there are 2 bands each optimally 324 ft long.* The inner hub has FIGURE 3 .5 inch radius and the outer casing has 5 inch internal radius. The above FIXED RATIO POWER RECIRCULAIIOH BAND geometry is that which is used in an TYPE VARIABLE INERTIA FLYWHEEL experimental apparatus being constructed. In these simulations the mechanism has an Unfortunately whichever type of BVIF/ estimated fixed inner moment of inertia load system is chosen the resulting differ- of .01 lb-sec^-in and an estimated outer ential equations are highly nonlinear. fixed moment of inertia of .1 lb-sec^-in. Attempts to linearize them have proven unacceptable in that the characteristics of the system have been lost. Thus the The initial conditions which must be equations have been programmed on a set for each run are the initial inner digital computer and step wise integrated angular rate, the initial outer angular using a Runge-Kutta technique. rate, and the initial position of the band material. This latter condition is given in terms of the initial value of the BVIF SYSTEM SIMULATION radius of the band material wrapped on the inner hub. Thus .5 inch would imply no To demonstrate the gross nonlinearity material is around the hub and all the of the equations representing a BVIF material is in the outer casing. This is system the Type I system will be analyzed. the assumed usual initial band condition. With T = 0 Equations 8 and 9 reduce to — (14) Ti+Tt dt and * "optimally" infers that the moment of inertia of material in the BVIF is such (15) that the difference between the maximum _x = — moment of inertia and the minimum moment where from Equationt dt s 10 and 11 of inertia is as large as is physically possible. The derivation of this result (16) will appear in a forthcoming paper. 190 As an example of the performance of history Fig. 5, shows the band transfer- the BVIF the outer casing initial angular ring from the outer casing to the inner rate was set at 100 rad/sec (995 rpm) and hub in 11 seconds ending the run. During the initial inner hub angular rate was this time the BVIF only releases about 25 varied below these outer rates. These percent of its energy to the load as shown initial conditions result in initial in the energy time history plot, Fig. 6. stored energy of about 3200 in 1b (267 ft The torque provided by the BVIF to the lb, .1 watt hr). The load conditions were load increases steadily with time as shown set with T= 0 (aero drag torque) and Tp in Fig. 7. (load constant friction torque) and I (load inertia) as variables. If the aero drag torque were not zero it would combine with the inertia to represent an essential- ly constant torque throughout acceleration and steady state operation. Thus varying only the constant load torque and its inertia are sufficient to understand the operation of the Type I BVIF. The base- line values used for the load variables are T^ = 1 in-lb and I = 1 lb-sec^-in. If the baseline load were connected to a fixed inertia flywheel of the same initial inertia and energy content as the EVIF (rad/uc) and rotating at 100 rad/secs the resultant angular rate time history of the load would be: 0) = 100 - .66t(rad/sec), FIGURE 4 a constant deceleration with u = 0 at 150 Load angular rate time history with seconds. This behavior can be compared to Initial outer casing angular rate of 100 that of the BVIF operating with different rad/sec and variation of Initial load inner hub initial angular rates, w. . The (inner hub) angular rate. Tf = .108 N-m cases to be considered are: o (1 in-lb), I. = .108 N'm-sec2 (1 1n-lb-sec*). Case (rad/sec) t (sec) > 20 11 N 40 15.5 D P 60 57 0 5 80 120 I T 95 132 I 0 H Note that the last column is the total time it took the band to reach one of its limits in winding or unwinding. In all data plots which follow, time is normalized with respect to this total run time. This makes visualizing the effects of parameter changes somewhat easier. .4 .6 HOUWLIZED TIME t/t» For Case 1, the initial angular rate FIGURE 5 of the load is only 20% of that of the Band Position time History with initial outer casing containing the band material. outer casing angular rate of 100 rad/sec As can be seen in Figure 4 the angular rate remains essentially constant through- and variation of Initial load (inner hub) out the first 2/3 of the run before angular rate. Tp = .108 N-m (1 in-lb), 2 2 increasing at the end. The band material IL = .108 N-m-sec (1 in-lb-sec ) 191 Physically, what is happening in case 1 is that the torque load is approximately equal the torque transfer- red by the bands, T of Eq. 7. This torque is produced By the reduction in moment of inertia of the band material in the outer casing. Thus the slower inner hub produces the torque T by causing the material to wind on the hub. This torque also balances the loss in angular momentun of the outer casing which is primarily due to change in moment of inertia while angular rate holding near constant. In other words the torque demand by the load causes the inner hub to resist accelera- tion resulting in a mass transfer to the («tt-hr) hub. This transfer increases the momentum of the hub and decreases that of the casing. Angular rate changes occur to keep the changes in momentum balanced with the torques. 0 .2 .« .« FIGURE 6 If the initial angular rate is increased to 40 rad/sec (case 2) there is BVIF Energy content time hi-story with very little change in the behavior of the Initial outer casing angular rate of 100 BVIF. The run now takes 15.5 sec and 53% rad/sec and variation of initial load of the energy is given up (most at the (inner hub) angular rate. Tp • .108 N-m end of the run). However, if the initial 2 2 (1 in-lb), IL ' .108 N-m-sec (1 in-lb-sec ) angular rate is raised to 60 rad/sec as in case 3 the behavior becomes drastically different. In case 3 the angular rate starts to decrease then increases and again decreases (Fig. 4). What is happening can be explained by examining Equ. 14. When t o the angular rate slope changes from a • 0 positive to negative or vice-versa then v I.(I). = 0 so that Equ. 14 can be rewritten t wxtfe T. = -TT as T i IJ 0 s (17) D 'l From equation 7 (•••] 2 2 2 2 T -(positive terms)(r, w -r to ).(18) t Thus Normalized Time t(positive terms) ?] -T ] (19) FIGURE 7 1 L Torque to'load time history with Initial The first reflex in the curve is outer casing angular rate of 100 rad/sec caused by a balance of the terms in Equ. and variation of Initial load (inner hub) 19. The second reflex is more interesting as the terms in Equ. 19 cancel leaving angular rate. TF • .108 N-m (1 1n-lb), 2 2 I. = 0. This can be seen in Fig. 5 where I, - .108 N-m-sec (1 1n-1b-sec ). tfte band stops winding onto the inner hub and reverses itself winding back into the outer casing. What is occurring is that the combination of the 192 2 2 r 0) and T terms have become greater than sec to decay. Cases 4 and 5 take 120 and 2 i L 132 sec respectively. Thus it can be 2 2 concluded for the loading used that at r U . With r~ increasing at a rate great- high initial angular rates of the hub the er ?han r_ and UK near the value of to BVIF acts like a standard flywheel. With this reversal can occur. lower initial angular rate the band winds After the change in the winding on the hub faster resulting in lower over- direction of the band the BVIF continues all run time, tn. to give up energy to the load (Fig. 6) K while the angular rate slowly decreases. As can be readily seen the behavior of the BVIF is variable depending on the It must also be noted in comparing initial status. Simulations have also Case 1 and Case 3 that the outer casing been performed with the load conditions as angular rate increases in Case 1 but variables. In these simulations certain decreases in Case 3. This accounts for loads result in a BVIF which gives up the higher energy release from Case 3 energy to the load while accelerating it. where wo final is only 26 rad/sec compared This implies that future work should lead to 139 rad/sec for Case 1. The reason for to a BVIF with truly usable characteris- this is that in Case 1 the torque from the tics. band, Tt is higher than in Case 3. In Case 1 this torque is higher than the FUTURE WORK angular momentum variation caused by inertia variation thus the angular rate is The above results are a part of a caused to increase to make up the differ- continuing study and in themselves ence. incomplete. Simulations are continuing on various fixed-ratio-power-recirculation As can be seen, explaining the causes systems such as the Type II configuration of the variations within the BVIF is mentioned earlier. It is anticipated that quite difficult with the continuous further simulation of these systems will balance of momentum change along with enhance the understanding of BVIF dynamics energy conservation determining the time history. If the initial inner hub angular As the proof of the viability of any rate is raised to 80 rad/sec, Case 4, then concept is in hardware a BVIF is being the angular rate decreases steadily as constructed with the dimensions used in shown in Fig. 4. While this is happening the computer simulation presented. This the band material starts by winding in, proof-of-concept model is designed to be reverses itself and starts by unwinding, configured as many different types of then reverses itself again and continues systems and can be inertially or friction- to wind in. In this case the relation ally loaded. between the band torque, T , and the load torque, T , is such that they change I. REFERENCES twice in Equ. 19. This configuration x releases most of its energy to the load as 1. Ullman, D.G., "A Variable Inertia Fly- does Case 5 with an initial angular rate wheel As an Energy Storage System," of 95 rad/sec. For both of these cases Ph.D. Dissertation, Ohio State Univer- the output is essentially that which sity, March 1978. would be of a regular fixed inertia fly- wheel. Namely, with a regular fixed 2. Ullman, D.G. and Velkoff, H.R., "The inertia flywheel the equation of motion Variable Inertia Flywheel (VIF), An would be Introduction to its Potential," 1977 Flywheel Technology Symposium Proceed- 10,. = -TL = -T (20) ings, Oct. 5-7, 1977, San Francisco, California. With I as the total initial inertia of the 3. Durouchoux, 0., U.S. Patent #3,208,303, BVIF, the solution to this differential 1965, U.S. Patent Office. equation is 4. Gulia, N.V., "Centrifugal Accumulator," - .62t (21) CO. USSR Patent #1131894/25-28,November 1969. 100 rad/sec then it Thus if i would take the fixed inertia flywheel 161 193 5. Gulia, N.V., "Variable Moment-of- Inertla Flywheel," USSR Patent #1182724/24-27, July 1969. 6. Gulia, N.V., "A Coiled Band Mechanism for the recovery of a Vehicle's Mechanical Energy," Jnl. Mechanisms, vol. 3, pp. 113, Pergamon Press. 194 PROJECT SUMMARY Project Title: Development of Cellulosic Flywheel System Principal Investigator: Arthur G. Erdman Organization: University of Minnesota Dept. of Mechanical Engineering 125 Mechanical Engineering 111 Church Street, SE Minneapolis, MN 55455 Project Goals: 1. To measure the long-term design strengths and fatigue properties of flywheel and rotors under vacuum conditions. 2. To develop self centering methods to assemble and balance rotors. 3. To develop a flywheel demonstration model. Project Status: The first quarter of the 12-month period project has been completed; experimental apparatus is being designed and built. Contract Number: 07-9072 Contract Period: Aug. 27, 1978 - Aug. 24, 1979 Funding Level: $41,239 Funding Source: Sandia Laboratories, Albuquerque 195 CELLULOSIC FLYWHEELS A. G. Erdman Associate Professor of Mechanical Engineering D. L. Hagen Graduate Research Assistant D. A. Frohrib Professor of Mechanical Engineering W. L. Garrard Associate Professor of Aerrrpace Engineering and Mechanics University of Minnesota 111 Church St. S.E. Minneapolis, Minnesota 55455 ABSTRACT Natural cellulosic materials have been shown to have moderately high tensile strengths, low densities and thus reasonably high specific energies. Thus they appear to be technically feasible and economically competitive for stationary flywheel energy storage applications where volume and total mass are not in question. An overview of previous and current research on cellulose flywheels is given. It has been found that one to two ton birch plywood rotors could be readily assembled and eight ton kraft paper rotors are available. Even larger rotors can be made using existing technology. The birch plywood rotors can be made at a cost of 4 wh/$ and the kraft paper rotors are available at 11 wh/$, with further improvements in costs predicted. Intrinsic specific energy of 294 kJ/kG (37 wh/lb) has been demonstrated for kraft paper. The objectives of current research are to establish the working strengths of cellulose in a vacuum as a function of moisture content and the cyclic fatigue expected in the rotor; to develop methods to attach hubs and balance cellulose rotors; and. to study the fatigue of repre- sentative rotors and hubs under actual cyclic operating conditions. Other objectives include the vibrational problems during operation, and the design and implementation of controls for the flywheel system. The efficiencies and use of continuously variable traction drives are also being studied. INTRODUCTION Table 1) The major difficulty with natu- ral materials is the distribution of Natural cellulose fibers have very defects and variation in properties from high tensile strengths on the order of one sample to another. By laminating 1.0 GPa (150,000 psi). Although the ac- multiple layers of veneers to make ply- tual strengths of clear wood are only wood, these effects can be minimized and about 160 MPa (23,000 psi), when combined averaged out to retain the high strength. with a low density of around 700 kg/m3 The effectiveness of this can be seen in these strengths result in a surprisingly the widespread use of glued laminated high intrinsic specific energy of around beams and plywood "I" beams for structu- 230 kJ/kg (29 wh/lb). The natural abun- ral use (replacing steel). Similar bene- dance of wood and the well established fits are obtained by separating the fibers industries produce comparatively low and reforming them as paper or hardboard. prices. These result in intrinsic energy storage per cost values of natural cellu- The possibility of using cellulosic losic materials competitive with high materials to make economically competi- performance expensive materials such as tive flywheel rotors was first recognized Kevlar, fiberglass, and graphite. (See by Rabenhorst in 1972.(1) Cellulosic 196- rotors appear to be suited to stationary practical problems of assembly and applications where economics rather than balancing raised by the initial research mass and volume are the governing factors. are beina pursued in current research The possibility has been mentioned in under continued funding from the Minne- several papers since then.(2) Tests of sota Energy Agency. We propose to scale wood rods and plywood disks spun to de- up to a 50 kg rotor mounted on a vertical struction at the Applied Physics Labora- axis. Some vacuum freeze drying equip- tory have confirmed the potential of using ment is being renovated to compare the cellulose in rotors.(3) Some similar strengths of oven and vacuum freeze dried tests are being performed at the Technical material using conventional methods. Institute of iurin. (4) A review of the Rotors under actual operating conditions literature revealed no work had been done may become even dryer. A chamber is to assemble and test full scale rotors being connected to a high vacuum pump to from wood. (2) dry samples out under higher vacuum con- ditions to see if this has any further Research was therefore begun under a effect on the strength. seed grant from the Minnesota Energy Agency to assemble a small demonstration The key question is still what is model of a flywheel energy storage system the design strength of cellulose under using a plywood rotor. (5) Figure 1 the sustained stress and cyclic loading shows a schematic of the demonstration conditions in a vacuum experienced by a model which houses a 14 inch diameter flywheel rotor. Studies of the basic (14 inch length) flywheel that rotates at stress strain phenomena are being pre- a maximum of 5,000 rpm. This pilot study pared under funding from Sandia Labora- is being used to gain familiarity with the tories. Measurements of the rate of plas- practical problems of assembling and tic deformation and the acoustic emissions balancing wooden rotors. Consideration will be made as functions of the stress was also given to the task of supporting and strain under dry conditions. the rotor and the associated vibrational Attempts will be made to correlate these problems anticipated in the operating with the duration of load effects. Vis- regions. Both electronic (Parajust) and coelastic effects compound the problem mechanical (a Zero-Max) continuously vari- and require high resolution measurements. able speed controls are used to transmit A PDP11-03 with a 14 bit (4% digit) ana- the power to and from the rotor. The ply- log to digital converter has been obtained wood rotor weighs 28 kg (62 lbs.). A for these tests. The acoustic emissions description of the system has been pre- up to 100 kHz will be recorded on tape sented elsewhere. (6) and digitized at slower speeds for anal- ysis. Data in the literature indicate CELLULOSE ROTOR that the rate of plastic deformation and acoustic emissions begin to increase A more detailed review of the liter- rapidly around the 50% of ultimate stress ature was made to assess in greater depth level which seems to correlate with the the potentials and difficulties of manu- long term strength. Measurements of the facturing large changes with the moisture fatigue strength in tension under dry content. (7) Very little data exists on conditions are also planned. the performance of cellulose under the vacuum conditions required for high speed An elastomeric interface between the rotors. Extrapolations from existing metal hubs and the plywood disks has been data indicate that wood becomes more found necessary to avoid shear failure brittle, and the strength and toughness between them under operation. (3) A apparently decrease as it dries out. cyclic rotary fatigue test bed is being Elevated temperatures in conventional designed to test the durability of the oven drying or kiln drying of wood hub to rotor attachment as well as that decrease the strength somewhat. Conser- of the rotor itself under actual opera- vative estimates of assembly costs con- ting conditions and fatigue regimes. A tinued to indicate that plywood rotors continuously variable drive will be used would be competitive with Kevlar or fiber- to cycle the energy between two identical glass rotors in stationary applications. disks. (7) The questions of the strength of wood at low moisture content and the 197 COSTS rotors are presently possible using com- mercially available materials. It is The large commodity markets of ply- projected that higher strengths can be wood and paper suggest that large flywheel obtained by minor modifications in opera- rotors could be made at low cost. Further ting conditions. Leboratory tests of research was conducted this summer (sup- highly oriented paper have demonstrated ported by the Applied Physics Laboratory) intrinsic specific energies of up to 294 on the economics of manufacturing flywheel kJ/kg (37 wh/lb). (8) Methods would need rotors. Discussions with several plywood to be developed to be able to approach manufacturers indicate that cylindrical these values in commercial quantities. rotors can easily be assembled from ply- The actual design strengths for flywheel wood disks using existing equipment. A rotors again are unknown and need to be one ton flywheel 4 feet in both diameter established by experiment. Of particular and height would cost $470 to assemble concern is the higher creep rates exhib- one hundred units, assuming 30% for over- ited by composite cellulose materials. head and profit. The material cost of Details of wrapping and supporting such rotors and the balancing and vibration birch plywood is $650 (including 23% problems would also need to be worked out. waste). This results in a storage cost on the order of 15.8J/$(4.3 wh/$). These actual quotes on assembly costs are half When costs of assembled rotors are those previously assumed. (7) The birch compared, cellulose rotors appear to be plywood is available in 1.5 m x 1.5 m (5 surprisingly competitive with high per- ft x 5 ft) sheets indicating that a two formance filament rotors, even in large ton flywheel 5 ft. in diameter and height volumes. The prices of graphite fibers could readily be assembled by any plywood are predicted to drop down to $23/kg by assembler having this size press. Reduc- 1985. The glass and Kevlar filaments are tions in assembly costs are possible with expected to increase with inflation higher production runs. unless large unforseen markets develop. Costs are for both the bare filamert Preliminary studies of the potential wound rotors being developed by of using kraft paper to assemble wound Applied Physics Laboratory and for the rotors also looks promising. Four ton composite rotors being developed by rolls are readily available at most mills, others. Current costs of paperboard are and one mill makes 8 ton rolls 315 in. obviously competitive with all the fila- long. Costs of lltf/lb to 17<£/lb in mill mentary rotors shown. Substantial in- lot quantities suggest that storage costs creases in the performance of paper are of 11 wh/$ in full scale cylindrical potentially possible, resulting in Table 1. Energy Storage Potentials and Costs Material Density Strength Fatigue Specific Material Labor Storage e P r. 0 Factor^ Energy b Costs Costs Costs kg/nT MPa f kJ/kg $/kq $/kg $/MJ Birch Plywood 700 120 0.6 17.6 0.65c .30-.45 54-63 Paperboard 1000 75 0.4 7.4 0.22 .03-.10, 34-43 Superpaper 1100 336 0.4 33. 0.22. .05-.30T 8-15 E-glass/epoxy 2100 1100 0.3 38.5 1.10° 1 .-10. 55-3409 9 S-glass/epoxy 2000 1750 0.3 64.3 4-50d 1.-20. 86-257 Kevlar 49/epoxy 1320 1800 0.6 200. 20.005 1 .-10. 105-1609 Graphite/epoxy 1520 1500 0.5 120. 70.00 7.-10- 590-680s Estimated for diurnal cycling for 30 years (10,00,0 cycles). eAssembly costs of winding rotors are for baKfS/p: K = Shape Factor = 0.35, S = bare filament rotors and for epoxy resin- Safety Factor = 0.70. filament composite rotors. cIncludes 23% waste. Not commercially available. dCost estimated by P. Ward Hill, graphite 9Epoxy resins estimated at $2.00/kg. prices to drop 65% by 1985, all others to rise with inflation. 198 significant increases in overall storage lity of connecting the variator directly costs. The major drawback of cellulose between the flywheel and the input power rotors is their low density and large source or output power generator without volume. This can be used to a benefit when shaft intermediate step up-step down cou- they are used as cores or hubs for com- plings. posite rotors using higher performance materials. One of the goals of the future test station will be to determine the suitabi- CONTINUOUSLY VARIABLE TRANSMISSION lity of the Cleveland variator as the cou- pling element between the flywheel and An efficient, low-cost means of apply- external components. Tests are planned to ing energy to and retrieving energy from determine 1) the efficiency of the varia- a flywheel presents a major hurdle that tor under typical operating conditions, must be overcome in the development of a and 2) the durability of the variator un- practical flywheel energy storage system. der the power, speed, and cyclic loading In order to input energy to a flywheel, conditions of a flywheel energy storage the rotational speed of an input shaft, system. If the Cleveland variator proves which must apply torque to accelerate the suitable, a prototype control system to flywheel, must be matched to the flywheel automatically adjust the CVT to match real shaft speed. This problem is magnified operating conditions will be developed. when the input shaft is driven by a power If the CVT exhibits problems in the fly- source such as a wind turbine, which in- wheel application, the feasibility of mod- herently possesses large speed fluctua- ifying the CVT to operate in this special- tions. A parallel problem occurs at the ized application or employing alternate output. The most useful final output of mechanical CVT systems, such as variable- a flywheel energy storage device is 60 Hz, pitch-pulley or free ball geometries, will AC electric power. Since constant AC be examined. frequency, variable shaft speed genera- tors are not yet available, the varying FLYWHEEL CONTROL SYSTEM speed of the flywheel shaft must be buf- fered to drive a constant shaft speed AC The control system for the flywheel generator. Therefore, an electric, hy- has been designed to accomplish two objec- draulic, or mechanical transmission capa- tives. These are: ble of fulfilling these restraints is a 1. Maintain the output angular vel- necessary component of such flywheel ener- ocity of the continuously variable trans- gy storage systems. mission (CVT) constant as the input velo- city from the flywheel varies. The current flywheel demonstration 2. Control the input velocity to the model at the University of Minnesota flywheel so as to simulate a typical duty (Figure 1) uses a mechanical CVT manu- cycle. factured by the Zero-Max Corporation in its output drive train. This CVT has The first object is accomplished by a exhibited promising performance, but it relatively simple feedback control system. is not presently available in sizes over The control system consists of a tachome- Ts HP. A larger flywheel test station ter, a DC motor (See Figure 2), and asso- now being designed will tentatively incor- ciated electronics. The control system is porate a 12 HP CVT manufactured by Eaton shown in functional block diagram form in Corporation. This particular transmis- Figure 2. The output angular velocity of sion, known specifically as the "Cleve- the CVT is measured by an analog tacho- land speed variator", is probably the most meter. The output of this tachometer is a popular all-metal traction drive system voltage proportional to the angular velo- presently available in the United States. city output of the CVT. This signal is The Cleveland variator has especially compared to a DC voltage proportional to great potential as a flywheel transmission the desired output angular velocity. The for two reasons: first, the variator difference in these two voltages is the features a top speed of 5,400 rpm, which error signal which activates a small DC exceeds the planned maximum speed of a servomotor. This servomotor drives a cellulose rotor; <=-"*ond, the variator has screw which cont ils the gear ratio of the an overall input-o > put ratio of 9:1 mak- CVT. For example if the error signal is ing it directly a, ,v icable to the expected positive (the output velocity is less than operating speed range of a cellulose fly- its desired value) the motor drives the wheel. These features raise the possibi- screw in such a way as to increase the 199 gear ratio. As the gear ratio increases, Of course in the real system unmodeled the output velocity increases, the error dynamics will increase system order and signal goes to zero, the motor stops and nonlinearities further complicate the pro- the system operates at its desired output blem; therefore, it may be necessary to speed. redesign the feedback control logic if difficulties such as hunting oscillations In practice, of course, the input develop in actual operation. velocity to the CVT will be constantly changing; therefore, the output control Microswitches are mounted at either system will operate constantly to regulate end of the operating screw on the CVT so the output velocity. The control system that the system can be disconnected when includes a variable gain and variable no further changes in gear ratio are pos- dead zone. One objective of the experi- sible. ments will be to determine if this control system can maintain the output speed con- The flywheel is driven by an AC in- stant within acceptable tolerances (^.5%). duction motor. The speed of this motor is controlled by a PARA-JUST controller. This An analysis of the simplified system controller is a solid state device which indicates theoretically that this system regulates the speed of an induction motor will perform as desired. The gear ratio by changing the frequency of the input of the CVT is defined as voltage to the motor. It maintains con- tact torque output from the motor up to (1) rated speed (power varies linearly with in speed). The PARA-JUST automatically accelerates the motor following a ramp in- the error signal is put ranging from 15 seconds to 150 seconds. It can also be adjusted to follow any com- u = K vW w (2) mand input voltage and has been modified error t desired " out' for manual control. In simulating the duty cycle a variable voltage will be in- and the angular velocity of the motor is put to the PARA-JUST which will then drive the motor and the flywheel so as to approx- w = = K K ((i) imate a typical input from an alternative motor terror m t desired " %uz^ energy source as a wind energy generator. (3) The gear ratio is ROTOR VIBRATIONS + w d A preliminary analysis of critical N = N(°) VoVt ^desired " out> ' speeds of the flywheel system mathematical (4) model indicates a fundamental natural fre- quency in the vicinity of 2400 rpm, midway °r S = Wt ("desired " wout> <5> in the originally planned operating speed range. The rationale in this analysis is where K = gain constant of the CVT that the natural frequency of the rotor- operation screw, unit changes elastic support system is a valid indica- in gear ratio/radian of opera- tor of critical speeds, as support asym- ting screw rotation metries, damping and bearing tolerances are small. The lowest critical speed con- K_ = motor gain constant, radians/ dition is assumed governed by the rotor ™ volt and its flexible support; the critical speed range is likely to be narrow, close Kt = tachometer gain constant, to that of the transverse natural frequency volt/radian of the assemblage. When Uout = wdesl>ed, dN= 0_ IfU()u t < Refinement of this analysis was con- ducted to include the effects of the iner- desired ^ > 0 and if u)out > desired ^ tia of the hubs supporting the rotor through elastomeric disc-like pads sand- < 0. Thus the gear ratio always changes wiched in the hub-flywheel interface. in such a way to reduce the error in angu- lar velocity to zero. Since the system is The model used in that analysis first order, no overshoots will exist. is shown in Figure 3. 200 As the lowest natural frequency is asso- commercially available cellulose materials ciated with essentially a quasi-static with energy storage costs competitive j distortion mode, a distortion was used with proposed rotors made from fiberglass derived from statics. An energy balance or Kevlar. Large multiton rotors can be then provided the prediction of critical readily assembled now in commercial plants speed: from plywood or paperboard. Such rotors show potential for stationary applications 2k. 1 where cost is the primary criteria. Ply- + ke/k3 wood also appears to be a promising mate- rial from which to construct inexpensive rotor hubs. Data on the general trends affecting 1 1 + 2 Q (l the long term strength is available. ke/k3 Experiments are in progress to measure the actual performance of the materials k k and assembled rotors under vacuum condi- tions. The practical details of assembly where k = . and balancing are being studied. The fac- e K1 + tors of outgassing and sustained high The effects of elastomeric pad stiffness vacuum are major unknown factors. in relationship to k , and hub-to-rotor mass ratio m/M, can be demonstrated Analytic expressions for critical graphically with the three-dimensional speeds are being refined and verified manifold (see Figure 4). experimentally. Continuously variable traction drives appear to be an efficient This demonstrates the sensitivity means of transferring power to and from of the lowest critical speed to hub the rotor. Control algorithms for the inertia and pad stiffness. For large pad operation of the flywheel system are being stiffness (K /K = 0) and sma1! hub mass developed and implemented. (jjj- = 0) the critical speed is gov- erned by the shaft and bearing supporting ACKNOWLEDGEMENTS the flywheel inertia. As these ratios increase from zero, a pronounced reduc- We would gratefully acknowledge the tion in critical speed occurs, and hence financial support for the different areas low ratios are desired if critical speed of research as provided by the Minnesota maximization is desired. Energy Agency, the Applied Physics Labor- atory, and the Sandia Laboratory. We In the prototype system, would also acknowledge the assistance of K /K = .18. Hence, u /|2ke/M the large number of students who have con- e 3 n tributed so much time and effort to the The subsequen' design trade-offs are project. Thanks are due also to the associated with the selection of a pad Eaton, Zero-Max, Onan, 3M, and Lenderink elastomer with inherent complex damping Companies, and to the Forestry, Physics modulus sufficient to attenuate critical and Chemical Engineering Departments who speeds, while simultaneously lowering have donated, or made available equipment these speeds due to the attendant elastic 201 Technical Conference, Montreal, ASME Paper No. 76-DET-96, Sept. 26-29, 1976. 3. Rabenhorst, D. W., "Composite Flywheel Development Program - Final Report," Johns Hopkins University, Applied Physics Laboratory: APL/JHU SD0-4616A NSF Grant No. AER 75-20607, May 1978. 4. Genta, G., Institute Delia Motorizza- zione, Politecnico Di Torino, Torino, Italy, Private Communication, Febru- ary 1978. 5. Erdman, A. G., Frohrib, D. A.,Garrard, W. L., Hagen, D. L., Carlson, T. P., "The Development of a Cellulose Fly- wheel System for Rural Wind Energy Storage: Final Report," University of Minnesota, for the Minnesota Ener- gy Agency, July 1978. 6. Erdman, A. G., Frohrib, D. A., Carl- son, T. P., Hagen, D. L., and Garrard, W. L., "The Design of a Wind Energy Storage System with a Cellulosic Fly- wheel," Proceedings of the 1977 Fly- wheel Technology Symposium, October 5-7, 1977. 7. Hagen, D. L., "The Properties of Natural Cellulosic Materials Pertain- ing to Flywheel Kinetic Energy Stor- age Applications," Proc. 1977 Flywheel Technology Symposium, October 5-7, 1977, San Francisco. 8. Stockman, V. E., "How Strong Can Paper Be Made?" TAPPI, Vol. 59, No. 3, March 1976, pp. 97-101. 9. Chiao, T. T., "Fiber Composite Mate- rials and Their Application to Energy- Storage Flywheels," Lawrence Liver- more Laboratories, 1978. 10. Hill, P. Ward, Hercules Inc., Cumber- land, Maryland, Private Communication, October 1978. 202 w\ \ \\\ \ \\ \ w wwwwwww Fig. 3. Vibration Model LEGEND A = Flywheel Housing B = Drive Motor (Below C = Clutch-Coupling D = Vacuum Pump (Below) /* / 7 i 4 2 /J E = Zero-Max Drive F = D.C. Control Motor G = A.C. Generator (Below) Fig. 1. Flywheel Demonstration Model Fig. 4. Natural Frequency Variations. Mou.t CVT radiometer out Transmission totio Motor Control 'Desired Error Fig. 2. Control System Schematic 203 PROJECT SUMMARY Project Title: Research and Development for Inertial Energy Storage Based on a Flexible Flywheel Principal Investigator: J. M. Vance Organization: Texas ASM University College of Engineering College Station, TX 77843 (713) 845-6225 Project Goals: Experimentally and theoretically analyze the rotational charac- teristics of proof-of-concept flexible flywheel configurations and devise methods for suppressing any undesirable whirl modes. Use the results of these analyses to develop the conceptual design of a flexible-flywheel energy storage system suitable for interfacing with a small size solar energy source. Project Status: The flexible flywheel support structure and associated equip- ment were moved from the University of Florida and installed in the newly established flywheel experimentation area at ASM. Installation of dynamic measurements instrumentation on the structure was begun. A 20-inch diameter flexible flywheel from the University of Flo.ida was spun on the structure and movies were taken (at both normal and high speed) of its non- synchronous whirl motion. The movies have not yet been screened for study. A concept has been developed for gimbal mounting the motor/ generator in a flexible flywheel energy storage system for the purpose of suppressing nonsynchronous whirl. The mathematical analysis of this concept has begun. Sandia has been requested to authorize the design and construction of a small table top model to explore this gimbal concept. Contract Number: 07-3693 Contract Period: Sept. 1978 - Sept. 1979 Funding Level: $39,708 Funding Source: Sandia Laboratories, Albuquerque 205 A CONCEPT FOR SUPPRESSION OF NONSYNCHRONOUS WHIRL IN FLEXIBLE FLYWHEELS Dr. John M. Vance Mechanical Engineering Department Texas A&M University College Station, Texas 77843 ABSTRACT An energy-storage flywheel sized for home or farm use is being developed at Texas A&M University. A unique feature of this "flexible flywheel" is its construction from flexible fibers (such as synthetic rope) with no bonding agent. The potential advantages are low cost, self balancing, and increased safety in operation. The one current dis- advantage, a whirling instability, is the major subject of the present contract. A ciesigp concept to suppress the instability is now being analyzed and is described herein. INTRODUCTION be a problem. In 1975, Dr. R. T. Schneider at the University of Florida conceived the idea of a flexible flywheel made of rope for energy storage. The idea was to develop a cheap, small scale, energy storage de- vice to make solar or wind-powered gener- ators practical for home or farm use. Beginning in 1976, a flywheel test facility was developed and the first rope flywheels were spun up. It soon became evident that one of the advantages of the flexible flywheel, its self-balancing fea- ture, had been bought at the price of a subsynchronous dynamic instability caused by internal friction, since the whirl crit- ical speed is well below operating speed. Fig. 1 Flexible Flywheel After it became evident that the maj- or (and apparently the only) technical The advantages claimed for this con- problem of the flexible flywheel is one of figuration are: rotor dynamics, the author of this paper 1. High strength fibers in pure ten- joined work on the project. The project sion with no bonding material to was subsequently moved to Texas A&M Uni- create mismatch in elasticity or versity when the author moved there. It strength. is less than two months since funding be- 2. Self balancing-highly flexible gan, at the writing of this paper. rotor operates at supercritical speeds. ADVANTAGES OF THE FLEXIBLE FLYWHEEL 3. Simple construction should be cheap to manufacture. Figure 1 shows a photograph of the 4. Gradual failure mode, with early flexible flywheel. Notice that the support warning by whiplash or hoop ropes carry only the weight of the fly- growth. wheel. Radial stiffness is provided by 5. Less destructive failure mode. gravity only, as in a pendulum. The ques- These advantages mean that the flexible tion of torque capacity and acceleration flywheel should be safer and less costly has been investigated and appears not to than conventional solid flywheels. 206 FLEXIBLE FLYWHEEL DESIGN EQUATIONS AND Applying the above equations and con- CONSTRAINTS straints to the design of a flexible fly- wheel allows the calculation of hoop Analysis has shown that to optimize sizes, weights and costs for various energy density, a flywheel should be con- choices of material. Results for Dacron R, structed of high strength materials to steel, and KevlarR are shown in Table 1. operate at high speeds. The maximum stor- age energy in a hoop flywheel is For successful energy storage, the most important parameter is the cost. How- E = TTRP ft-lb (1) ever, before the low cost advantages of a flexible flywheel can be realized, the E = (.3768)(10 )nRP KW-hrs. technical problem of rotor dynamic sta- bility must be successfully dealt with. where R = hoop mean radius, ft. P = cumulative strength of all ROTOR DYNAMICS hoop fibers, lb. It was recognized early that the sig- Notice that the material mass density nificant technical problems associated does not appear in the equation. Mater- with the flexible flywheel would be in ials with high mass density (heavy mater- the area of rotor dynamics, specifically ials) do not optimize energy density, the problem of subsynchronous whirling which is contrary to popular intuition. due to internal friction. The advantage of self-balanced operation at high super- The mass density does, however, affect critical speeds must be purchased with the speed at which the maximum energy is the price of suppressing or avoiding a stored.- This speed is given by self-excited dynamic instability. This is a challenge with a high payoff for success and one which the author believes rpm (2) can be met. where w = hoop specific weight, lb/ft Ever since Jeffcott's analysis of g = 32.2 ft/sec2 synchronous rotor whirl in 1919, rotor dynamicists have known that a flexible Although there are many advantages to rotor displays a "critical speed inver- using super-strong fibers to take advan- sion," in which the center of mass comes tage of equation (1), the resulting high inside the whirling rotor center at rotational speeds (most high-strength fi- supercritical speeds. bers are not heavy) pose rotor dynamics and bearing problems which must be proper- The flexible flywheel, by virtue of ly appreciated in the preliminary design its low stiffness rotor, always operates phases of any modern flywheel. For exam- at speeds which are highly supercritical, ple, contemporary electrical motors and where Jeffcott showed that synchronous generators are designed to operate well whirl (due to unbalance) amplitudes are below the speeds dictated by equation (2) minimized. Experiments to date have for a Kevlar R wheel. verified that the flexible flywheel pro- duces extremely low levels of synchronous There are also practical constraints vibration, with no precision balancing on the dimensions of a flywheel to be used required. in the home or on a small farm. Figure 2 gives dimension limits for a fiber hoop Not long after Jeffcott's results flywheel. became widely known and applied, it was learned that rotating machinery can be- come dynamically unstable in subsynchron- ous whirl at supercritical speeds, if the ratio of internal friction (in rotating parts) to external damping is high enough. Early tests of the flexible flywheel showed subsynchronous whirling which tend- ed to grow with speed and/or time. Figure 3 illustrates the mechanism of the inter- Fig. 2. Dimension Constraints nal friction excitation. For subsynchron- 207 ous whirl, the spin speed ft is faster A DESIGN CONCEPT than the whirl speed $. As support rope 3 moves around to position 1, its rate of Figure 4 shows how the motor/gener- strain is a maximum at position 2, thus ator can be gimballed on nonintersectlng generating the friction force F on the axes to provide the low support stiffness, hoop which is tangential to the whirl stiffness asymmetry, and bearing support orbit in the forward direction. mass (the motor itself), which are the parameters important to whirl stability. Since the motor/generator will be put inside the vacuum chamber, a cooling sys- tem is required. The coolant can serve double duty as the bearing lubricant. The motor bearings will be designed to support the flywheel, thus minimizing the number of bearings and eliminating the necessity for a clutch. Work is now in progress to verify this concept, both experimentally and Fig. Instantaneous Configuration of analytically. Rope Ring A table-top-sized model is being con- Rotor dynamics theory and analysis structed for preliminary evaluation of has identified several ways of suppressing whirl stability characteristics. The self-excited subsynchronous whirl. They distance from the y axis down to the motor are: center of mass is being made adjustable, 1. Flexible bearing supports. so as to be able to vary the stiffness 2. Asymmetric bearing support stiff- asymmetry. ness. 3. Bearing support damping. A mathematical model with four de- 4. Bearing support mass (dynamic grees of freedom is being analyzed to pre- absorber effect). dict the whirl stability characteristics. The characteristic equation (eighth order) Before considering how to apply these will be checked for conditions which methods to a flexible flywheel., it is use- guarantee all real parts of the roots to • ful to also look at the other "system" be negative. design considerations. It is hoped that this two-pronged FLEXIBLE FLYWHEEL DESIGN CONSIDERATIONS approach will yield a stable design with- out resort to a damping mechanism, which 1. A shaft seal through the vacuum Is often difficult to keep in proper chamber wall is expensive. adjustment over a long period of time. 2. A disconnect clutch also increases the total system cost. Once a stable design has been achiev- 3. A new motor/generator must be ed, the existing larger-scale prototype developed to match flywheel will be redesigned correspondingly to torque-speed characteristics. demonstrate the potential of home or farm 4. - Low friction bearings must be energy storage using a flexible flywheel. designed for the application, to operate in a vacuum environment. 5. The total number of bearings should be minimized, for lowest cost. When these design considerations are coupled with the design requirements to suppress the subsynchronous_whirl, a design philosophy for the flexible flywheel em- erges. The basic elements of this philos- ophy are shown in Table 2. Fig. 4. Flexible Flywheel on Gimballed Support 208 Table 1. Parametric Values for a 10 kw-hr Flexible Flywheel (Safety Factor = 2) MATERIAL HOOP WEIGHT MAX. RPM LENGTH L MAT'L COST lbs. ft. $ Dacron 1298 5,167 1.74 2654 (1/4") Steel 2123 4,039 .642 1938 (1/2" IWRC) Kevlar 162 14,606 .287 1380 (1500 Den.-"29") Table 2. Flexible Flywheel Design Philosophy Design Factors or Contraints Solution or Conclusion *Shaft seal is expensive Put motor inside vacuum chamber *New motor/generator required *Clutch increases cost Support flywheel directly from *New bearings required motor shaft *Minimize no. of bearings *Need low support stiffness Gimbal motor/generator on *Need asymmetric support stiffness nonintersecting axis *Need support damping and mass 209 PROJECT SUMMARY Project Title: Conceptual Design of a Flywheel Energy Storage System Principal Investigator: F. C. Younger Organization: William M. Brobeck & Assoc. 1235 Tenth Street Berkeley, CA 94710 (415) 524-8664 Project Goals: The development of a conceptual design for a flywheel energy storage system suitable for on-site interfacing with small scale solar energy sources. The basic design objective is to provide a generous margin of safety and above average reliability and efficiency at the lowest practical cost. Project Status: A basic concept for a flywheel energy storage system to interface with solar energy sources is being studied. This concept utilizes a constant voltage motor/generator directly connected to a flywheel rotor. The field current of the motor is controlled to maintain the voltage level at the optimum operating voltage of the array of photovoltaic cells. When a surplus of electrical energy from the solar source occurs, the voltage of the array tends to rise above the set voltage of the motor/generator, and surplus power drives the motor to accelerate the flywheel. When a deficiency in solar energy occurs, the voltage of the array tends to drop below the set voltage, and power is drawn from the flywheel via the generator. A variety of motor types and rotor types have been studied to arrive at a conceptual design to meet the program objectives. A fiber- composite rotor driving a separately-excited dc motor appears to satisfy the basic requirements. Design details are being examined to confirm the dynamic stability of the rotor suspension system and to determine the energy losses and system efficiency. Optimization studies, design layouts, detailed specifications, and cost estimates have not yet been made. Contract Number: 07-3663 Contract Period: July 1978 - June 1979 Funding Level: $139,053 Funding Source: Sandia Laboratories, Albuquerque 211 CONCEPTUAL DESIGN OF A FLYWHEZL ENERGY STORAGE SYSTEM Francis C. Younger William M. Brobeck & Associates 1235 Tenth Street Berkeley, California 94710 ABSTRACT A conceptual design of a flywheel energy storage system suitable for on-site inter- facing with small scale solar energy sources is being developed. The basic design ob- jective is to provide a generous margin of safety and above average reliability and efficiency at the lowest practical cost. This paper describes the basic design concept and the general approach to the conceptual design of the energy storage system. The basic concept to interface with solar energy sources utilizes a constant voltage motor/ generator directly coupled toa flywheel rotor. The voltage level of the motor is maintained at the optimum opera+ing voltage of the array of photovoltaic cells. Power is drawn from the flywheel-driven generator when the voltage output of the array drops below its optimum value and power is delivered to the flywheel via the generator (then operating as a motor) when the voltage rises above the optimum value. A variety of motor types and flywheel rotor types has been studied to arrive at a conceptual design to meet the program objectives. A fiber-composite rotor driving a separately-excited dc motor appears to satisfy the basic requirements. To obtain acceptable efficiency and low run- down losses, the flywheel operates in a vacuum and has a combination of magnetic thrust support and ball bearings to reduce bearing friction in an economical fashion. INTRODUCTION This paper describes an effort for the development of a conceptual design for The Flywheel Energy Storage System a FESS suitable for on-site interfacing (FESS) is intended to enhance the value with small scale solar energy sources. The and utility of small solar power generating new design will have as objectives a gen- plants. The timely development of econo- erous margin of safety, high reliability mical energy storage concepts is crucial in and efficiency obtained at the lowest many solar power applications. The variable practical cost. The design will be for a nature of sole • energy severely limits its stationary application, in which no cost applications if some practical storage premiums are jllowed for minimizing size system is not made available. Batteries, and weight. The basic design objective pumped-hydro, compressed air, thermal and is a system which can be economically mechanical energy storage all offer the manufactured in existing industrial facil- potential for suitable systems applica- ities using conventional production methods tions. Economic factors may favor one Commercially available components will be system in a particular application and yet used to the maximum practical extent. another in a different application. Thus, it is critical to subject all such systems Attainment of the highest practical to careful examination of performance and charge/discharge efficiency will be a goal cost characteristics. However, at present of the highe-st level priority. This re- such an examination is difficult because quires very careful attention to a myriad an adequate basis for cost and performance of design details, as well as the selectior analyses does not exist for emerging of the most appropriate engineering concepi technologies. New designs with generous margins of safety have not yet been demons- BACKGROUND trated for mechanical energy storage systems. A concept selection and design program The flywheel energy storage system foi which could lead toward a practical demons- use with a small solar energy source must tration is presented. satisfy several specific requirements in addition to being economically attractive. The program objective is the design of storage units suitable for use with small strength-to-we^ght ratio. In many engineer- to intermediate scale solar sources. The ing applications light weight is very nominal energy storage capacity is in the important and a premium may be paid for a range of 5 kWh to 100 kWh; the specific material with a high strength-to-weight requirements are for the detailed design ratio. However, for a utility application, of a 10 kWh storage system and for a de- additional expense merely to save weight tailed cost estimate for a 50 kWh system. would be unreasonable. The appropriate The required power rating of the 10 kWh design criterion is maximum energy at system is 5 kW. These units are to have an minimum cost. Fiber-composite using E-glass electrical interface for 60 Hz/220 Volts appears to satisfy this criterion. single-phase power. The control system must maintain the desired frequency and CONCEPTUAL DESIGN voltage over the full rotor operating speed range. The candidate flywheel energy storage system is shown in Fig. 1. It consists The round-trip efficiency for charging of a filament-wound fiber-composite fly- and discharging must exceed 70%. This wheel directly connected to a drive motor requires that the one-way efficiency exceed which acts as a generator during discharge 84%. The rate of loss of stored energy due of the system. The flywheel is contained to friction and windage must not exceed in an evacuated enclosure to keep the aero- 5% per hour. dynamic drag low. The enclosure also provides a means of support and containment for The design must have a generous factor fracture debris in case of failure. of safety and high reliability. It must be designed for a life expectancy of at least The flywheel (Ref. 1) shown represents twenty years during which a minimum of 10,000 a fiber composite concept which has been greatly charge/discharge cycles may be encountered. de-rated to achieve a high factor of safety. These cycles cover the full speed range of The energy density and resulting stresses the design. Allowance may have to be made are but a fraction of that which m.ght be for any overspeed tests which might occasion- obtained at some future time when fiber- ally be performed. composite flywheels are more fully developed. PROJECT DESCRIPTION The vacuum requirements are being set by trade-off studies where the benefits The project is divided into four main associated with low aerodynamic drag are tasks covering (1) the basic system concept; evaluated relative to the added cost, (2) optimization; (3) design of a 10 kWh complexities and power consumption of the unit; and (4) estimate for a 50 kWh unit. vacuum pump. These studies show that a hard vacuum is not required. A vacuum of 10"'* The basic system concept effort and Torr is quite adequate. subsystem options consist of an evaluation of workable engineering concepts and the The drive motor acts to accelerate the narrowing down of choices to the best sub- flywheel during charging and acts as a system options. The product of this work generator to decelerate the flywheel during will be a documentation of the comparative discharge. A variety of options for analyses and the selection process. This controlling the power flow to and from a is being done in the context of a provi- motor are available. These must take into sional layout of the total system. Sub- account the fluctuating nature of the load, system options include flywheel configura- the output characteristics of the solar tion, materials, flywheel housing and generator, and the speed variations of the containment, seals, type of motor/generator, flywheel. suspension system, mounting, controls, auxiliaries, and instrumentation. Problems The selection of a motor type is of safety, efficiency, losses and parasitic dependent upon the power control method to loads, manufacturability, maintenance, and a large extent but is also dependent upon economics are being identified and solved. other factors such as the operating speed range, voltage range, and operating envi- Various flywheel configurations and ronment. material combinations have been considered for energy storage systems. It has been For the conceptual design shown in shown that high energy density is potentially Fig. 1, a range of sizes and speeds is achievable with materials possessing a high required to satisfy the range of energy 213 storage desired. Table 1 shows the sizes E-glass shows a significant loss in strength and ratings for the various components of a due to static fatigue. Much of this loss candidate design for a 10 kWh demonstration is believed to result from the chemical model and for a 50 kWh unit. action of moisture in the air which may be prevented by exclusion of moisture or MATERIAL PERFORMANCE, COST ANALYSIS AND by operating in a vacuum. Long-term spin OPTIMIZATION testing in a vacuum is one way to determine the static fatigue limit for E-glass. Until For a given configuration, the energy such testing is completed, the only safe density in Wh/kg is proportional to the approach is to allow an adequate reduction strength-to-weight ratio of the rotor in the design stress. material; thus, the energy storage per dollar of material is directly proportional The Kevlar 49 compusite materials have to the strength-to-weight ratio divided by a higher specific strength than the fiber- the cost per pound. The cost of the vacuum glass materials and do show less sensiti- enclosure is dependent upon the size of the vity to static fatigue; however, these rotor which, fora given energy storage and composite materials are much more expensive. configuration, is dependent upon the absolute In spite of the higher cost of Kevlar 49 strength of the rotor material. The size a limited use of such a material may pro- of the rotor rather than its weight is a vide significant design advantages. The primary factor influencing the fabrication high modulus of elasticity and low density cost. The heavier material, in addition of Kevlar 49 makes it an ideal candidate to permitting the use of a small flywheel material for use as an overwrap to reduce for a given energy, may also allow a lower the radial tensile stress in a thick-walled rotational speed which, in turn, may yield ring by creating a radial compression at lower friction and windage losses. The the interface. overall effect of rotor material properties and cost upon tne complete system cost are A biannualte rim, shown in Fig. 2, being examined for a complete system design consisting of an inner ring of E-glass/ in which the cost of the enclosure, bearings, epoxy with an overwrapped outer ring of and motor are included. These costs depend Kevlar 49 can permit a considerable in- to a considerable extent upon the properties crease in radial thickness for the complete o'- the rotor, principally rotor shape and rim without giving excessive radial tensile material. stress. This large radial thickness also gives good space utilization. The ratio Of prime importance in the selection of Kevlar to fiberglass in the biannulate of flywheel materials is high strength rim will be determined by design studies. because the energy which can be stored per Increasing the amount of Kevlar may increase unit of volume is directly proportional to the cost faster than it increases the energy the allowable stress. A survey of flywheel storage. It appears from preliminary cal- projects in progress shows that there have culations that only a small amount of beon no complete life cycle tests and eval- Kevlar is required to give an optimum uation of high-strength high-energy density design. flywheels to date. Increasing the amount of Kevlar in the Several types of flywheels using differ- rim at the expense of the fiberglass has ent types of materials are being examined certain disadvantages. In the limit of an including concepts based on isotropic all-Kevlar rim, the maximum radial thick- materials (metals) and orthotropic materials ness must be quite small to prevent excessive (fiber-composites). Analysis of perform- radial tension. The utilization of space ance and total costs and projections of would then be poor. For a given energy, reliability and safety to personnel are the flywheel would be much larger and would being considered. At this time, it appears have a higher rim speed. Consequently, that a fiber-composite flywheel is the best the aerodynamic drag would be higher and choice. the vacuum enclosure larger. The Kevlar is a more expensive fiber and the larger Table 2 shows the properties of several enclosure for a Kevlar flywheel would also candidate fiber composite materials, of be more expensive. these, the E-glass is the least expensive. Although it has the lowest strength, the ratio of strength to cost makes it the most attractive candidate. Fatigue test on 214- FLYWHEEL HOUSING AND CONTAINMENT The design of the system should be such that in the event of flywheel failure only This component serves a number of im- a small fraction of the system momentum portant functions: and energy must be rapidly dissipated. The larger remaining fraction can be • An evacuated enclosure to reduce dissipated very gradually. windage losses. Two controlled braking methods are • An encasement to prevent corrosion being considered. One will provide a and mechanical damage to the controlled deceleration of fracture debris flywheel. rrom the failure of the flywheel rim and the other will simultaneously provide a • A structural member to transmit very much slower deceleration of the fly- loads. wheel hub. The first of these braking systems consists of a containment ring A structure for mounting the surrounding the flywheel rim. This ring bearings and seals. is permitted to rotate within the vacuum chamber with substantial frictional re- • A means of containment in the sistance to its rotation being used to event of failure of the wheel provide braking. Fracture debris from a or bearings.* rim failure would impinge upon this con- tainment ring which would be strong enough For a demonstration unit, the housing to hold the failed fragments and keep them could be a steel or aluminum alloy weld- rotating around on its inner surface. The ment. The housing for a production version fracture debris and the containment ring "baseline" flywheel module is expected to may both be rotating within the vacuum be a casting. A steel casting will be a chamber until the friction eventually prime candidate. Another housing candidate, brings them to e. stop. which will be considered in view of its potential for low costs, is a fiberglass The second mode for stopping the hub composite material. provides aerodynamic braking by permitting the pressure in the vacuum chamber to rise. The highest loads which must be trans- This braking is very gradual and will not mitted through the housing are those result in excessive torque or excessive associated with a flywheel failure of some temperature rise. type. High torques and high internal radial loads are expected if failure occurs. The containment ring must be strong enough to resist the initial impact of the The flywheel energy storage system fracture debris and the continued centrifugal has a large energy and momentum storage. force of the rotating debris as its speed A sudden and uncontrolled release of this gradually reduces. Analysis shows that energy and momentum would have serious the required strength is related to the consequences, unless, by careful design, kinetic energy of the fracture debris. this energy can be released as heat. The The relationship for the ratio of energy reduction in angular momentum will produce containment to weight of containment ring a torque proportional to the rate of change is the same as that for a rotating ring. in angular momentum. That is: For a flywheel with adequate inertia to £ _5- store 10 kWh of energy, the torque required W " 2w to remove all of its angular momentum in one second would be in excess of 68,000 lb-ft. where E = contained energy This is an enormous amount of torque and W = weight of containment ring quite likely is greater than the flywheel a = allowable stress in the ring mounting structure could support. It w = density of containment ring seems clear that efforts to limit the rate The importance of the strength-to- of change of momentum are required. weight ratio of the containment ring is shown in this equation. It seems clear *Placing the FESS in an underground con- that a high strength-to-weight ratio crete-lined pit may provide the most cost- of the material may be just as important effective containment means. here as it is in the flywheel itself. A 215 high strength-to-weight ratio fiber-compo- It is desirable to keep the pressure below site could be used for the containment 10"° atmosphere where loss is around 30 ring. A composite of E-glass may be more watts. If the pressure is allowed to rise economical than high strength steel for to 10-4 atm. the loss become excessive. this application. To maintain a pressure below 10"° atm. will require a mechanical pump. ENERGY LOSSES Vacuum seal friction is due to the Table 3 lists the categories of losses rubbing contact between the fixed and and parasitic loads of an electrically- rotating faces of the seal. This loss will coupled flywheel energy storage system. vary from 50 watts to 100 watts over the speed range. The bearing lubricating oil Table 3. System Losses and Parasitic will also lubricate and cool the seal. The Loads. seal will be designed to balance most of the atmospheric load so as to limit the rubbing Flywheel Unit Losses: force to a low value. Flywheel Bearing Losses MOTOR SELECTION STUDY Flywheel Windage Vacuum Seal Friction A successful motor application depends Vacuum Pump Power primarily on selecting a motor that satisfies, Motor Windage as nearly as possible, the kinetic require- Motor Bearing Loss ments of the driven machine without exceeding Motor Brush Friction* the temperature or torque limitations of Motor Excitation the motor. The first step in motor selection Motor Iron Loss is to determine the load characteristics- Motor Armature Copper Loss power, torque, speed, and duty cycle. Lubrication Pumping Power In this application the motor drives Power Conversion Losses: the flywheel shaft to accelerate during the charging time and acts as a generator Semi-Conductor Losses during discharging time. The motor must Other Component Losses provide good efficiency in both modes of Control Power operation. The most widely used motor Auxiliary Conversion Losses system which provides good efficiency as a motor and as a generator is a dc motor. However, ac motors can also be used as ac *If brushes are used. generators but the electrical control equipment is costly and will complicate Flywheel Unit Mechanical Losses. The the system. An economic study is being friction of the flywheel ball bearings is considered to test the feasibility of being estimated from the bearing manufac- both systems. turer's test data and from tests of fly- wheels of comparable design. The schematic diagram (Fig. 3) illus- trates the different modes of operation. For the flywheel candidate design, the ball bearing loss will be 80 to 120 The functional schematic diagram watts for the 10 kWh unit at 10,000 rpm. shown represents the block diagram of the FESS. The different blocks shown in the Windage loss of the flywheel proper diagram are as follows: depends, for a given geometry and speed, on the pressure and composition of the gas i) Flywheel; which is mechanically in which the flywheel runs. The dependence connected to a motor system. of windage loss upon pressure and flow regime has been calculated and used to estimate ii) Solar input; which has an the aerodynamic drag on the flywheel. output of 250 Vdc ±50V. In the pressure range in which the iii) Power flow controller; which flywheel operates, the mean free path of controls the power flow in the molecules of gas is long compared to and out of the motor system the wheel clearance in its housing, and and to the load. the windage loss is proportional to pressure. 216 iv) Power conditioner; which Losses calculated for motors have been inverts the power from dc divided into load-dependent and load- or ac with any frequency to independent losses. The latter determine exactly 60 Hz/220V single the standby power required by the flywheel phase ac. and is considered separately from the load dependent losses. The eddy current and v) Load; which can take power at hysteresis losses are dependent upon field 60 H2/220V single phase. excitation of wound-field machines. We have assumed that the field current will There are four modes of operation: be turned-off during standby to reduce run-down losses, similar to the design of 1. Only the solar source is Refs. 2 and 3. feeding the load. CONTROL OF POWER FLOW 2. The solar source and the flywheel both are feeding Generally, if the solar power genera- the load. tion exceeds the load power requirement, the excess power will be diverted to the 3. The solar source is feeding flywheel; and if the solar power generation the load and charging the is less than the load requirement, the flywheel. deficiency will be made up from the fly- wheel . 4. Only the flywheel is feeding the load. Generation of solar power is due to the photovoltaic effort (PVE) which occurs Note that in Modes 2 and 4, where the in semiconductors. Charge separation flywheel is discharging and feeding the in the photovoltaic cell will cause an load, the motor is acting as a generator. electrostatic potential difference across its P-N junction, and by placing cells The motor has a dual function, i.e., in parallel or series, we can obtain the it acts as a motor when it is charging required voltage or current for a specific the flywheel and it acts as a generator load. Four parameters are of interest in when feeding the load during discharging. analyzing the performance of a photovol- Separately-excited dc motors are easily taic cell. They are the short circuit switched from the motor configuration to current (I,-); the open circuit voltage the generator configuration and can handle (V ); the current for matched load, i.e., the speed range for this application. An current under maximum power transfer ac type motor can also be used. However, (operating point) conditions (ImD) and since the electrical frequency will vary the corresponding voltage (V _). as the speed of the flywheel changes, there will be a need for electrical power and Figure 4 shows typical current- frequency control devices to convert the voltage characteristics of a P-N junction high and variable frequency of flywheel solar cell array under a given amount of source to the low frequency of the load. sunlight. Since the maximum power output occurs when the current and voltage is at A brush!ess dc motor is very attrac- ^D and ^nro' °Perating at this point makes tive because lower losses and higher available pthe maximum .power. The control reliability are expected. system should, to the extent possible, keep the solar output voltage at V . If the There are several ways to reduce power load demand exceeds the solar poutput losses in the motor. One is to reduce losses either because there is inadequate sunlight in the core, either by adding more material or because there may be a high load condi- to the magnetic core structure or by using tion, then the voltage would start to drop a steel with improved core-loss properties below Vmp. To prevent this voltage drop and thinner lamination. Another method is additional power must be supplied by the to increase the cross-sectional area of flywheel energy storage unit. On the conductors to reduce resistance. Another other hand, if the load demand is lower alternative is to shorten the air gap than the solar output, then the voltage to reduce the magnetizing current required. would rise above V . To prevent this voltage rise the pexcess power must be used to recharge the flywheel. As examples, consider the case where the sunlight is such 217 that the maximum output is 3kW and V is SUSPENSION SYSTEM 200 Vdc. If the load demand is 8 kW^the control system would bring the flywheel The suspension systems being considered output to 5 kW to keep the solar output provide sufficient flexibility to permit voltage at V . If the load demand is 1 kW, the flywheel to rotate about its own mass the control psystem would divert 2 kW to center. The total deflection needed to recharge the flywheel to keep the solar satisfy this requirement can be quite output voltage at V . small and is dependent upon the unbalance of the flywheel. The mode of operation During normal operation the sunlight which permits this deflection of the mass fluctuates going to zero at night, also center involves rotational speeds greatly the load fluctuates making abrupt changes in excess of the first critical speed. as electrical equipment is turned on and Operation at such speeds would be unstable off. As these fluctuations in output and if special precautions to insure stability demand occur, the control system senses are not taken. Means of preventing whirl the changes in voltage and regulates the instability are known and will be incor- motor output or input to discharge or porated into the suspension system (Ref. 4). recharge the flywheel unit and maintains the voltage. In the design example with A variety of suspension systems are a separately-excited dc motor, the field possible. It is difficult to prove that excitation would be raised or lowered by one system has an overwhelming advantage over the control system in order to keep the all others. Our experience (Refs, 5 & 6) with a desired voltage. The usual dynamic prob- pendulous mounting shows that this system lems of feedback control systems will be is acceptable when precision ball bearings solved in the conventional way. are used. A stradle mounted system may be preferable where magnetic bearings are used. For other types of motors, the control problems are similar. The control system The analysis of run down losses indi- will regulate the motor output or input cates that some magnetic support to reduce power of the flywheel unit to keep the bearing loads is required. Thus, a voltage output of the solar array at V . suspension system using a combination of magnetic and ball bearings will be designed. If the flywheel is completely charged Such a system will have magnetic support or if it is discharged, then the control above the flywheel which will be rotated system will not be able to regulate the about a vertical axis. The magnetic voltage. In this case, the voltage must support will be passive. The electrical be allowed to rise or fall. If the FESS power requirements for such a bearing is too small, excess power from the solar appear, from preliminary calculations, to unit will be lost after the FESS is fully be very small. Of course, the power must charged. be smaller than the frictional loss of conventional bearings in order to justify the use of the magnetic support. The state of charge/or discharge of the flywheel is simply determined by its REFERENCES speed. When the flywheel is being charged from the excess power of the solar unit, an 1. Younger, F. C., "Tension-Balanced overspeed sensor can be used to shut off Spokes for Fiber-Composite Flywheel the flow of charging power. Similarly, Rims," 1977 Flywheel Technology when the flywheel is being discharged, the Symposium, CONF-771053, March 1978. under speed sensor (or inadequate voltage) can be used to shut off the output power of 2. Brobeck, W. M., "Design Study for a the motor. However, the normal operation Flywheel-Electric Car," IEEE 28th should be such that the flywheel is dis- Vehicular Technology Conference, charged only to its minimum rated speed, Denver, Colorado, March 22-24, 1978. where the motor can no longer generate the required output voltage at maximum 3. "The Application of Flywheel Energy field excitation. Storage to Electric Passenger Automobiles," William M. Brobeck & Associates, Report No. 8600-28-R1, February 1977. 218 4. Thomson, W. T., Younger, F. C, and Gordon, H. S., "Whirl Stability of the Pendulously Supported Flywheel System," Journal of Applied Mechanics, Vol. 99, No.s, June 1977. 5. Brobeck, W. M., "Flywheel Energy Storage for Utility Applications," IEEE Power Engineering Society Summer Meeting, Paper No. A 77 652-1. 6. Brobeck, W. M., "Flywheel Development for the Electric Power Research Institute," 1977 Flywheel Technology Symposium, CONF-771O53, March 1978. 219 MOTOR- GENERATOR Figure 1. Candidate Flywheel Energy Storage Unit 220 Figure 2. Candidate Fiber-Composite Flywheel Rotor 221 SUNLIGHT SOLAR FLYWHEEL INPUT / POWER FLOW CONTROLLER to MOTOR/ POWER GENERATOR CONDITIONER si O g Figure 3. Functional Schematic Diagram Figure 4. V-I Characteristic of Photovoltaic Device 223 Table 1. Characteristics of a Candidate Flywheel Energy Storage System. 10 kWh Demonstration 50 kWh Unit Unit Energy, kWha 10 50 Power, kW 5 10 Speed, rpm 10,000 5,850 Flywheel Diameter, in 50 8505. Flywheel Height, in 13 22.2 Flywheel Weight, lbs 613 3,065 Motor Type dc separately dc separately excited excited Converter dc to ac dc to ac inverter inverter Efficiency - round-trip 71% 72% Run Down Losses 4«/hr 4%/hr Output 60 Hz/220 V 60 Hz/220 V Nominal Energy Rating ac to dc rectifier may also be needed if utility power is used for recharge 224 Table 2. Fiber Composite Properties Fiber: E-Glass (Owen-Corning) S2-G1ass (Owens-Corning) Kevlar 49 1420 denier Vol% Fiber: Nominally 65 vol% Nominally 60 vol% Nominally 60 vol% 3 Density: 2.1 gm/cm 2.08 1.38 Mechanical Properties Elastic Constants: Longitudinal Young's Modulus, E-|i 52.15 t 0.89 GPa 56.2 ±2.7 GPa 81.8 ± 1.5 GPa Transverse Young's Modulus, E22 14.03 t 0;61 GPa 15.7 ± 1.0 GPa 5.10 ± 0.10 GPa Shear Modulus, G]2 6.3 t 0.5 GPa 7.41 ± 0.56 GPa 1.82 ± 0.09 GPa Major Poisson's Ratio, u-j2 0.207 t 0.016 0.282 ± 0.031 0.310 ± 0.035 Minor Poisson's Ratio, U21 0.056 i 0.011 0.079 ± 0.014 0.0193 ± 0.0014 Ulti mates: Tension ro Tension Tension to ui Longitudinal Strength 1108 t 25 MPa 1615 ± 127 1850 ±50MPa Longitudinal Ultimate Strain (2.16 t 0.11%) ~2 2.23 ± 0.06% Transverse Strength 7.5 t 1.1 MPa 41.0 ± 6.3 7.9 ± 1.1 MPa Transverse Ultimate Strain 0.054 t 0.009% 0.292 ± 0.064 0.161 ± 0.023% Compression Compression Compression Longitudinal Strength 530 tllO MPa 460 ± 60 MPa (235 ± 3 MPa Longitudinal Ultimate Strain (1.11 :t 0.27%) 0.92 ± 0.06% (0.48 ± 0.3%) Transverse Strength (78 t 4 MPa) 111.8 ± 2.2 MPa (53 ± 3 MPa) Transverse Ultimate Strain (0.68 :t 0.10%) 2.93 ± 0.32% (1.41 ± 0.12%) Shear Stress at 0.2% offset 22.4 :t 1.7 MPa 30.4 ± 0.98 24.4 ±-2.4 MPa Shear Strain at 0.2% offset 0.546 :t 0.045% 0.620 ± 0.041 1.55 ± 0.16% Matrix: 100 parts Dow Chemical DER 332 (bisphenol-A epoxy resin), 45 parts Jefferson Chemical Jeffamine T403 (polyether triamine) Cure: 16 h @ 60°C (E-Glass & S2-Glass) Cure: One day at room temperature and 16 h at 85°C (Kevlar) Reference: L. L. Clements and R. L. Moore, Lawrence Livennore Laboratory, Report UCRL-79262 (1977. PROJECT SUMMARY Project Title: Residential Flywheel with Turbine Supply Principal Investigator: T; W. Place Organization: Garrett AiResearch 2525 W. 190th Street Torrance, CA 90509 Project Goals: Develop the conceptual design of a cost effective flywheel energy storage system that interfaces with a wind turbine energy source. The system is to be sized for residential use. Design emphasis is to be on the use of current rather than future technology. Accordingly, the manufacturing methods are to involve conventional processing to the maximum practical extent. Project Status: A literature search was conducted for evaluating promising system concepts. This has culminated in the selection of a baseline system concept for analysis. This concept consists of five major subsystems. These are a gearbox, generator, variable speed transmission, flywheel rotor, and controls. The input to the system is mechanical shaft power from a vertical axis wind turbine. The system output will be based on the load demands of an all-electric single family residence with a floor area of 139 m2 {1500 ft2). Time varying input power and electrical load data have been furnished in card-deck form by Sandia and will be used as guidance in evaluating the conceptual system. Trade-off criteria were established for the comparative rating of subsystem options, and several variations have been selected for each subsystem for cost and performance evaluation. A computer code is being written for analyzing the baseline system with various combinations of subsystem options. These analyses are scheduled for completion at the end of November. Contract Number: 07-9093 Contract Period: July 1978 - July 1979 Funding Level: $135,899 Funding Source: Sandia Laboratories, Albuquerque 227 RESIDENTIAL FLYWHEEL WITH TURBINE SUPPLY Theodore W. Place AiResearch Manufacturing Company of California A Division of The Garrett Corporation 2525 W. 190th Street, Torrance, California 90509 ABSTRACT This paper examines a flywheel system that stores energy from a wind turbine source and converts the energy to 60-Hz, 220-v output for residential use. The typical residence has a 1500-sq ft floor area with a maximum power level of 5 kw. The purpose of this study is to determine the cost/benefits of storing wind energy in a flywheel and using it on a demand basis. The study examines the systems and the flywheel rotor materials that offer the greatest promise for reducing the initial cost. The paper presents the progress to date on th i s program and 'Jescr i bes the work planned to complete the study. INTRODUCTION technology. The method of manufacturing is to be conventional processing. The study The vertical wind or Darrieus turbine is expected to stimulate industry to pro- converts wind energy to mechanical energy. duce and market a mechanical energy storage The mechanical energy is stored in a fly- system that will produce a significant wheel for use as electrical power in a reduction in the use of utility power. The 1500-sq ft house. Calculations have shown energy storage system will also have an that the typical residential demand in an acquistion cost that produces a net savings area such as Albuquerque, New Mexico, is in life-cycle costs. 2.4 kw, and that the demand slowly undu- lates as a function of time. The first BACKGROUND step in applying the stored energy is to change it to electrical power and use it In 1925, G.J. Darrieus,1'2 of Paris, on a demand basis in the home. France, applied for a U.S. patent for a new type of windmill to generate power. Tests conducted on the wind turbine The vertical axis turbine shown in Fig. 1 have shown that the wind energy is inter- has several blades of symmetric airfoil in mittent and that the available energy can cross section, and is curved in the shape be much greater than demanded by residen- that a flexible cable of uniform density tial electric use. One solution is to and cross section would assume if spun store this pulse of energy in a flywheel about a vertical axis. The blade shape that has been sized to accept the avail- has been designated troposkien. The advan- able wind energy, and to use it to supple- tages of the Darrieus turbine over a con- ment the utility power. These sizing ventional propeller type wind turbine are studies will be conducted to determine its ability to accept wind from any direc- the system effectiveness. tion, its lower and simpler construction costs (since its qenerator is located on PROJECT GOALS the around), and the resulting lower maintenance. The project goal is to identify a cost-effective meciianical storage device. Several tests have been run in which the In addition, the device must be safe and meteorological data were collected at vari- reliable for 20 years. The study con- ous test sites and used to calculate both straints include a 5-kw output generator supply (wind velocity) and demand (solar and a 10-kwhr flywheel system. The insulation, air temperature, and wind mechanical storage system is to use velocity). These data are illustrated in current technology rather than future Fig. 2 for the Blue Hill test site in Massachusetts. 228-3CZT kevlar. These devices have been construc- ted for applications such as subway rail vehicles or passenger vehicles. These data will be used in the study. AIRFOIL SECTION PROGRAM PLAN The program is divided into four func- tional tasks plus a documentation task. These are shown in Table 1. The first task is to select the basic system and to establish the component trade options. The second task will evaluate performance, cost, and reliability. This will be done by evaluating selected com- ponents in the baseline system, using a computer study. The test data collected at Bh.e Hi I I and Albuquerque w!! I be input. Task III will provide a specification and estimate of a 10-kwhr system. In Task IV, a 50-kwhr system will be extrapolated and est i mated. PROGRAM SCHEDULE WORK DESCRIPTION IK I 1.0 Of VIII SASIC SVSTEM i.o SELECT IUSSYSTCM OPTION!: I*. FLYWHEEL ENEROV STORAflE UNIT U.IIICTIIICAL MOTOR/ GENERATOR 2c. THANSMII DON 2*. CONTROL! TAIK II Fig. 1. Vertical axis wind turbine .0 UTUP COftT-MRFORMANCE MOO1L !.Q ESTIMATE PERFORMANCE showing modified troposkien configuration. 3.0 ESTIMATE SAFETY AND RELIABILITY IASK III ULYOUT AftD SPECS noKWi SELECT KANOrACTURERB COST ESTIMATE ANNUALIZE COST TASK IV EXTRAPOLATION TO ID KW INPUT ESTIMATE COiT ANNUALtZCD COST At REPORT STATUS i Table 1. Schedule for Sandia flywheel energy storage system study. PROGRAM METHODOLOGY The method that has been chosen for the study is as follows. First, a litera- ture search has been accomplished in order TUES WED THRU to evaluate both the baseline system and the components. Second, a baseline system will be selected and preliminary sorting of Fig. 2. Wind energy (input) versus residential demand (output). components will be estimated. The assump- tions will be defined along with the The present study will use data for a 365- charging and discharging cycle. day year for both a minimum wind site such as Albuquerque, New Mexico, and a maximum A performance and cost computer model wind site such as Blue Hill, Maine. will be established for the baseline system. Selected components will be Inserted into Flywheels have been constructed and the program, and the performance and cost tested that have been made either of iso- data wilI be evaluated for input wind and tropic material such as steel or of com- •electrical demand data. The baseline sys- posite material such as fiberglass or tem with selected components will be traded 229 off to optimize performance in terms of FLYWHEEL cost. The model will then be evaluated for safety and reliability. The flywheel will have the following variables: The 10-kwhr system preliminary design will be finalized and defined by cross sec- (a) Shape tion drawings and specifications. Finally, a 50-kwhr system will be extrapolated. The (b) Material results will be documented in a final report. (c) Vacuum SYSTEM DESCRIPTION (d) Bearings The mechanical energy storage system for a residential application consists of (e) Seal five major components. These are a gear- box, a generator system, a variable speed transmission, a flywheel storage system, Flywheel Shape. The variation of shapes' and controls. A typical system is illus- for the steel or isotropic material fly- trated in Fig. 3. In the system analysis, wheel will be limited to a truncated coni"- candidates for each of the above components cal disc, solid disc, and pierced disc. will be substituted into the performance The composite material will use a concen- and cost analysis to select the system. tric ring approach. Fig. 4 illustrates the shape factor Ks that is used in calculating energy density. The second variable for CLUTCH consideration is the length-to-diameter DARK EUS FLYWHEEL ratio. The selected minimum diameter is MIND |/\J CONTROLLER TUHB NE 2.0 ft and the selected maximum diameter BOX is 6.0 ft. , JOTOR/ SPEED 60 HZ TRANS- TRUNCATED t MISSION 220 V VOLT. _ rnmrAi ^~-~ 1 0.8 REG DISC "-5^; TAPER RATING = .3 2 1 0.6 Fig. 3. System description with ac generator. -A I—i.c 0.3 3 msr. 1 1 1 The input to the system is mechanical f— o. o. I shaft power supplied at either a constant 12t>- or 75-rpm speed, depending on the test location of the wind turbine. The wind CONCENTRIC CYLINDERS J 0.3 power has been supplied for a worst case (Al RESEARCH) ~| and best case in the form of test data! supplied by Sandia Laboratories. The output of the system (electrical Fig. 4. Flywheel rotor candidate shapes. demand) has been calculated assuming an all- electric house with a floor area of 1500 sq ft. These data have also been supplied Flywheel Material. The available material on a card deck which describes the yearly for constructing flywheels is listed in demand of the residence for 60-Hz, 200-v Table 2. The table shows the specific power. energy per pound of flywheel material. Considering present-day technology and COMPONENT DESCRIPTION demonstrated success, several candidates were selected for preliminary cost analy- The selection of components wilI be sis. Those candidates marked with an accomplished in two steps. The first step asterisk will be examined in the system is to select candidate components and performance and cost model. Table 3 shows determine their characteristics. the flywheel rotor cost comparisons. 230 Table 2. Flywheel material candidates. Energy density versus ;;hape factor Density Allowable Mater i a 1 Ib/cu in. stress m 0.8 0.6 0.3 *High-strength steel 4340 0.283 140,000 12.4 9.3 4.7 •Maragtng steel 18 Ni + 300 0.289 150,000 13.0 9.8 4.9 *High-strength load 0.283 too,ooo 8.8 6.6 3.4 alloy 100 AR Aluminum 7075 0.101 26,000 6.5 4.9 2.4 Tivanillin 6A1 4V 0.160 70,000 11.0 8.2 4.1 Depleted uranium 0.683 25,000 0.9 0.7 0.3 *E-glass-epoxy and 0.080 120,000 14.0 aluminum hub S-glass-epoxy and 0.079 130,000 15.5 aluminum hub Kevlar 49 and epoxy 0.052 170,000 30.7 and aluminum hub Plywood and birch 0.022 8,000 6.8 3.4 ^Selected for system evaluation! Table 3. Flywheel rotor comparison (preliminary) Costs dollars/1b Flywheel Cost per kwhr, Material E w-hr Labor Too 1i ng Unuseable Mater i a 1 dollars Hiqh-strenqth 12.4 0.25 0.10 0.07 1.32 140 steel 4140 Maraginq steel 13.0 0.25 0.10 0.10 1.92 182 18 NI -300 High-strength 6.6 0.36 0.05 0.07 0.32 121 low al loy 100 AR E-g1ass-epoxy 14.0 0.84 0.05 0.01 0.87 126 and aluminum hub E-g1ass-epoxy 12.6 0.84 0.05 0.01 0.72 128 and steel hub Kevlar 49 and 30.7 0.84 0.05 0.01 6.75 249 epoxy and aluminum hub 231 Vacuum for Flywheel. The fIywheeI system between the 1200-rpm gearbox and the fly- windage losses will be insignificant if the wheel, and since the design horsepower cavity pressure is maintained between 1 and is approximated 26 hp, then it appears 10 microns. Fig. 5 shows that the loss is that either the traction or the variable- insensitive to the lei.^th-to-diameter ratio. pulley rubber belt would satisfy the The variables for the study will then be design. The relative costs chart listed in Fig. 7 shows the variable-pulley rubber the cost impact of a hermetic sealed, be 11 to be most attract i ve on a f i rst-cost magnetic coupled flywheel versus a sealed basis. These two selections will be shaft, and the related pump down equipment. carried on for final analysis. The second part'of the transmission tradeoff is the cost of a transmission. The transmission will vary from zero to fu 11 speed versus a speed change of 3 to 1, plus the special motor that must be designed to tolerate the heat and large slip angle. Fig. 8 shows the candidates for zero- speed versus the straight 3-to-1 belt system. 100,000 10,000 Fig. 5. Effect of length/diameter ratio versus a. windage losses. Bearings for the Flywheel System. The u candidate bearings'* for the performance and | cost analysis will be ball bearings, roller ° bearings, and hydrostatic bearings. The § evaluated performance and cost will be the £ criteria for selection. The magnetic bear- 3 ings have been eliminated since the loss is j| low for the lower cost bearings. Seals for Flywheel System. The seals4 for the flywheel system that have been selected as candidates are the carbon seal, and the hermetic sealed/magnetic coupled unit. 1,000 10,000 100,000 The ferro fluid seals have been eliminated RATED OUTPUT SPEED, RPM because of hiqh losses for the large- Hm7 diameter seal. The life-cycle cost will be Fig. 6. Design rationale for selection evaluated as part of the vacuum system of transmission options. analysis. 100,000 Structure for Flywheel System. The fIy- FLUID wheel containment housing4 must be sealed for vacuum. The candidates for cost evalu- ~ 10,000 ation are a hermetic sealed metal container, a metal and concrete container, and a metal METAL-CHAIN AND WOOD-BLOCK BELT container with earth embankment or pit. 1,000 The vertical axis and the earth oriented axis will be examined for the flywheel. TRANSMISSION 100 The purpose of the transmission is to provide a variable speed input to the fly- wheel system so that mechanical energy may 1 10 100 1,000 JO, 000 be stored and extracted as a resuIt of a speed change of the flywheel. The trans- RATED POWER CAPACITY, HP MI7M mission type was first selected between the candidates listed in Fig. 6. Since the Fig. 7. Cost rationale for selection transmission is placed in the sysTem of transmission options. 232 MOTOR/GENERATOR alternator, it is less stable under vari- able loads and it operates at a lower power The main function of the motor/ factor. generator (M/G) is to provide a.nominal 220 rms (110-v, line-to-neutral, single- The separately excited dc generator is phase) at a nominal 60 Hz. In the motor readily voltage-regulated, reasonably effi- mode it can be used to start the wind tur- cient, capable of absorbing overloads, and bine, using the utility as a power source. capable of wide speed variation under loaded conditions. It must be augmented The M/G could also be applied as a with a solid-state inverter to provide the velocity governor for the wind turbine if household 60-Hz power. The beneficial it is referenced to the utility to maintain operating characteristics are offset by an accurate 60-Hz frequency. If it is not the high cost of the required inverter. referenced to utility frequency, the fly- wheel and CVT must be controlled to provide FUTURE WORK the frequency required. The frequency would then be accurate enough for all A base Iine computer model has been household loads except synchronous clocks. established. In this way, the candidate components may be substituted into the sys- The three M/G types that are commer- tem and input/output performance examined cial I y available as well as technically for rotational losses. In addition, a life- acceptable are: cycle cost analysis will be accomplished to select the lowest cost system. (a) Salient pole synchronous alternator After the system is selected, a layout (b) Squirrel cage induction alternator drawiny and specifications will be accom- plished for the lu-kwhr storage system. A (c) Separately excited dc generator cost estimate will be prepared for both a 10-kwhr and a i>G-kwhr extrapolated system. The salient pole synchronous alter- REFERENCES nator is the most easily controlled of the three candidates. With slip rings to trans- 1. J. F. Banas, W. N. Sullivan: fer dc field current to the rotor, voltage Engineering of Wind Energy Systems regulation can be readily accomplished. It (Sandia Laboratories, SAND75-0530, cannot control suddenly applied overloads January 1976). since it will lose synchronism if pullout torque is exceeded. Amortisseur windings 2. B. F. Blackwell: The Vertical Axis are required if it is to be used as a Wind Turbine, How It Works (Sandia starting motor. Laboratories, SAND74-0160, December 1974). TRANSMISSION TRADEOFFS 3. D. A. Towgood: An Advanced Vehicular ! i 1 Flywheel System for ERDA Electric VARIABLE VARIABLE BELT VARIABLE Powered Passenger Vehicle (U.S. ANP SPEED TOROIDAL BELT AND BELT AND SPEED INCREASER DIFFERENTIAL Department of Energy, CONF 77J053, INCREASER WITH FLUID GEAR DRIVE COUPLING October 1977). 4. Economic and Technical Feasibility Study for Energy Storage Flywheels (U.S. Department of Energy, HCP/ M1066-01 , May 1978). Fig. 8. Transmission tradeoffs. The squirrel cage induction alternator is capable of continuous near-synchronous operation despite short term overloads. However, its efficiency is not as high as the synchronous alternator unless it is controlled to a low slip operation. In general, when compared to the synchronous 233 SESSION V: SUPERCONDUCTING MAGNETIC ENERGY STORAGE 235 PROJECT SUMMARY Project Title: Superconductive Energy Storage Principal Investigator: R. W. Boom Organization: University of Wisconsin 531 E.R.B. Madison, WI 53706 608/263-5026 Project Goals: To develop hardware components, produce engineering system designs, procure manufacturing and assembly equipment designs, revise cost estimates, and assess operational efficiencies for large underground superconductive storage systems. Project Status: Component development in the areas of cryogenics, conductors, and structures is under way and to be completed in FY-81. Rock mechanics design and experimentation is under way and to be com- pleted in FY-81. Parallel efforts supported by the Wisconsin Utilities are under way on the same schedule in five other areas: Magnetics, environmental studies, system design, electrical, and safety. Contract Number: EY-76-C-02-2844-000 Department of Energy Contract Period: FY 1970 - FY 1981 Funding Level: $600,000 FY-78 D.O.E. * Funding Source: Department of Energy, Division of Energy Storage Systems Total Funds: $3,174,000 FY-70 to FY-78 ERDA, DOE $1,600,000 NSF 715,000 IKW. 554,000 Wisconsin Utilities 250,000 General Electric 55,000 'Included in this project are the following: S. W. Van Sciver, "Recent Component Development Studies for Super- conductive Magnetic Energy Storage," page 247. 237 SUPERCONDUCTIVE DIURNAL ENERGY STORAGE STUDIES R. W. Boom Engineering Experiment Station University of Wisconsin Madison, Wisconsin 53706 ABSTRACT A general description of a large central energy storage unit employing super- conducting magnet technology is presented. Work on this concept has been under way at the University of Wisconsin since 1970. Economic and engineering optimization has shown that a superconducting energy storage magnet system would be competitive with other storage schemes for units larger than 1000 MWh. The device consists of a large superconducting magnet of approximately 100 m radius interfaced with the power system with a three-phase Graetz bridge. Economic analysis dictates that the magnet be a single layer segmented solenoid buried in bedrock and operated in super- fluid helium. An overall discussion of progress on the Wisconsin project is pre- sented including recent cost estimates and time schedule for commercialization of the technology. INTRODUCTION According to the virial theorem, Energy storage studies have been M > § E, the structural mass M per kWh of under way at the University of Wisconsin stored energy E must be greater than since 1970. The originating Wisconsin 264 lbs/kWh for stainless steel with idea was that a three-phase Graetz 0 = 50,000 psi and p = 7.86 g/cm3. The bridge could be used to convert dc cur- weight per pound can be reduced only by rent in a superconducting storage magnet using lighter material at higher stress into ac1current in a three-phase power levels. The practical result is that system.' The Wisconsin work has con- such structural requirements absolutely centrated on large central storage units forbid any purchased structure, only in- for diurnal use while a subsequent energy expensive bedrock is available. Fly- storage effort at Los Alamos (see com- wheels also suffer from the same absolute panion papers by J. Rogers et al. in this prohibition if $/kWh is significant. session) has evolved to include a concen- tration on small storage units for utility The three-phase ac/dc Graetz bridge stabilization purposes. The efforts at and the dc energy storage magnet coil Wisconsin and at Los Alamos represent form an inductor-converter (I-C) unit. the major worldwide activity in super- Typical uses for large I-C units are conductive storage although there are 2-12 hour discharges at rates of 100 to indications that Japan and Russia are 2000 MW for nighttime charging and day- becoming interested. time discharging. The concept is that large but Large central storage I-C units are simple superconducting solenoids store unique in that 90-95% efficiency is pos- energy in the form of dc currents in an sible. Another unique property is the inductance equal to V2LI2, where L is speed of response; within 50 milli- the inductance and I the dc current. The seconds an I-C unit can change its power solenoid turns are superconducting to level from full charge to full dis- eliminate I'R losses, where R is zero charge. We are not aware of any other resistance for superconductors. The use storage system with these two advantages. of superconductors necessitates the use of cryogenic systems and a liquid helium A total of $3,174,000 has been coolant. The solenoid would be mounted expended at the University of Wisconsin in bedrock which is the least expensive from 1970 through FY-1978 from the mechanical support structure available. various sources listed in Table 1. The University of Wisconsin and General generally estimated to be between 5% and Electric Corporation provided initial 10% of the peak available power from funding followed by National Science generators. The duration of the power Foundation support which led to the delivered from storage would vary from present ERDA-DOE grants. Throughout the 2 to 10 hours in different utilities eight years the seven electric utility across the country with a trend towards companies in Wisconsin have provided needing 12 hours from storage. The peak working staff in addition to the funds power period starts in the morning after listed which were transferred to the 8 A.M. and persists through the late University from the Wisconsin Electric afternooon on weekdays. Occasionally Utility Research Foundation. peak needs for storage even arise during a weekend. The peak power in Wisconsin two years ago was on a Sunday that was Table 1. Aggregate funding for the exceptionally hot and humid. University of Wisconsin Energy Storage Project from 1970 to October 1, 1978. As an example let us take the state Amount of Wisconsin whose power requirements are Source about average for the 50 states. The University of Wisconsin $ 554,000 peak power is about 8000 MW which implies General Electric Corp. 55,000 that 800 MW would be desirable from Wisconsin Utilities 250,000* storage for about 10 hours, as has been N.S.F. 715,000 determined by our utility collaborating E.R.D.A. 300,000 engineers. Thus 8,000 MWh might be D.O.E. 1,300,000 needed for the average state. We predict economic competitiveness for I-C units $3,174,000 larger than 1000 MWh and therefore * recommend for Wisconsin, as an example, Also provided in house studies and two I-C units of 4000 MWh each. Larger staff. units are less expensive per unit of storage but lack the reliability of During this period faculty and staff redundant smaller units. Thus compromise from five engineering departments have between size, cost and redundancy would worked on the storage project. There be made after operating experience is have been 13 faculty, 12 research asso- obtained. ciates and assistant scientists (post- doctoral s), 5 permanent staff, and 32 The magnitude of U.S. storage needs is research assistants (graduate students). iaken as 50 states x 4000 MWh * 200,000 MWh, The project was fortunate to have avail- which assumes that the average state wants able specialists in all necessary disci- storage at half the Wisconsin rate. plines: superconductivity, cryogenics, Wisconsin is fortunate in having 30% of its rock mechanics and geology, metallurgy power capacity from inexpensive new base and materials, power engineering, load generators which are available 90% of mechanical design, and stress analysis. the time, a notably reliable performance. As expected, storage couples well with This paper is one of two describing efficient generation. In addition, storage the Wisconsin effort. In the second couples well with intermittent generation, paper by S.W. Van Sciver the new and as would be available from future photo- recent research and development is pre- voltaic cells, for example. sented. In this paper is presented the more general aspects of the Wisconsin DISPERSED STORAGE project development with emphasis on The magnitude of the utility storage system specifications, system design, needs seems to preclude widely distributed electrical power system design and rock small storage units. Most utility advisors mechanics. Tentative cost projections would not want 80 I-C units of 100 MWh each and project schedules are discussed. and would prefer only two or three units of equivalent total energy. Small units are NEED FOR SUPERCONDUCTIVE STORAGE generally inefficient and costly to build per unit power. I-C units costs scale as REQUIREMENTS Ed/i, where E is the total stored energy. The desirable amount of power from In Table 2 these cost trends are illus- storage in an electric utility system is trated. We can predict 200,000 MWh of 239 storage countrywide in 50 large units time of day metering. It will require would cost 27% of the cost for 2500 $100 M to completely install metars 1n smaller units. In addition, maintenance, the WPL system. We estimate that $100 M siting and environmental controls might would buy a 1500 MWh I-C unit. Such scale even less favorably for many small storage is 10% of WPL peak power for units. 10 hours and would probably eliminate the need for time of day metering with its implied disruption in life style. Table 2. Relative capital costs of I-C storage units. ADDITIONAL CREDITS Size Capital Cost/MWh Supplementary uses for I-C units 10,000 MWh 1.00 in power systems, such as AGC (automatic generation control), transient stability 8,000 MWh 1.08 regarding major disturbances and voltage 5,000 MWh 1.26 regulation have been discussed in our reports. The major use, of course, is 2,000 MWh 1.70 the diurnal storage and release of energy. 1,000 MWh 2.16 What makes I-C storage high quality is its speed of response. No other storage 500 MWh 2.72 system can reverse power direction within 100 MWh 4.63 50 milliseconds. Such speed might prevent system blackouts following losses of load or generators. Load following second by POLLUTION AND SITING second can be provided by an I-C unit One of the very few pollution prob- simultaneously with its major charge- lems associated with I-C storage is the discharge function. Such load follow- electromagnetic interference arising ing is otherwise unavailable and should from the three phase Graetz bridges. greatly reduce the wear and tear on This problem is common to most forms of "old" generators which normally provide electrical storage, especially storage load following functions. batteries, which rely on ac/dc conver- sion. The bridges must be shielded to In summary, load leveling in electric prevent interference on telephone systems. utility systems by large central storage It is much easier to shield 50 interfer- units is needed. Superconductive I-C ence sources than 2500 sources, which again storage may prove to be the best option to mitigates against widely dispersed storage. pumped hydro storage and time of day metering. There is no need to locate storage units near generators, it is only SUPERCONDUCTIVE ENERGY STORAGE SYSTEM required that adequate transmission DESIGN STUDIES lines exist between the storage unit and the power system. EARLY RESULTS The early work between 1970 and 1976 ALTERNATIVES was primarily a feasibility study which indicates that superconductive storage is The two main competitors for load technically and economically feasible. The leveling today are pumped hydro storage results of these early studies have been and load management through time of day published in Vol. I and Vol. II, Supercon- metering. Pumped hydro storage is ductive Energy Storage Reports, University economic and would be attractive wherever of Wisconsin and in the first 19 papers the terrain allows for upper/lower bodies listed in the Wisconsin bibliography. of water and environmental standards can The major results are: be met. There are very few sites avail- able in the central U.S. and environ- 1. Bedrock structure is needed. mental disadvantages are extensive. 2. Cryogenic stability for the con- ductor is required. Time of day metering is a costly metering snd billing process which adds 3. Pool cooling with superfluid nothing productive to a system. Wisconsin helium is preferred. Power and Light, which is the 60th largest utility in the country, is a leader in 240 4. An aliiminum. stabilized NbTi com- Magnet safety will be achieved by posite is planned. subdividing the unit into several 5. The conductor, dewar, dnd asso- separate vacuum and helium sections so ciated structure are to be rippled at that any local problems are isolated to approximately 1 meter radius of curvature. a fraction of the total unit. Energy can be removed quickly from a section by 6. A one layer, thin wall, high mutual coupling to other magnet sections current, solenoid is probably best. in conjunction with some external energy 7. A multi-tunnel, sectored sole- discharge through the Graetz bridge. In noid results in less cold structure and this way magnetic fields and magnetic greater safety. fluxes are maintained approximately constant, thus avoiding induced voltages Since 1976 the project has entered and ac losses. the component research and development phase, as reported in papers 19 to 55 of ROCK MECHANICS the bibliography and in the 1976 Annual Report published in May, 1977. In the The preferred design is a multiple companion Wisconsin paper by S.W. Van Sciver tunnel unit shaped for a favorable field the study of conductor, structure and and force distribution. One example studied is the "C" shaped three tunnel cryogenic systems is given. Here we in- 36 clude the electrical system and rock 1000 MWh unit sketched in Fig. I. mechanics studies. This three tunnel unit minimizes shear loading on the walls by arranging the ELECTRICAL turns in a "C" shape. The magnetic load is transferred to the outer wall of the The external electrical circuitry central tunnel and the inner horizontal consists of three-phase Graetz bridges, walls of the outer tunnels as shown by such as those used on dc transmission the arrows in Fig. 1. lines. By arranging the converters in series and parallel large dc currents and real and reactive power control can be achieved.1 Superconducting short circuit switches across the storage magnet are not planned because 10 hour charge and 10 hour discharges use up the time available. The energy loss in the leads and bridges is a minor loss and can be tolerated for 24 hour periods. If a short circuit is needed then a thyristor bypass switch is probably the best choice. The internal magnet electrical con- siderations are unique to superconducting magnets. Magnet dc voltages up to 10 kV and currents up to 330,000 A are planned for storage magnets with fields up to 2.5 T. Because of the large currents it Fig. 1. Three dimensional view of three is necessary to limit the number of tunnel magnet system. Arrows denote direc- external leads, preferably to only two. tions in which the rock is loaded. The Voltage breakdown in helium liquid and central tunnel is 15 m high; its inner wall vapor, and through and across insulators, radius is 65 m (not drawn to scale). limits internal design voltages to a few hundred volts. The lowest voltage break- Site studies and associated labora- downs occur in helium vapor. Our experi- tory work are under way in order to ments show that 1 kV is the maximum determine criteria for I-C storage units. voltage across reasonable separations of a Site investigations include geologic few mm.5Z Me plan for voltage drops to be mapping, hydrologic studies, core hole taken across solid insulators so that the studies, laboratory and field mechanical helium will not be subjected to large studies, in situ rock stress measurements, potential differences. and finite element analysis. Four core holes have been drilled; in granite, 241 guartzite, rhyolite and carbonates Fracture Geometry and Hydrology. Fracture (dolomite) to depths of between 200 m geometries combine with stress distribu- and 300 m. The data obtained is used to tion to cause unstable zones in the conduct the finite element analysis of tunnels; of particular importance are the rock structure around the three zones where joint systems parallel the tunnel unit. One result is shown in inner walls of the central tunnel and Fig. 2, where surface tensile stresses or the area between the tunnels in the 90° surface shear failures are reduced with profile. Rock masses must have low appropriate rock bolting.36 permeabilities controlled by fractures which can be successfully grouted. Bolting may be necessary in some parts of the inner walls of the central • ft ffli • tunnel, the roof of the upper tunnel, and the floor of the lower tunnel to avoid joint deformation which might C flfl C Ca cause joint aperture increases and water • a flow problems. COST ESTIMATES Over the years cost estimates and O'-W cost reduction engineering research has 0«45* been emphasized. In 1976 a particularly careful cost optimization and design was undertaken. The following tables out- Fig. 2. (a) Potential shear failure zones line engineering progress and develop- before reinforcement, 0° profile, (b) ment over the years. The cost basis in potential shear failure zones before rein- all cases is the 1976 dollar and changes forcement, 90° profile, (c) rock bolt in costs result only from engineering pattern; note that the bolt pattern improvements. The cost in mills/kWh change's with profile, (d) shear failure stored is based on delivering 90% of the zones after reinforcement, 0" profile, stored energy in a 10 hour day. The (e) shear failure zones after reinforce- yearly cost of the unit is taken as 16% ment, 90° profile. of the original capital cost for interest, taxes, dividends, maintenance, etc. and Some general conclusions from the 20% for interest during construction. rock mechanics work are that three major These were typical rates for 1975-76. factors affect the stability of storage tunnels: In Table 3 the initial cost esti- Rock Strength. The rock mass must be mate for storage is 101 mills/kWh stiff enough to contain the magnetic delivered. The copper in the composite loads without deforming so much as to conductor and the stainless steel damage the conductor. Because of joints structure are too expensive. The storage and discontinuities, we have found some cost in mills/kWh delivered is a better reinforcement is necessary even in hard measure than $/KW because different rock masses. discharge times drastically effect the kilowatt rate for a given storage magnet. In Situ Stress. Stress distributions in The reader i s referred to Vol. II for a the tunnel vary strongly with profile complete set of cost presentations. about the tunnels. The least stable In Table 4 the design status in zone appears to be the 90° profile which 1974 shows substantial improvement Is parallel to the greatest horizontal resulting from replacing copper with in situ stress. The 0° profile is less weight aluminum and steel structure perpendicular to the greatest horizontal with bedrock structure. The storage in situ stress. Depths of at least 300- cost estimate of 27.1 mills/kWh is 400 in are necessary to avoid vertical tolerable when one notes that in 1976 and tangential tensile stresses in parts a cost of 60 mills/kWh was attached to of the tunnels. gas turbine peaking power. 242 by better and less expensive reinforced TlMt >. lallltl Kaae «>l|» l> 1171. polyester struts. Mriacttr Moata tlualia Jtatlltala 1.1 Mlli/llk Tiaia t. anlna anl|a, 1170. Cwtactar hillatu n.« Strvctafa can staiaitii suel JMIlaala u.o OaaKUr Ml ItntrntM MMIWW. Nw cajripaiat CMVtrttr Irittn Stantara. J.I ftallaala CwaKUr •1 Rw&aB1"* t.t trjw11' CMUIwr SUtHlfU Stn) ' ftnlitklt t.t Stractara kKt l.t •afrlMratar lf«iM Hailaa •ntlaalt O.I K •ata: kttn iwrMiul Oxaitaaau tf aaaaa Caa««ur. Slnrt aaj ~~*" Hallda Kaat frantfar an. Iraai»art Canclailani Cawoiwt ttlll" an. Ttit Wtl IMaita». »»t Ma> Ir Davaiepaant af rakricatian Elt Ctnlactar aall AvaiUMi 1.1 anil/Ma) Canfcctor M» ESSES- t.t Stnctarw HackCafara SHclaiiitt t.l EKavatiaa Both engineering and cost progress •War Cqutpmt Canarur Irtejti [ItctHul a>M t.t are seen in the above tables. The cost Wigt Statlai Cr>a|aiilc CaatalMr •laalnaa Alia; <>allaila. 0.0 basis is given in Chapter II, Vol. II SBSU- and is purposely conservative. atfrllarator SwarfliiM Hallaa 0.1 uHrj""^ PROJECT SCHEDULE tl«aM Haltaa Mtllakta 0.4 •Stmtt Oa^waM t.l COMPONENT DEVELOPMENT 17.1 allll/u* •all: Contuctar, strati. Hack Hacktnici ant HtfrliaraKr »Mf For the next two years, FY-79 and FY-80, component development and test CaMlinlan: Alwlnui anf Rack tavar. aaia fieoa Prvttactf ncca»ta.li will be completed for the conductor, strut, and cryogenic system. Rock In Table 5 the benefits of engi- studies will continue with emphasis on a neering optimized component design is site near Madison. In FY-81 a substan- evident. Costs are now reduced to tial portion of the funds will be used 18.7 milis/kWh due to the proper to procure the Industrial design of selection of magnetic field, strut component manufacturing and assembly length and conductor. equipment (conductor, struts, cryo- genics). Specialized rock excavation inn «. Ortiatirt fciin. mi. equipment may need to be designed. The system construction equipment CMactar Ml l.talllt/kW will be designed in FY-81. CaiakclaF alialma nu, cirrmit t.l u Rnwractarv Smctan Hack Haaa-i atallcatlta Otyaiopacnt Following the development of the IkMtr Cwjiaatat IrlHaa •vallaala 1.1 components and the design of the « Allay Urtt. A.lllrtla t.l assembly equipment it will be possible •tfrliaratar Si«trf1a<. Htllui t.l to accurately assess the cost estimates lltalt Ikilaa t.l and recommend the future construction Strata t.t of the first model unit. 11.7 allls/UD MODEL UNIT Halt: Cantoctar, stnt. Itfrlgtratar ana Hacl Ikchmln Unaar Omlaaaaiit lawlulaa: Coiavcur Oaaccaataki. At this time we anticipate that the model should be in bedrock and that all In Table 6 the current revised ideally developed fabrication and con- design costs slightly more at 19.2 struction equipment should be used. The milis/kWh. The main advance 1s that model will not be built to show that it the conductor is now deemed to be operates. The magnet is so conservative manufacturabie. Heat transfer data is that operation is absolutely guaranteed. measured to be better than previously For example, the composite conductor predicted and epoxy struts are replaced uses the well developed Fermi National 243 9. "Superconductive Energy storage Inductor-Converter Units for Power Systems," Peterson, H, A., K. Hohan and R. W. Boon, HEE Transactions Laboratory conductor which carries on Power Apparatus and Systems. Vol. PAS-94, Ho. 4, July/August It/*, current without all of our aluminum. pp. 1337-13*6. The additional aluminum insures 1U. "A Look at Superconductive Storage," Peterson, H. A., It. U. lorn, reliability and trouble free operation. N. C. Storck and W. C. Young, Electrical World. Harch 1, Wi, 11. "Wisconsin Superconductive Energy Storage Project," Peterson, tl. A., R. W. Boon and U. C. Young, Proceedings of the American Power The model is to be built to demon- Conference, Vol. 37, 1975, pp. 1046-1057. 12. "Magnet Design for Superconductive Energy Storage for Power 5ystems," strate that the proper construction Bom, R. w., H. A. Hilai, ft. W. Hoses, G. E. Kclntosh, H. A. Peterson, R. L. Willig and W. C. Young, Fifth International Conference on Magnet equipment has been used and to verify Technology (HT-5) Proceedings Roma. (EUltjV Italy, April 21-25, 1975. that costs are known. Laboraton National) del CNLN, July 1975, pp. 477-483. 13. "Configurattonal Design of Superconductive Energy Storage Magnets," Hoses, R. U., Jr., Advances in Cryogenic Engineering. Vol. 21, Plenun Scaling is easy. Since only full Press, 1976, pp. MET*?: 14. "Optionalten of Mechanical Supports for Large Superconductive Magnets," scale conductor is used at full current Hiial, M. A. and R. U. Boom, Advances In Cryogenic Engineering. Vol. 22, at full field at full radius of curva- Plenum Press, 1976. pp. 224-23~2l 15. "Cryogenic Design Elements for Large Superconductive Energy Storage ture there is no extrapolation needed to Magnets," Hllal, H. A. and G. E. Hclntosh, Advances 1n Crvocenic larger sizes. The first unit could be Engineering. Vol. 21, Plenum Press, )976, pp. 69-7?. 16. "Electrical and Mechanical Properties of Dilute Alumiiium-Gold Alloys in the 10-50 MWh range and the demon- at 300, 77 and 4.2 K," Hartwig. K. T., F. J. Worzala and H. E. stration unit to follow in the 100- Jacltson, Advances in Cryogenic Engineering. Vol. 22, Plenun Press. 1000 MWh range. 1976, pp. 472-476. 17. "Low Temperature Resistance Studies on Cyclically Strained Aluntnw," Segal, H. R. and T. G. Richard, Advances 1n Cryogenic Engineering. CONCLUSIONS Vol. 22, Plenum Press. 1976, pp. 486-469. 18. "Superconductive Energy Storage for Tokarok fusion Reactors," L«» J. W., H. A. Peterson and R. W. Boom, Proceedings of the Sixth Symposium on Engineering Problems of Fusion Research. 1975. pp. 291-95, Superconductive energy storage is IEEE Pub. Ho. 75CHI09J-5-NPS. technically and economically attractive, 19. •Superconductive Energy Storage Inductor-Converter Units for Power Systems," Peterson, H. A., EFC Session on SuperfiywhceU and Super- according to the Wisconsin development conductive Storage, Asilonar, California, February 8-13, 1976. and design studies. Storage efficiencies 20. "Magnet Design for Superconductive Energy Storage for Electric Utility Systems," Boon, R. u., B. C. Haimson, H. A. Hllal, R. W. up to 95% seem probable. The further Hoses, d, E. Hdntosh, H. A. Peterson, R. L. Hlllig and U. C. Young, development of components and the design EFC Session on Superflywheels and Superconductive Storage, Asilomar, of manufacturing and assembly equipment California, February 8-13, 1976. 21. "Dilute Aluminum alloys: Their Potential in Superconducting Devices," will provide a more advanced confirmation Hartwig, K. T. and F. J. Woriala, 105th A1KE Annual Meeting, Las Vegas. of this favorable assessment. The first Nevada, February 22-26, 1976. 22. "Determination of Magnetic fields at the Conductor for Solenoids «nd model to be built in bedrock with the Toroids," Moses, R. W. and R. L. WilHg, Proceedings of the CQHPUHAfi Conf. on the Computation of Haonetic Fields. Oxford. England. proper equipment will provide the final March 31-April 2, 1976. confirmation of the use of superconductive 23. "Oilute Aluminum Alloys: Their Potential in Superconducting Devices," Hartwig, K. T. and F. J. Worzala, Proceedings of the Sixth Interna- storage for load leveling in utility tional Cryogenic Engineering Conference. Grenoble. France. Hay 1976. systems. pp. 406-410. 24. "Free Convection Heat Transfer to Supercritical Helium," Hllal, M. A., R. w. Boom and M. H. El-Hakil, Proceedings of the Sixth International Wisconsin Bibliography (1972.78) Cryogenic Engineering Conference_and Exhibition. Grenoble. France. Hay 1976, pp. 327-329. Reports 25. "Transition and Recovery of Cryogenically Stable Conductors," HJU1, 1. Wisconsin Superconductive Energy Storage Project Volume .1, University M. A. and R. M. Boom, Proceedings of the 9th Symposium on Fusion of Wisconsin. July 1, 1974. Technology. Pergamon Press, Oxford and New York, 1976, pp. 67-93. II, Wisconsin Superconductive Energy Storage Project Volume 11, Univer- 26. "Optimization of Current Leads for Superconducting Systems," Hllal, sity of Wisconsin. Januiry 1, 1976. M.-A., IEEE Transactions on Hagnetics. Vol. HAG-13, Ho. 1, January 1977, pp. 690-693. 111. Wisconsin Superconductive Energy Storage Project Annul! Report, Un1vors>ty of Wisconsin, Hay I, 197/. 27. "Reinforced A'uminum as a Superconducting Magnet StaMIIier," Segal, H. R., [EFE Transactions on Hagnetics. Vol. HAG-13, No. 1, January Publications 1977, pp. 109-1. \ 1. -Superconductive Energy Storage for Power Systems," Sow, R. V. and 28. "Problems Associated with the Use of High Purity Aluminum In the H. A. Peterson. IEEE Transactions on Higneties. Vol. HAG-B. Ho. 3, Design of Composite Conductors 1n the Elasto-Plastic Region," September, 1972, pp. 701-701. Ladkany, s. G. and W. C. Young, IEEE Transactions on Magnetics. Vol. HAG-13,. Ho. 1, January 1977, pp. 105-108. ~ 2. "Superconducting Energy Storage," Boom, H. If., G. E. ffclntosh, H. A. Peterson and W. C» Young. Advances in Cryogenic Engineering. Vol. 19, 29. "Properties of RF Sputtered Nb,Ga Superconducting Films," Burt, ft. J. plenu* Press 1974, pp. 117-126. and F. J. Worzala, IEEE Transactions on Magnetics. Vol. HAG-13, Ho. 1, January 1977, pp. 323-326. 3. "Inductive Shielding for Pulsed Energy Storage Magnet's," Hoses, R. U.,Jr. and J. K. Ballon, IEEE Transactions on Magnetics. Vol. HAG-11, No. 2, 30. "A Method for Preventing Pressure Oscillations In Tubes Connecting Harch 1975, pp. 495-*96. Liquid Helium Reservoirs to Room Temperature," Hllal, H. A. and G. E. Kclntosh, Cryogenics. February 1976, p. 122. 4. "External Field Reduction of Superconducting Energy Storage Solenoids," Ballou, J. K. «nd R. H. Moses, Jr.. IEEE Transactions on Hagnetict. 31. "Superconductive Inductor Storage and Converters for Pulsed Power Vol. HAG-11. No. 2. ««rch 1975, pp. 497-499. Loads," Hohan, N. and H. A. Peterson, IEEE International Pulsed Power Conference, Lubbock, Texas, November 9-11, 1976, 5. "Cellular Concrete - A Potential Load-Bearing Insulatton far Cryogenic Applications." R1ch»ro, T. G., J. A. Dobogai, T. 0. Gerhardt and "Site Characterization for Tunnels Housing Energy Storage Magnets," W. C. vouns. IEEE Transactions on Hagnetics. Vol. HAS-11, Ho. 2, Haimson. 6. C, T. W. Doe, S. R. Erbstoesser and G-F. Full. Proceedings of the 17th U.S. Symposium on Rocfc Mechanics. Snowbird, lftan\ Harctt 1975, pp. 500-502. August, 1976, pp. 34-1-9. C. "Flu* Olffuslon Losses 1n StabilUed Conductors," Hllal, M. A. and ft. H. Soon. IEEE Transactions on HagnetJcs. Vol. KAG-11, Ho. 2. "Th1n-Walled Solenoid End effects," WllUg-, R, L. and R. W. Hoses, Or., Harch 1975, pp. 444-54/. presented at the 15th Annual International Magnetics Conference, Intermsg 77, Los Angeles. Calif., June 1977. 1. "Superconductive Energy Storage for targe systems," BOM, R. If., 8. C. Hiinson, C. E. Mclntosn. H. A. Peterson and W. C. loung, UK "Inductor-Converter Superconductive Energy Storage Systems for Electric Transactions on Hagnc.ics. Vol. MAS-11, No. 2. Harch 1975, pp. 4TC4B1. Utility use," Peterson, H. A., R. W. Boom and W. C. Young, presented at the World Electrotechnical Congress, Moscow. USSR, June 21*25, ¥977. B. "Superconductive Inductor-Converter Units for Pjlsed Power Loads," Peterson, H. A., N. Hohan. W. C. Young ind R. w. 8otm. Energy Stc—• "Structural Design for Large Superconducting Magnets," Young. If. C, Compression, and Switching. Plenum Press, 1976, pp. R. W. Boom and S. G. Ladkany, presented at the World Electrotechntcal Congress, Moscow. USSR, June 21-25, 1977. 244 K, "Design of Underground Tunnels for Superconductive Energy Storage," Fun. G. F., T. One and I. C. Haimsor, 18th U.S. Sy*po>1u> on lock Mechanics, Keystone, Colorado, June 22-24, 1977. 37. "Constant Tension and Constant Field Solenoids," M. *. El-Derlni, R. W. tool and N. A. Hilal, Advances 1n Cryogenic Engineering. Vol. 23, Plenum Publishing Corp., Hew Yorto, 1970. pp. B8-l». 31. "Hew Crltorla for Refrigerator and L'quefler Cycles Design," H. A: Hilal, presented at CEC-ICHC. Boulder, Col., Aug. 2-5. 1977. 99. "Heat Transfer to Subcooled Liquid Helium," Ibrahim. E. A.. R. y. Boom and G. E. HclntDsh, Advances In Cryogenic Engineering. Vol. 23. Plenum Press, 1978, pp. 333-339. 40. "Dielectric Strength of Helium Vapor and Liquid at Temperatures Between 1.4 K and 4.2 K," Hwang, K. F.. Advances In Cryogenic Engineering. Vol. 23, Plenum Press, 1978, pp. 110-117. 41. "Resistance to Strain Degradation 1n Preliminary IIWWC TF Coll Conductors for Fusion Reactors," Kong, S. 0., P. F. Hichaeison, I. H. Sviatoslavsky, and U. C. Young. Advances In Cryagenic Engineering. Vol. 23, Plenum Press, I9)&\ 42. "ThermodynMlc 0pt1m1ial1on Study of the Helium Hulti-Englne Claude Refrigeration Cycle," Khaiil, A. and G. E. Hclntosh, Advances In Cryogenic Engineering. Vol. 23, Plenum Press, I97B, 43. "High Current Al-T1Nb Composite Conductors for Large Energy Storage Hignets." Ladkany, S. G.. Advances 1n Cryogenic Engineering. Vol. 24, Plenum Press, I97B, 44. "bnpressive Strength of Glass Fiber Reinforced Composites at Room Temperature and 77 K,' Stone, E. L. and w. c. Young, advances 1n Cryogenic Engineering, vol. 24, Plenum Press, 1978, 45. "Kapitia Conductance of Aluminum and Heat Transport from I Flat Surface Through a Large Diameter Tube to Saturated He II," Van Sciver, S. U-, Advances In cryogenic Engineering. Vol. 23, Plenur. press, 197B, pp. 341-348. 46. "The Effect of Thermal Treatments on the Critical Current Oensity of a Commercial NbTI Filamentary Superconductor," Larbalesticr, 0. C, R. Flach and D. G. Hawksworth. Proceedings of the Sixth International Conference on Magnet Technology. Aug. 29-Sept. 2. 1977. 47. "Formation of Superconducting A-15 V-Ge Compound by • Compositt- Olffusion Process," Tachlkawa, K., K J. Burt and K. T. Kartirlg. Journal of Applied Physics, vol. 48. No. 8, August 1977, pp. 3623-25. 48. "Geotechnice! Investigation and Design of Annular Tunnels for Energy Storage," Haimson, B. C.. T. Doe and G. F. Fun, presented at the International Symposium on Storage 1n Excavated Rock Ctntmt September 1977. 49. "Underground Caverns for Energy Storage Using Superconductive Magnets," Hilmson,.B. C, K. T. Hartwig and T. w. Doe, Underground Space. Vol. 2, Pergamon Press, 1978, pp. 137-142. 'Cryogenlcatty Stable' Holloat CoConductorn s Cooled sy Supercritical leliutt." Hllal., H. A. and R.. uW.. IBoon, Proceedings of the Seventh SymposiuSymposium on Engineering ProblemProblem:s of Fusion Research, Knoxvtlie, Tent... October 25-28, 1977, Vol. 1. pp. 695-699. IEEE Pub. Ho. 7JCH12SM-WS. SI. "Design of Energy storage Solenoids for Tokamak Reactors," Cl-Derinl, «. «., K. u. loon and H. A. Peterson, Proceedings of the Seventh 5yanos1un on Engineering Problems of Fusion Research, Knoxvilic, Tern., October 25-28, 1977. Vol. II, pp. 1367-1370, IEEE Pub. No. 77011267-4-HPS. £2. 'Dielectric Breakdown of Liquid and Vapor Hellun In Bulk and Across Epoicy Insulation," Hwang, K. F. and S. 0. Hong, Proceedings of the Seventh Symposium on Engineering Problems of Fusion Research, Knoxville, Tenn., October 25-28, 1977, Vol. II, pp. 1531-1534, IEEE Pub. r». T7CH1267-4-NPS. 53. "F;t1gue Tests with Small Colls of Filamentary ffl>-Sn,r Larbalestler, D. C. and S. 0. Hong, Proceedings of the Seventh Symposium on Engineer- ing Problems of Fusion Research, KnoKiHlle, Term., October 25-28. 1977. Vol. II. pp. 1260-1262. IEEE Pub. No. 77CH1267-4-HPS. 54. "Compressive Strength of Glass Fiber Reinforced Composites at Room Temperature, 77 K and 4.2 K," Stone, E. L. and V.. C. Young, Proceedings of the Seventh Symposium on Engineering Problems of Fusion Research, (nonville. Term.. October 25-28, 1977, Vol. II. pp. 1510-1512. IEEE Pub. Ho. 77CH12S7-4-KPS. 55. "CryosteblHution of Large Superconducting Magnets using Pool Boiled Helium II," Van Sclver, 5. ».. Proceedings of the Seventh Symposium on Engineering Problems of Fusion Research. Knoxville, Tenn., October 25-28, 1977, Vol. I, pp. 690-694, IEEE Pub. Ho. 77CHI267-4-W5. 245 RECENT COMPONENT DEVELOPMENT STUDIES FOR SUPERCONDUCTIVE MAGNETIC ENERGY STORAGE S. U. Van Sciver Engineering Experiment Station University of Wisconsin Madison, Wisconsin 53706 ABSTRACT For the past two years the Wisconsin Superconductive Energy Storage project has been in the component development phase. Work during this period has been principally directed toward development of the superconducting composite conductor and the fiber reinforced strut needed to transfer the loads from the magnet to the bedrock structure. Conductor work has involved design as well as research into manufacturing techniques suitable for mass production. In a related activity, heat transfer studies in super- fluid helium has shown potential for substantially better conductor stability than occurs in normal helium. Compressive testing of fiber reinforced composite materials has been employed to evaluate the commercially available products. A current survey of progress in the component development studies at Wisconsin provides the main content of the present paper. The direction of future activities is also discussed. INTRODUCTION presented here to provide a basis for discussion of the components. In 1976, The Wisconsin Superconduc- tive Energy Storage Project entered the A superconductive energy storage component development phase. This magnet for electric utility load leveling activity, which has followed the early and peak shaving has been estimated to be system optimization and conceptual commercially competitive with alternate design studies, has required initiation storage schemes for units larger than of development programs in several key 1000 MWh.6 In order to make the device areas. Based on the conceptual design cost competitive, it must be buried in studies, development has centered around bedrock, which provides the structural the superconductive composite conductor support at the rather low cost of exca- and the fiber reinforced composite vation. A schematic of a superconductive structural supports of the storage magnet. energy storage unit is shown in Fig. 1. The principal goal of this phase is to The dimensions of a 1000 MWh magnet are identify the best materials and manu- approximately 70 m radius and 30 m facturing techniques so that equipment height, for a magnetic field at the can be designed to mass produce the center of the solenoid of 5 T. The components for a full scale energy solenoid is segmented into three storage magnet. separate magnets to minimize the cold structure and to reduce the shear It is the purpose of the present loading on the rock wall. paper to provide an up-to-date survey of the progress that the Wisconsin group has made during the component development phase of the project. The discussion will include the directions for future activities. COMPONENT DESIGN The Wisconsin approach to the design of superconductive energy storage magnets has been discussed in detail in the adjoining article' as well as in numerous papers in the literature. "° A Fig. 1. Conceptual design of a super- brief review of the overall concept is conductive energy storage facility. 247 A detailed view of the magnet and sup- the superconductor since it is currently port structure are shown in Fig. 2. -The available in industry. magnet consists of a single layer solenoid wound from aluminum stabilized-NbTi super- HIGH STRENGTH ALUMINUM SHELL conducting composite conductors cooled by pool boiling superfluid helium at 1.8 K. The conductors have a round cross section and are rippled in a radius of curvature of approximately 1 m. The ripple is necessary to reduce the hoop stress in the HIGH STRENGTH conductor. The radial load is transferred ALUMINUM WEB to the bedrock through fiber reinforced composite struts. If the magnet were not rippled in the above fashion, it would SUPERCONDUCTING require a more complicated and costly sup- FILAMENTS port structure. HIGH PURITY ALUMINUM STABILIZER Fig. 3. Composite NbTi-aluminum conductor design. Stability of a magnet is a basic requirement so that it can operate without unplanned quenching. Full cryo- genic stability must be the approach for large superconducting magnets since the total energy stored is very large. By full cryogenic stability we mean that the conductor contains sufficient high purity normal metal to carry the current should the superconductors go normal for any reason. We determine the required normal metal by using the cryogenic stability criteria10 \c p/A = qS O) Fig. 2. Detailed view of magnet support where the joule heat generated per unit structure. length, I2 p/A, must be dissipated in The superconducting composite conduc- the cryogenic coolant through the con- tor requires an active research and develop- ductor surface, S. The allowable heat ment program. The end product of this transfer, q, is determined by the program must be a high current conductor properties of the helium. which can be manufactured in continuous lengths and at moderate cost. The Wisconsin An important design decision early answer to the above requirements is shown in the project was to propose to cool schematically in Fig. 3. The diameter of the magnet with liquid helium at 1.8 K the full scale conductor is 8.0 cm which rather than at its normal boiling point has a design current of 286,000 A.9 The of 4.2 K. This decision was motivated overall current density of the conductor by two considerations. From the stand- is 5700 A/cmz. point of cost, the superconductor required for 1.8 K operation is roughly 11 The conductor has three principal half that needed for 4.2 K. This components. The superconductor, located reduction in NbTi more than offsets the near the outer surface, consists of a added refrigeration for the lower braid of NbTi copper composite very temperature operation. The other aspect similar to that used by the National to the decision was that at lower tem- Accelerator Laboratory in the Energy peratures, T < 2.2 K, helium becomes Doubler Project. At this time, no superfluid having substantially better development programs are required for heat transfer characteristics. This cooling concept requires further research 248 to evaluate the heat transfer charac- in a polymer matrix such as polyester or teristics of superfluid helium. epoxy. The ultimate strength's of these materials ure typically greater than High purity aluminum has been chosen 100,000 psi (690 MPa) while thermal as the stabilizing material in the energy conductivities are in the range 0.1 to storage magnet conductor. This choice is 1 W/mK. principally because aluminum can be easily and cheaply manufactured in the very pure state. In addition, aluminum has a lower magneto-resistance than copper. The dis- advantage of using high purity aluminum is that it has a very low yield strength. Under the loading which is present in the energy storage magnet, the high purity aluminum yields completely and must be contained. The conductor is designed with a skin of intermediate strength aluminum to contain the stabilizer. The structural cruciform, the third component of the composite conductor, is necessary to carry the hoop tension between the support struts. Also, the conductor must be able to transfer loads to the struts requiring bearing stress capabilities. The star cross section has been chosen to insure against collapse of the conductor at the point where it transfers loads to the strut. SECTION *-* Fig. 4. Conductor, dewar and strut Finally, the conductor has been assembly. designed with a circular cross section to permit better manufacturability. Each component can be extruded separately and COMPONENT DEVELOPMENT assembled with a combination of drawing CONDUCTOR and swaging processes. These procedures can be carried out nearly continuously At Wisconsin, we have demonstrated allowing very long lengths to be manufacturability of the conductor concept manufactured. discussed above. In a small production facility, we have been able to produce The design of the fiber reinforced quarter scale test samples. A photograph composite strut is shown in Fig. 4. of a completed conductor sample is shown Since the strut separates the conductor in Fig. 5. The cruciform is extruded in windings at liquid helium temperatures industry from 6061 aluminum. Wedges of from the ambient temperature rock wall, high purity aluminum are then inserted it must have both a high strength and a with the superconductor, also purchased low thermal conductivity. from industry. The entire assembly is subsequently inserted into a tube of in- The design scheme calls for cooling termediate strength aluminum, such as the strut at intermediate points, 11 K 6063, and swaged together to form a and 70 K, to minimize the heat load at tight monolytic unit. Consideration is liquid helium temperatures.IZ Refriger- being given to alternative aluminum ation power required to recover the alloys particularly for the skin where losses through the strut amounts to the strength and thermal conductivity must major loss mechanism in the energy both be high. storage magnet. In addition, the choice of material must allow easy low Future programs in conductor cost manufacture since many struts are development are directed toward studying required for an energy storage unit. mechanical, electrical and thermal pro- perties of the conductor samples which Apparently the best material for are being produced. A 1,000,000 lb the strut is a composite of glass fibers testing facility and a 3 T large bore 249 AT(K) Fig. 5. Quarter scale conductor sample Fig. 6. Heat transfer curves for super- manufactured at Wisconsin. fluid and normal fluid helium. superconducting magnet are available to This choice is principally due to their investigate the performance of this high strength to thermal conductivity conductor. ratio (a/k), which is a figure of merit for cryogenic structural materials. At CRYOGENICS Wisconsin, we have been investigating commercially available glass fiber rein- The volume of high purity aluminum forced composites to determine which would required to stabilize the magnet conduc- be best for a strut. This survey has in- tor is largely determined by heat transfer volved compressive testing at room as well to the helium bath. The maximum design as cryogenic temperatures. Both ultimate surface heat flux (often the recovery strengths and cyclic fatigue data have been heat flux in a boiling experiment) must collected on five different composite not be exceeded in steady state. At materials. Wisconsin we have been evaluating the heat transfer behavior of superfluid Ultimate compressive strength data helium to compare its performance with at room temperature, 77 K and 4.2 K are normal pool boiling at 4.2 K. Plotted shown in Fig. 7. In general, th'.- uni- in Fig. 6 are heat transfer and boiling directional composites are stronger than curves for superfluid and normal fluid the cloth reinforced. Strengths of the helium. For normal helium at 4.2 K, the unidirectional composites are typically recovery heal flux is around 0.2 W/cm2.13 100,000 psi at room temperature in- Substantially better heat transfer is creasing to nearly 200,000 psi at 77 K observed in superfluid helium where the and below. Cloth reinforced materials recovery heat flux is about 0.7 W/cm2 at are roughly half as strong ranging from 1.9 K and 0.02 atm (saturated vapor 50,000 to 100,000 psi ultimate strength pressure). Further enhancement in the at reduced temperatures.16 recovery heat flux is achieved by pres- surizing to around 1 atm, where An important observation of this recovery is observed to occur at z 14 lb data is that the strengths of materials 1.9 W/cm . ' This improved heat are relatively independent of whether transfer means that less normal metal is the matrix is epoxy, polyester or vinyl required to stabilize the conductor than ester. This effect is important since would be needed in normal helium, further the cost of polyester is substantially reducing the overall conductor cost. lower than epoxy. In addition, the polyester is more durable under cooldown STRUCTURE being less inclined to crack. Thus, based on the ultimate strength measure- The best materials for the structural ments it appears that glass reinforced member in an energy storage magnet appear polyester is the best choice for the to be glass fiber reinforced composites. strut. 250 Future activities in the structural =115 support research and development program 200 - are to test model struts under compressive and shear loads. Initially the work !« * will be carried out at room temperature, -a <=> with later modifications allowing for 150 I D cooling the strut to low temperatures. V m- 10 ~ • , Half scale struts will be tested up to 1,000,000 lbs total load. O o 100 § : "5 CONCLUSION x • • " The principal achievements during the " OEP0XY- CLOTH - 5 first two years of the component develop- ment phase of the Wisconsin Superconduc- A POLYESTER A, ROD 1 tive Energy Storage project have been ZopOLYESTER A, ROD 5 - V POLYESTER B summarized above. Activities in the - O VINYL ESTER development of the superconductor com- 0 i i posites conductor have shown manufactur- 0 100 200 300 ability of a fully stable aluminum-NbTi TEMPERATURE (K) composite to be cooled by 1.8 K superfluid helium. Structural support development Fig. 7. Average ultimate compressive has tentatively selected glass rein- strengths versus temperature for fiber forced polyester as the strut material. reinforced composites. Testing of the conductor and strut are planned in the near future. The ulti- Since the energy storage magnet mate goal of these activities is to varies its loading over the diurnal develop manufacturing equipment in cycle, it is necessary that the strut be industry for producing the components of designed to withstand the 10b cycles an energy storage magnet. that occur during the lifetime of a unit. Therefore, cyclic fatigue measurements have been carried out REFERENCES on the glass-fiber reinforced polyester. Results of these measurements are shown 1. R.W. Boom, paper V-l, this con- in Fig. 8 which is a plot of the peak ference. stress versus median fatigue life of the sample. The allowable peak stress under 2. Wisconsin Superconductive Energy lO1* cycles is approximately 70% of the Storage Project, Vol. I, Engineering ultimate strength for both the room Experiment Station, University of temperature and cryogenic temperature Wisconsin, July 1974. data.17 3. Wisconsin Superconductive Energy Storage Project, Vol. II, Engi- neering Experiment Station, Uni- versity of Wisconsin, January 1976. 4. Wisconsin Superconductive Energy Storage Project, Annual Report, Engineering Exper:ment Station Report No. 47, University of Wis- consin, May 1977. 5. S.W. Van Sciver and R.W. Boom in Proceedings 21st Midwest Symposium on Circuits and Systems, pp. 615- 619 (1978). 6. H.A. Peterson, R.W. Boom and W.C. 10' 10* 10° Young in Proceedings of World MEOIAN FATIGUE LIFE (CYCLES) Electrotechnical Congress, June 21-25, 1977, Moscow. Fig. 8. Peak compressive stress versus media fatigue life for glass/polyester composite. 251 7. R.W. Boom, G.E. Mclntosh, H.A. Peterson and W.C. Young, Advances in Cryogenic Engineering 19, 117 (1976). 8. R.W. Boom, B.C. Haimson, G.E. Mclntosh, H.A. Peterson and W.C. Young, IEEE Trans, on Magnetics, HAG 11, 475 (1975). 9. S.G. Ladkany, Advances in Cryo- genic Engineering, 24 (to be pub- lished). 10. Z.J.J. Stekly, Journal of Applied Physics 37, 324 (1966). 11. P.E. Hanley and M.N. Biltcliffe in Proceedings Fourth Interna- tional Cryogenic Engineering Conference, p. 224-226 (1972). 12. M.A. Hilal and R.W. Boom, Advances in Cryogenic Engineer- ing 22, 224 (1977). 13. D.N. Lyon, Advances in Cryogenic Engineering J£, 371 (1965). 14. S.W. Van Sciver in Proceedings 7th Symposium on Engineering Problems in Fusion Research, pp. 690-694 (1977). 15. S.W. Van Sciver, Cryogenics JjS, 415 (1978). 16. E.L. Stone and W.C. Young, ref. 14, pp. 1510-1514. 17. E.L. Stone, L.O. El-Marazki, and W.C. Young in Nonmetallic Materials and Composites at Low temperatures (to be published). 252 PROJECT SUMMARY Project Title: Power System Stabilization Using Magnetic Energy Storage Dynamic Characteristics of the BPA System Principal Investigator: R. L. Cresap Organization: Bonneville Power Administration P. 0. Box 3621 Portland, OR 97208 503/234-3361 Ext. 4419 Project Goals: To determine the dynamic characteristics of the Bonneville Power Adnini strati on (BPA) electric system in order to establish the feasibility of using a superconducting magnetic energy storage (SMES) unit to damp power oscillations in the western U. S. power system. Project Status: Field tests on the BPA system with the dynamic brake were conducted to determine the dynamic characteristics of the western U. S. power system. The analysis of the test data and the operating experience with v*e modulation of the DC intertie show that a small magnetic energy storage unit will provide damping of power oscillations for the pacific AC intertie. The AC intertie power variation together with the system transfer function allowed the sizing of a SMES unit. It was found that a 30-MJ/10-MW unit is adequate for damping 99% of the AC intertie power oscillations. The project has been completed. Contract Number: None Contract Period: None Funding Level: Not Applicable Funding Source: BPA (in-house) 253 POWER SYSTEM STABILITY USING SUPERCONDUCTING MAGNETIC ENERGY STORAGE DYNAMIC CHARACTERISTICS OF THE BPA SYSTEM R. L. Cresap and J. F. Hauer Bonneville Power Administration P.O. Box 3621 Portland, Oregon 97208 ABSTRACT . Modulation of the Pacific HVDC Intertie has shown that a small amount of control can provide damping for the western system, which has a history of negatively damped oscillations. Because an extended outage of the DC Intertie could reduce the capa- bility of the AC Intertie an alternate source of damping is desirable. This could be done with a small special-purpose SMES unit. In order to establish the size of the unit, field tests were conducted to determine the dynamic characteristics of the western system. Analysis of the data indicates that, under normal operating conditions, a 10-MW device would provide damping equivalent to dc modulation. INTRODUCTION can be achieved by control of an ac-dc converter . The availability of this The western power system has a long additional damping was a key factor in history of negatively damped syn- permitting a 400 MW uprating of the chronizing oscillations. This tendency Pacific AC Intertie. The possibility to oscillate is a factor which can exists that loss of the DC Intertie, such significantly reduce the maximum trans- as occurred in 1971 when an earthquake fer capability of the Pacific AC Inter- partially destroyed the Sylmar Converter tie. Because surplus hydro-generation station, could cause a reduction of the in the Northwest is often used to dis- AC Intertie transfer capability. place oil-fired steam generation in the Southwest it is important that this Due to the use of an ac-dc converter, Intertie have the maximum possible SMES has excellent dynamic response rating. characteristics. Experience with dc modulation has shown the amount of Modulation of the Pacific HVDC Intertie control necessary during normal opera- has shown that large amounts of damping tion of the system to be quite small 254 -. (5 MW peak to peak). As a result it 1,400-MW interconnection between the appears possible to provide an alternate Pacific Northwest and the Los Angeles source of system damping using a small area. Figure 1 shows the Pacific AC and special purpose SMES unit. Pacific HVDC Interties. DESCRIPTION OF THE Portion" - ,j WESTERN POWER SYSTEM \\ \OBEOON The western power systen, as comprised Pacific ^SS\\ NEVADA by the Western Systems Coordinating Ocean Council (WSCC), includes all or part of the 14 western states, and British Columbia. These interconnected systems operate about 75,000 miles of trans- mission lines, 115 kV and higher, and have a total installed generating capa- city of about 92,500 MW. Fig. 1 Pacific Northwest-Pacific The WSCC system can be characterized by Southwest intertie system. its diversity. The Pacific Northwest is a winter peaking region with predomi- During the spring and early summer high natingly hydro generation, while the stream flows in the Northwest dictate a Pacific Southwest is a summer peaking heavy export of surplus hydro generation system which has primarily oil-fired to the Southwest. Because this dis- steam generation. In order to take places high-cost oil-fired steam genera- advantage of this seasonal load diver- tion in the Southwest, large economic sity and the availability of surplus and energy conservation benefits are Northwest hydro generation, the northern realized. At other times, this system and southern portions of the WSCC of interties provides a capability to system are connected by a system of exchange off-peak energy. interties. The 500-kV Pacific AC Inter- tie is the primary alternating current In order to enhance the transient sta- interconnection between the Northwest bility of this interconnection, BPA and Southwest. This Intertie is com- installed a 1,400-MW braking resistor in 2 posed of two lines with a total rating its system . During a large transient of 2,500 MW. A system of 230-kV and swing the brake is applied to the system 345-kV ties extending around the eastern for 30 cycles, thereby decelerating side of the system also connects the two Northwest generation. Stability studies regions. In addition to the ac inter- have shown that this improves the trans- ties, the + 400 kV bipolar, direct ient stability of the north-south inter- current Pacific HVDC Intertie provides a connection by about 900 MW. 255 CAUSES OF POOR SYSTEM DAMPING The primary source of negative damping of the one-third Hz swing mode associ- Traditionally, the stability of power ated with the Pacific AC Intertie is the systems has been assessed in terms of response of generator excitation sys- maintaining synchronism between the tems. The negative damping occurs when various parts of the system, both in excitation systems correct changes in steady-state (steady-state stability) machine terminal voltages caused by and following a severe disturbance such swings in power angle between the inter- as a 3-phase fault (transient stabil- connected systems . These fluctuations ity). Usually, if the system survives in excitation cause sizeable negative the first swing, the transient subsides damping torques to be exerted on the because of the natural system damping rotating masses. Other less important resulting from machine windings, tur- sources which contribute to the negative bines, loads, etc. damping of this swing mode include the high-frequency response of hydro gover- Control systems on generators can lead nors and certain types of loads. to another type of instability. Norm- Figure 2 shows a typical example of a ally, in tightly-connected systems, the negatively damped oscillation. frequency of electromechanical swing 1820 • modes range from about 1 Hz to 2 Hz and 2 1780 - are usually well damped. However, when large systems are connected by long, relatively weak interties, lower fre- quency modes result. The response of generator controls to the synchronizing swings associated with these low- frequency modes can produce sufficient negative damping to cancel the natural 0 10 20 30 40 50 60 70 80 90 100 110 120 damping of the system. When this hap- TIME IN SECONDS pens, oscillations of increasing ampli- Fig. 2 Negatively damped Pacific AC tude will occur (dynamic instability). Intertie oscillation Although a system could operate beyond its transient stability limit provided SIZING THE SMES UNIT no . severe disturbances such as 3-phase faults occur, it could not operate Because the rating of the SMES unit beyond its dynamic stability limit. The depends on the size of the synchronizing constantly occurring small changes in swings to be damped, the types of dis- operating conditions (loads, voltage turbances which occur in a power system levels, etc.) would give rise to need to be considered. Disturbances oscillations of increasing amplitude. range from random load changes and 256 switching of lines and shunt compensa- In general, the response of the system tion occurring during normal operation, to modulation of a SMES unit would be to large disturbances such as faults and different than its response to dc modu- loss of major generators. Experience lation. As a result, the rating of the with dc modulation has shown that, SMES unit required determination of a during normal operation, strong damping system transfer function obtained from a of the swing mode associated with the test, design of a suitable controller, Pacific AC Intertie can be accomplished am. measurements of the statistical by small changes in dc power. However, characteristics of load variations. transient stability studies indicate that to provide the same degree of LOCATION OF THE SMES UNIT damping for swings caused by serious AND DETERMINATION OF A faults would require very large changes TRANSFER FUNCTION in dc power. The efficient control of a swing mode Fortunately, large disturbances occur requires that the associated torques infrequently, and system nonlinearities exerted on machine rotors be maximized. cause large swings to be better damped As a result, the best location for than small ones. In any case, it would control of the AC Intertie swing mode is be prohibitively expensive to provide near the center of northwest generation. sufficient capacity in a dedicated SMES unit to strongly damp a large swing. On Transient stability studies to determine the other hand, random load changes a site for the dynamic brake showed that occur constantly and the resulting a good location was Chief Joseph Sub- swings must be damped. station in North Central Washington. Favorable operating experience with the As a result, a criterion was used to brake, together with its availability to size the SMES unit that would insure perform the testing needed to establish damping during normal operation. Be- a system transfer function, also made cause of the stochastic nature of Chief Joseph an attractive location for Intertie swings, caused by random load the SMES unit. variation, it was decided that the rating of the SMES converter should be 3 A 0.5-second brake application produced standard deviations, or be inside its the AC Intertie response shown in limits 99.7-percent of the time. Figure 3, and the Bode plot shown in Figure 4 was obtained from a Fourier While experience with dc modulation analysis of this data. A transfer provides some insight, it cannot be used function was fitted to the frequency directly to establish the rating of a domain data using a criterian which SMES unit. The response of a system included both gain and phase . Fitting depends on the location of the input. was done in the frequency domain rather 257 Table 1. Poles and zeros for AC Inter- tie response to Chief Joseph brake application. Poles Zeros -0.295 0.084 -0.422 + jO.374 -0.336 + jO.782 -0.254 +j2.214 -0.386+ j3.240 TUM (N HCONOS -0.445 + J3.560 -0.365 + J4.051 Fig. 3 Time response of AC Intertie -0.519+ j4.989 -1.135 + J5.700 phase current to Chief Joseph brake -1.123+ j6.650 -25.484 +j7.294 application. -1.570 +j7.338 -0.134 +j7.670 -0.788 + J9.089 -0.456 + J9.168 Gain = 0.01065 AC Intertie Phase Amps/ Brake MW The third pole is the AC Intertie swing mode. Remaining poles are needed to describe the multitude of high frequency swing modes associated with tightly coupled groups of machines. This trans- fer function has the time domain re- Fig. 4 Frequency response of AC Inter- sponse shown in Figure 3. Because tie to Chief Joseph brake application. "Hanning" was used to obtain smooth frequency response data, the damping of than the time domain because accurate the transfer function response is some- gain and phase replication was needed what greater than the measured response. over a wide frequency range in order to design the compensator. Complex values COMPENSATOR DESIGN for the poles and zeros are listed in Table 1. To get maximum benefit from the SMT'f; unit it is necessary that control sig- The poles describe the damping and nals have the proper gain and phase frequency of the modes of the system, relationship to AC Intertie swings. while the zeros provide the proper model Because the objective is to damp a composition. The first two modes are specific swing mode, it is desirable believed to be due to the dynamics of that components of the control signal hydro-plant water passages, and the due to other swing modes and noise be composite response of speed governors. minimized. These objectives were suc- 258 cessfully achieved with dc modulation. ginary axis of the s-plane (undamped). As a result, it was anticipated that a Also the zeros closest to these poles similar control system would be used for were moved to the imaginary axis. The the SMES unit. compensator parameters employed for dc modulation were used as the starting Figure 5 shows the compensator used for point for determining values for the dc modulation . For the SMES system the SMES system. Figure 6 shows the cor- input to the compensator would be phase responding root-locus where the sub- current magnitude derived from fast scripts c and s have been used to dis- responding current transducers located tinguish compensator poles and zeros at the northern terminal of the AC from those of the system. The "best Intertie. This signal would be differ- gain" was choosen to maximize the damp- entiated and telemetered to Chief ing of the AC Intertie mode. An exten- Joseph. The low-pass filter is used to sive parametric study showed that, with control bandwidth, and the lead compen- a suitable gain change, dc modulation sator provides additional signal shap- compensator values were also best for ing. High-frequency swing modes (1 Hz the SMES system. This is due to similar to 2 Hz) are removed by the notch fil- system transfer functions aad a compen- ter. The output is applied to the sator design which is insensitive to regulator of the SMES converter. parameter changes. ' ""LD1 "LP - •2 • V • -2m \tt- LOW-PASS DIFFERENTIATOR CAIN FILTER :o.o rldluu/tacond 20.0 ndlus/sacond "LDl o.o radlws/stcond "IB 1.5 rsdlans/second 50.0 ^rsdUns/sacond)* 50.0 (ridl.ni2 -1 I.a seconds 10.0 IVFRSF OECAV TIME CONSTANT IN SEC Fig. 5 DC modulation compensator. Fig. 6 Root-locus of dc modulation The SMES unit was assumed to have an compensator contolling SMES unit. instantaneous response over the fre- quency range of interest. The system C - Compensator dynamics transfer function, listed in Table 1, S - System dynamics was pessimistically modified by placing 0 - Best gain (0.46) the AC Intertie pole pair on the ima- 259 NOiSE CHARACTERISTICS OF THE Table 2. Poles and zeros for spectrum AC INTERTIE of AC Interie power variations. Fourier analysis of AC Intertie power Poles Zeros variations, measured under a wide range of operating conditions, provided spec- 0 -0.167 + j0.996 tral characteristics of the response of -0.136+ jl.028 -0.137 + jl.854 the AC Intertie to random load switch- -0.290 + J2.266 ing. Figure 7 shows a typical spectrum with HVDC modulation off, fitted with a Gain = 5.8 5th-order model (Table 2). At low frequencies the effect of random load RATING OF THE SMES UNIT switching is evident, producing the initially linear decline with log The AC Intertie noise spectral informa- frequency like the spectrum for integrated tion, together with the system transfer white noise. The peak at 0.35 hertz is function and compensator, permits estima- amplification of load switching noise tion of the converter rating of the SMES by the AC Intertie swing mode. The unit. For analysis purposes noise activity at 0.17 hertz is of undetermined spectra can be modeled as the output of origin. However, sustained oscillations a transfer function T (s) with white have recently been observed at that noise as the input. Referring to frequency. Figure 8, modeling the feedback effect of the SMES unit requires conversion of the noise spectra to be an input at Chief Joseph. Tg(s) is the system transfer function listed in Table 1 with a gain change to convert A I._ to APA(;. Tg(s) is Tg(s), possibly modified to represent changes in system dynamics such as making the Intertie mode nega- a 1 is- tively damped. '°1 Fig. 7. Spectrum of AC Intertie power Fig. 8 Model for sizing SMES unit, variations 260 This technique was tested by using noise spectra obtained with dc modulation out of service to predict the spectra which would have been observed with modulation in service. These predictions were compared to later measurements made 20 with modulation in service, with excel- lent results. Consequently, it appears to be a highly accurate method for determining the converter rating of the SMES unit. Figure 9 shows the effect of the SMES noise at compensator output, unit on AC Intertie noise. Also shown with loop closed is the spectrum for the output of the SMES unit. By Parseval's Thecrm the area under the power spectrum of the output of the SMES unit is its variance, 1.0 or standard deviation squared. Standard FREQUENCY IN HERTZ deviations were calculated for a number of noise spectra, obtained unaer a wide Fig. 9. Noise Spectra range of operating conditions, and with postulated changes in system dynamics. Values ranged from 2.5 MW under normal Also such an application would provide operating conditions, to 3.9 MW for valuable information about the relia- high system noise and assuming a nega- bility and maintainability of super- tively damped Intertie mode. As a conducting equipment in a power system result a 10 MW converter would provide environment. 3-standard deviations for most system conditions. While the primary purpose of SMES is load-leveling, it is well suited to CONCLUSION control applications. Because stability problems often impose serious operating System measurements and operating constraints, this capability can be experience with dc modulation show that valuable. Early demonstration of the a small special purpose SMES unit coma control capabilities of SMES will add an provide danping for the western system. important dimension to the technology, Because the damping capability of dc and may lead to the development of a modulation would be lost in the event of variety of special-purpose control an outage of the DC Intertie, an alter- devices. nate source of damping is desirable. 261 KEFERENCES R.L. Cresap, D.N. Scott, W.A. Mittel- stadt, and C.W. Taylor, IEEE Trans. Power Apparatus and Systems. PAS-97, 1053 (1978). 2M.L. Shelton, P.F. Winkelman, W.A. Mittelstadt, and W.J. Bellerby, IEEE Trans. Power Apparatus and Systems. PAS-94, 602 (1975). 3J.F. Hauer, BPA Report. (19?7). 4F.P. Demello and C. Concordia, IEEE Trans. Power Apparatus and Systems. PAS-88, 316 (1969). 262 PROJECT SUMMARY Project Title: "Energy Storage Systems and Identification of Power Systems Equivalents" Principal Investigators: D. P. Carroll and D. M. Triezenberp Organization: School of Electrical Engineering Purdue University West Lafayette, IN 47907 317-493-3813 Project Goals: One major objective of this project is to investigate techniques for identifying power system configura- tions, load magnitudes, and generation levels, from time series measurements taken at a limited number of points in the system. The other major objective of the project is to develop simplified system models and to explore techniques for controlling energy storage devices to improve the transient performance of power systems. A significant part of this latter objective is the study of low power superconducting magnetic storage systems for damping controls. Project Status: In the research area of power system identification, algorithms are under development for multiple output ARMA process parameter estimation. Some success has occured in estimation of simplified model parameters using data taken from detailed system simulations. In the area of energy storage applications, simpli- fied analytical models have been developed and veri- fied for line-commutated and force-commutated power converters su-itable for interfacing batteries and superconducting magnets with power systems. Simplified and detailed computer simulations are being used to study the dynamics and control of candidate systems involving energy storage. A preliminary design has been completed for a damping control using a low power superconducting inductor-converter unit. Contract Number: EC-77-S-02-4206.A000 Contract Period: Jan. 1977 - Dec. 1979 Funding Level: $517,818 Funding Source: Department of Energy 263 HYBRID COMPUTER STUDY OF A SMES UNIT FOR DAMPING POWER SYSTEM OSCILLATIONS P.C. Krause and O.M. Triezenberg Energy Systems Simulation Laboratory School of Electrical Engineering Purdue University ABSTRACT The work conducted at Purdue University regarding the use of SMES to damp oscilla- tions in the BPA system is reported. This investigation was conducted in parallel with work at LASL and BPA and serves as a verification of the work at BPA. In particular, the dynamic characteristics of the BPA system were simulated on the hybrid computer using the response obtained from the Chief Joseph Dynamic Brake test. The SMES unit was simulated using information supplied by LASL. The compensator was simulated using the design developed by BPA. The hybrid computer study illustrates the dynamic response of the BPA system with the SMES unit in service for sinusoidal variations in AC Intertie power and for random variations of AC Intertie loading. INTRODUCTION In mid 1977 an existing DOE contract from LASL, and the transfer functions G(s), with Purdue University was extended to in- H(s), and C(s) from information provided clude an investigation of a possible ap- by BPA. plication of SMES. In particular, the study considered damping of power oscil- The IC unit is connected to the 230 lations on the West Coast North-South AC kV Chief Joseph Substation through a 12.75 Intertie with an SMES unit and an AC to DC MVA transformer bank with 1 kV line-to- converter. If this application proved to line voltage on both secondary transfor- be feasible, this inductor converter (IC) mers (12 pulse converter). The inductance, unit could provide backup for the Pacific L, Is 2.4 H and the maximum inductor cur- HVDC Intertie which is presently being rent, l(c, Is 5000 A. The commutating modulated to damp the AC Intertie power reactance, Xf, is 0.00784 ohms which is osci1lations. 10% of base impedance. The harmonics are neglected in the representation of the Personnel from LASL, BPA and Purdue converter. That is, the output voltage of University, with H.A. Peterson acting as the converter, V,Q, and the inductor or a consultant, were involved in this inves- output current, I\Q, are the average val- tigation. Important, to this study was ues of the actual variables. The IC unit the background of BPA personnel who had is equiped with an open-loop power control designed a modulation control for the which senses I|Q apd calculates the nec- Pacific HVDC Intertie. Purdue's principal essary voltage, Vref, to satisfy the de- role was to provide verification of tt e sired or reference power, Pref. In this work being conducted at BPA [1]. This was case the reference power is zero where- to be accomplished from a hybrid computer upon the signal from the compensator, Pc, study of a simplified representation of determines the desired power. The refer- the western system. The results of this ence voltage and the output current, are study are recorded in this paper. used to calculate a control signal, ec, which establishs the firing angle of the SYSTEM STUDIED converter and thus V|r,. The power output of the IC unit is denoted Pjr,. That is The system studied on the hybrid com- the power consumed or supplied by the IC puter at Purdue University is shown in unit at Chief Joseph Substation. block diagram form in Fig. 1. The IC unit is represented from information received 264 From information obtained as a result of an 0.5 sec. application of the dynamic brake at Chief Joseph, 6PA established the transfer function between Chief Joseph Substation injection and AC Intertie cur- rent magnitude [1]. This transfer func- tion, which is given in [I], Is G(s) with where one exception; AC Intertie power P|f, Is used rather than AC Intertte current. R(x) = E{x(t)x(t-T)} Although not explained herein, this Is This random signal is generated by samples accomplished by multiplying current In of a random number generator held for 0.1 amperes by 1.07 to obtain AC Intertie sec. each, with the variance of the random power in HW. number generator corresponding to (2.5 Fig. 1 Block diagram of BPA system- transfer function approach The transfer function H(s) provides MW)2. Three sets of data for H(s) were a means of modeling the semi-random power furnished by BPA corresponding to thrrc oscillations recorded on the AC Intertie different inertie loads; 1680, - 1'sC and Monitor. H(s) consists of three blocks in 920 HW. The 1680 MW load (export) Is the series. The output, P^, is added to Pi-!-, case considered In this study. For this the output of G(s), to yield the AC In- condition H(s) is tertie power, Pjy. The input to the trans- fer function H(s) is a bandwidth limited white noise with an average random power Function of 0.25 (MW)2-second, uniform for fre- quencies in the range of 0 to 6 rad/sec. 15.6 Average random power for the signal x(t) s is here defined as 265 1.138 REFERENCES s2 + .3108s + 1.A31 [1] R.L. Cresap and J.F. Hauer, "Power s2 + .9428s -i- 5.565 Systems Stability Using Super-conduc- ting Magnetic Energy Storage Dynamic s2 + .AO53s + 5.395 Characteristics of the Bonnevilie System," presented at the First Annual Mechanical and Magnetic Energy Storage The compensator design was performed Contractors1 Information-Exchange by BPA [1]. The output signal of the com- Conference, Luray, Virginia, Oct. 2k- pensator is added to the power reference 26, 1978. to establish the desired power output of the IC unit. ACKNOWLEDGEMENTS COMPUTER STUDY The authors greatly acknowledge the assistance of H. Boenig, R.L. Cresap, and A frequency response analysis was per- H.A. Peterson. formed with a linearized representation of the converter, making the transfer ratio of P|T(S) to PN(S) equal to H(s)/(l-C(s) G(s)). From this analysis It was estab- lished that for a noise frequency of 2.29 rad/sec. (about 1/3 Hz) the ratio Pp (s)/ Pj-](s) has an amplitude of 2.1k. This fre- quency response calculation is verified in fig. 2 wherein traces of Pn(t), PfJ(O, P|C(t), PjT(t). P.T(t), P (t), V. (t), and l)c(t) are given for PN(t) a 1/3 Hz sinu- soid. In steady state the amplitude of PN(t) is approximately 12 MW while the steady state amplitude of P|T(t) is ap- proximately 5 MW, a reduction by a factor of 2.4. Figures 3 and k show the system with Pn(t) represented as a "bandwidth limited white noise" with S(w) = 0.25 MW2-sec, uniform for w from 0 to about 6 rad/sec. The same mode of operation is depicted in both figures. The recording scale is ex- panded in Fig. k for the purpose of portra- ing the noise input. In all the studies performed, the output of the compensator Pc did not exceed 10 MW and the converter voltage, V|C, did not exceed 2.5 kV. CONCLUSIONS The material presented in this paper provides a verification for analysis and recommendations of BPA. Hence, the con- clusions of [1]. are essentially the con- slusions of this paper. Clearly, if the transfer function between power at the Chief Joseph Substation and the AC Intertie and if the variations in the AC Intertie power are as used in this study then, as recommended by BPA, a 30 MJ 10 MW IC unit (SMES) placed at Chief Joseph bus should be sufficient to damp AC Intertie oscil- lations. JtCCUCHART niliu'mtnl eytttmt DMllon j ! ; t'' ' I i ' ' ' ..4 : . -.. ' • vi ••!••! -i j I--) t- '• -'-+ ; i I i . i J .i . t I • • T • - " PIC r MftnMA/iAAMArtAM MW J VvUVVv Vv v wvi/UVJvVv 25.0 P.T 0 -25.0 E- 12.5 -12.5 .2.5 0 FI9. 2 Response of system sham In FI9. 1 to 5fnuso|d«l noise Input, 267 25.0 p "IT 0 L -25.0 L- .2.5 ?K>J\f,' "• -...sL-^/lijuA 12.5 , Fig. 3 Response of system shown in Fig. I to bandwidor Hroftetf White noise input, 268 ACCUCHART OouM me.. InMnmwnt SyMMii -12.51 25.0 PIT o 4-.. H-LHu KW E- -25.0 '-!•+• t-ri^-:..i"r" 12.5 f.T 0 -12.5 t:rL.L.i:tt"T.ri:rr-: : 12.5, Pc 0 -12.5 1 IP.... •jf :of • •••*-*-+ Ftg. 4 Same as Fig. 3 with expanded time scale. PROJECT SUMMARY Project Title: Superconducting Magnetic Energy Storage (SMES) Principal Investigator: John D. Rogers Organization: Los Alamos Scientific Laboratory University of California Los Alamos, NM 87545 505/667-5427 Project Goals: The goals of the SMES program are two-fold, the first is to design, fabricate, and place into operation a 30-MJ, 10-MW SMES unit for electric utility transmission line stabilization on the Bonneville Power Administration (BPA) system by 1982-83. The second goal is to design and have constructed a 10-to-50-MWh SMESunit. This unit will be for a completely detailed engineering prototype demon- stration of an electric utility diurnal load leveling system. Project Status: The conceptual engineering desiqn of the BPA SMES system is complete. An engineering specification design for the 30-MJ superconducting coil will be complete by 10-1-78, and an RFQ for fabrication design and actual coil fabri- cation will be issued within 60 days. The superconducting wire for the entire 30-MJ coil and 50- and 100-m lengths of prototype 5-kA superconducting cable are on order. The SCRs and sinks for the converter are on order. An RFQ for the converter fabrication will be issued soon. An entirely new reference design for a 1-GWh diurnal load leveling SMES unit is underway. Emphasis is being given to a better understanding of the dewar, the dewar support structure, rock mechanics, excavation details and costs, stabilization of the excavation walls, and high purity aluminum superconductor matrix costs. Contract Number: W-7405-ENG-36 Contract Period: Continuing Funding Level: $1,050,000 FY78 Funding Source: Department of Energy, Division of Energy Storage Systems and Division of Electrical Energy Systems. 271 SUPERCONDUCTING MAGNETIC ENERGY STORAGE* J. D. Rogers and H. J. Boenig Los Alamc Scientific Laboratory of the University of California Los Alamos," NM 87545 ABSTRACT Superconducting inductors provide a compact and efficient means of storing elec- trical energy without an intermediate conversion process. Energy storage inductors are under development for diurnal load leveling and transmission line stabilization in electric utility systems and for driving magnetic confinement and plasma heating coils in fusion energy systems. Fluctuating electric power demands force the elec- tric utility industry to have more installed generating capacity than the average load requires. Energy storage can increase the utilization of base-load fossil and nuclear power plants for electric utilities. Superconducting magnetic energy storage (SMES) systems, which will store and deliver electrical energy for load leveling, peak shaving, and the stabilization of electric utility networks are being devel- oped. In the fusion area, inductive energy transfer and storage is also being devel- oped by LASL. Both 1-ms fast-discharge theta-pinch and l-to-2-s slow tokamak energy transfer systems have been demonstrated. The major components and the method of op- eration of a SMES unit are described, and potential applications of different size SMES systems in electric power grids are presented. Results are given for a 1-GWh reference design load-leveling unit, for a 30-MJ coil proposed stabilization unit, and for tests with a small-scale, 100-kJ magnetic energy storage system. The results of the fusion energy storage and transfer tests are also presented. The common tech- nology base for the systems is discussed. INTRODUCTION Energy storage units can be used to meet the peak-power requirements and to The nondissipative dc-current carrying absorb excess available during periods of capability and the low ac-current gener- low-power demand. To date, only pumped- ated losses of a superconductor place the hydro storage, with units up to 15,000 design of high current inductors within Mwh, has been used very effectively.1 reach of useful application. Only in the Other energy storage technologies include last 15 years has the superconducting chemical storage in the form of batteries technology been sufficiently advanced to and hydrogen, thermal storage, compressed make the applications covered in this pa- air storage, and magnetic storage.'-4 per become practical. Only more recently Economic considerations eliminate iner- has energy transfer within a few seconds tial storage in flywheels for utility or less, into and out of a superconducting applications. energy storage coil been demonstrated. Superconducting magnetic energy Electric utilities experience periodic storage has several advantages. SMES load variations on a seasonal, weekly, and units will have fewer site restrictions. daily basis. The daily maximum and mini- Large SMES units can be constructed for mum loads of a power company typically structural support in the rock formations differ by about a factor of two. The poor near most large load centers, and exten- load factor is an economic burden to the sive new transmission systems will not be utilities because their installed cr acity required. SMES units will have a re- must be capable of meeting the peak demand sponse characteristic of less than a cy- and much of the generating capacity is cle to power system demand, which can idle during periods of low demand. Inex- improve power system stability. With a pensive but inefficient units, such as high efficiency of 90% the energy is peaking gas turbines, are used to meet the stored electromagnetically without an peak loads. intermediate mechanical or chemical ener- gy state. The cost for a large SMES unit (10 GWh) is estimated to be about $30 to *Work done under the auspices of the DOE. $35/kWh. 272 - The LASL and the University of Wiscon- cycle, the converter rectifies ac power sin (UW) are developing SMES systems for to dc for charging the coil. Stored en- electric utility applications.5>6 The ergy is returned to the utility bus for superconducting coils for these systems peak-load demands by operating the con- range in size from small units a few me- verter as an inverter. Commercially ters in diameter and height, which will available thyristors are used as the store as litle as 30 MJ (8.3 kWh), up to switching elements in converters. Fig. 1 large installations several hundred meters shows the SMES system components in block in diameter and height, which will store diagram. as much as 10 GWh. A full-wave Graetz bridge, as shown A technology development program for in Fig. 2, is the fundamental building pulsed superconducting energy storage sys- block of a line-commutated converter. tems for fusion applications has been un- Phase-angle control of the thyristors in derway at LASL since 1968. Both high- the converter determines the dc-output theta-pinch? and low-3 tokamak ohmic- voltage. Such a converter requires reac- heating°~l" systems will need nondissi- tive power from the ac bus during both pative energy storage to achieve over-all modes of converter operation. A reactive power balance. Liners, Z-pinches, lasers, power compensation network, such as a and pulsed electron beam machines are ex- capacitor bank, a synchronous condenser, amples of fusion devices which require or a static, reactive-power controlling large, fast energy delivery systems. The device is needed to provide power factor toroidal Reference Theta-Pinch Reactor correction. (RTPR) would require about 60 GJ delivered in 30 ms, the linear theta-pinch fusion- SMES APPLICATIONS IN ELECTRIC fission hybrid reactor needs about 25 GJ UTILITY SYSTEMS in 2 ms,H and a liner reactor may re- quire about 10 GJ in 1 ms. The ohmic- The types of energy storage systems heating coils in present US designs of for the utilities may be separated ac- tokamak experimental power reactors have cording to the duration of the load vari- about 1-2 GJ of stored energy, and the ation. On a seasonal basis the utilities storage currents must be reversed in 0.5 typically use some form of fuel storage to 2 s to induce plasma current.12-16 to meet the winter or summer peak load. The daily and weekly load variations are Feasibility experiments for Magnetic met by pumped-hydro storage, gas tur- Energy Transfer and Storage (METS) systems bines, and old, inefficient fossil-fired with 1-ms discharge from 300-kJ to 540-kJ power plants. At present, the short-term superconducting coils have been success- load variations are met by adjusting the fully demonstrated for delivery of energy power output of one or more power plants to an adiabatic theta-pinch plasma com- on the system. Each of these short-term pression coil for fusion.1? Pulsed en- load variations is discussed below, and a ergy simulation of both the tokamak plasma SMES unit which might meet their power ohmic-heating and burn cycles has also and energy requirements is described. been demonstrated with a superconducting energy storage coil and a dc-commutated mechanical capacitor.18 LOAD LEVELING Power demands are generally met by a SUPERCONDUCTING MAGNETIC ENERGY combination of three or more types of STORAGE SYSTEM DESCRIPTION power generation, including the base load generation consisting of the more effi- The SMES coil is immersed in a liquid cient fossil fueled or nuclear power helium bath in a dewar, which keeps it plants; the intermediate load generation superconducting at a temperature below 4.5 (midr^ige peaking), consisting of older, K. A closed-cycle refrigeration system smaller, less efficient fossil-fueled cools and liquifies the boiloff helium gas plants anu energy storage units; and the and returns it to the liquid bath. For peak-load generation,'consisting mainly economic reasons, the inductor is a short of gas turbines and energy storage solenoid, a coil with a ratio of height to units. A representative weekly load diameter of about 1/3. A transformer and curve, which is for the Michigan Electric a converter connect it to a 3-phase utili- Coordinated Systems is shown in Fig. ty bus and regulate the power flow. Dur- ing the charge phase of the energy storage 273 Traraformer Smarter Siper conduct ino Three coil pferse bus Reactive power SCR M F ipC nSQi IQfl 1Coil firing circuit protection LHe Wigeratw Controller I Control sicnoHo/from electrical power system Fig. 1. Components of a superconducting magnetic energy storage system. 3-PHASE AC SYSTEM COIL TRANSFORMER GRAETZ- BRIDGE Fig. 2. Schematic of a full-wave, 6-pulse Graetz bridge. 14,000 12.000 §10,000 ~ 8,000 O 6,000 4,000 I2M I2M I2M I2M I2M I2M I2M I2M I2N I2N I2N I2N I2N I2N I2N SUN MON SUN WED THU FRI SAT Fig. 3. Typical weekly load distribution for an electric utility. A SMES unit with the same capacity as the magnetic fields in the vicinity of a the pumped-storage unit in Ludington, MI, SMES unit. It is possible to store large which has a storage capacity of 15 GWh and amounts of energy in a relatively small a power capacity of 2076 MW, would be a volume because superconductors allow high solenoid about 340-m diam and 114 m high. magnetic fields. For example, the Whereas Ludington cost $351 x 106 in Ludington plant occupies about 10 1973 (or $503 x 106 in 1978 based on 7% (km)2. The equivalent SMES unit would per year inflation), the estimated cost of require only 1.5 (km)29 including all an equivalent SMES unit is about $480 x the land area within a shield coil. 106. SYSTEM STABILITY AMD SHORT-TERM LOAD The superconducting coil could be con- VARIATIONS structed several hundred meters under- ground in solid rock, which acts as a The output of power generation structural material to contain Lhe magnet- plants must be adjusted to balance random ic forces on the coil. For a 10-GWh unit and periodic load variations, such as a shield coil with a radius of about four those caused by steel rolling mills, arc times that of the main coil would be con- furnaces, etc. An energy storage unit structed just below ground level to reduce 274 capable of leveling these short-term power Table 1. Design Parameters of a 30-MJ variations would be of great value to a SMES System Stabilizing Unit utility. The SMES units for diurnal load leveling might have a converter capacity Maximum power capcity 10 MW of 1000 to 2000 MW. As the response time Operating frequency 0.35 Hz of the converter to a power demand is less Energy exchange 9.10 MJ than a cycle, it will be possible to meet Maximum stored energy 30 MJ these short-term power demands by varying Coil current at full charge 4.9 kA the power in the converter by a few per- Maximum field at full charge 2.8 T cent. This particular function could also Maximum coil terminal voltage 2.2 kV be satisfied by a small SMES unit that Coil operating temperature 4.5 K stored only 100 to 500 MJ which had a con- Coil lifetime verter capable of delivering only 20 to 50 Heat load at 4.5 K 150 W MW. Inductance 2.5 H Mean coil radius 1.5 m Occasionally, load variations and the Coil height 1.2 m subsequent generation response cause an Winding thickness 0.42 m electrical power system to become unsta- ble. System instabilities can be avoided by limiting the load variation, by chang- Spinning Reserve. The electric power ing the electrical characteristics of utilities are required to have a minimum transmission lines, by reducing the time spinning reserve capacity which amounts response of the generation plants, and/or to about 10% of the load or 1.1 times the by providing system damping. One specific largest generation unit on line. Addi- location where an energy storage device tional converter capcity on a large SMES might improve the stability of a power unit may substitute for the spinning re- system is on the intertie between the serve. During the periods of low-power Pacific Northwest and southern Califor- demand, spinning reserve on the system is nia. Two ac lines and one dc line trans- achieved through the ability of the con- mit power along this corridor. Under verter to change from charge to partial certain conditions an instability arises charge or discharge in less than one on the ac line.20 This instability has cycle. During the times when the unit is been overcome by installing a feedback neither charging nor discharging, the system that controls the converter power SMES unit will be a substitute for spin- at the northern terminal of the dc line, ning reserve. During the longer periods thereby damping the power oscillation on of high-power demand the capacity of the the ac line. This solution is not com- unit will normally only be used at a pletely satisfactory as the power flow on fraction of the maximum converter capa- the ac line depends on the dc line working city. The excess discharge capacity of properly. If the dc line fails, the power the system may then take the place of the flow on the ac line should increase to spinning reserve for the utility system. take up the load, rather than decrease because of reduced stability. A 1-GWh SMES SYSTEM DESIGN A small SMES unit, storing 30 MJ and The LASL is developing a reference having a 10-MW converter, could damp the design for a 1-GWh SMES unit for diurnal oscillations which occur at a frequency of load leveling, One of the major purposes about 0.35 Hz. The fast response of a of a reference design is to provide a SMES unit should improve system stability starting point for detailed engineering and provide for spinning reserve^! (dis- designs. Some of the parameters of this cussed below). Table I shows the major unit are given in Table 2. design parameters for the 30-MJ stabiliz- ing unit. Analysis of the Bonnevi^e ENERGY STORAGE COIL AND SUPPORT STRUCTURE Power Administration system by P. C. Krause at Purdue University has recently The coil is a thin-walled, 132-m- confirmed the earlier work of Cresap and diam, 44-m-high solenoid as shown in Fig. coworkers at BPA.setting the energy stor- 4. The size and shape are the result of age, power rating, and control behavior of a cost optimization and the dimensions such a system. Much of the technology are determined by the maximum field. base for the 30-MJ coil has been estab- lished as a part of the fusion program pulsed inductive energy storage work. 275 Table 2. SMES 1-GWh Reference Design En- allowable stress in the conductor, about ergy Storage Coil Specifications 70 MPa (10,000 psi), then they are trans- mitted through struts to the rock, the Energy stored at full charge coil is placed below the surface of the 3.96 x 1012 J (1.10 GWh) earth where the compressive stresses in Energy stored at end of discharge the rock can be used to offset the mag- 3.36 x 1011 (0.1 GWh) netic loads of the coil. Current at full charge 50 kA Current at end of discharge 15 kA CONDUCTOR Maximum power output or input 250 MW Terminal voltage "to provide Pmax at end Superconductor for a SMES coil must of discharge 16.7 kV be reliable (this includes but is not Inductance 3.17 kH limited to stability considerations), Maximum field at conductor at full charge must cost as little as possible, must be 4.5 T capable of being fabricated with existing Operating temperature 1.85 K techniques or extensions of those tech- Mean coil radius 66 m niques, and must be flexible enough to be Coil height 44 m wound into a coil in a 3-m-wide tunnel. Coil radial thickness 0.30 m Fig. 4. Artist's concept of large SMES unit constructed underground. The magnetic forces must be contained Operation at 1.8 K rather than at 4 by rock to reduce the cost of the system. to 6 K and the use of NbTi rather than If stainless steel bands were used in a Nb3Sn keep the total system cost low. SMES coil, their cost alone would far ex- The use of high purity aluminum instead ceed the cost of other types of storage of copper as the current stabilizer is systems. A set of struts and rods is re- less costly and reduces the size of the quired to transmit the forces from the conductor. A 1-GWh SMES unit is estima- coil at 1.8 K to the rock at about 300 K. ted to cost about $80 million or $80/kWh. The stresses and deflections associated with the thermally induced contraction of To fabricate the conductor with ex- the coil during cooldown and the magnetic isting techniques, a design has been cho- (Lorentz) forces on the conductor are sen in which the NbTi is extruded in cop- taken up by a short length curvature in per and the aluminum is added in subse- the dewar and conductor repeated on about quent fabrication steps. To meet the l-to-2-m centers. Axial loads are al- flexibility criterion, a conductor design lowed to accumulate until they reach the is considered in which several insulated 276 subconductors are in parallel electrically Figure 5 shows one possible circuit and are cabled to reduce hysteretic losses. configuration of a converter. To reduce the reactive power requirement, the con- CONVERTER verter is designed as a series connection of four 12-pulse modules each with its Line-commutated, solid-state convert- own power transformer. Each module is ers are used extensively in high-voltage, designed for 4.5 kV and consists of two dc-power transmission. In comparison, 6-pulse bridges connected in parallel by converters for large SMES systems will an interphase reactor which balances the have medium to high voltage and current current flow in the two bridges. Two ratings. A 1-GWh energy storage capacity modules have a current rating of 50 kA, and a 4-h charging or discharging period and the two remaining modules have a cur- will require a power of 250 MW. Because rent rating of 30 kA and 20 kA. At maxi- of the purely inductive load and the re- mum coil current, each 6-pulse bridge quirement that the maximum power be avail- will provide 25 kA dc. Each 12-pulse able at all operating currents, the maxi- module can be bypassed by a mechanical mum voltage and maximum current do not switch when the module voltage is zero, occur at the same time. Thus, the con- and then the module can be disconnected verter has to be designed for a power from the 3-phase bus. This improves the greater than the maximum power flow ever overall converter efficiency by removing expected through the converter. For the the forward voltage drop of at least four 1-GWh unit the voltage rating is 16.7 kV series-connected thyristors. and the current rating is 50 kA. The installed converter power rating Phase-controlled converters generate and the converter cost can be decreased harmonics and absorb reactive power. by designing those modules which are 3-PHASE BUS 1 AC SWITCH TRANSFORMER 6-PULSE BRIDGE INTERPHASE REACTOR DC SWITCH SUPERCONDUCTING MAGNET Fig. 5. Four series connected 12-pulse converter modules forming a converter. Advanced converter circuits are used to switched out of the circuit first (i.e. minimize these unwanted effects. The har- at a low current) for the current at monic content of the ac-line current is which they are switched off, rather than reduced by using 12-pulse or 24-pulse mod- for maximum current. Theoretically, if ules. Tuned filter networks can remove there were an infinite number of con- the remaining harmonics. The reactive verter modules, the converter could be power requirement can be reduced by sub- designed for a rating dividing the converter into several series + connected modules. During operation, the Pmaxtl 1" (WW phase-delay angles of all but one module SMALL-SCALE SMES SYSTEM TEST are kept at 0°. That one module has a phase-delay angle that depends on the A small scale system, which included voltage requirement. All those converter all the components shown in Fig. 1 except modules at 0° require only a small amount the refrigerator, was tested to evaluate of reactive power caused by commutation. the electrical characteristics of a SMES unit.22 The system contained a 70-H superconducting coil built by stacking 277 eight 3000-turn coils in series. The Reactor (SFTR),23 for adiabatic com- quench current above which the 70-H coil pression of a fusion plasma. A design lost its superconducting property was optimization study2** for the METS sys- 45 A. A 12-pulse, solid-state converter tem'5 led to modular energy storage and a power transformer with a 6-phase coils of approximately 400-kJ size. secondary winding interfaced the coil to These were to be charged in series and the 3-phase laboratory bus. The maximum discharged in parallel for nearly 10055 converter output voltage used in the ex- efficient energy transfer. periment was 150 V. The control system for the model SMES unit was designed with The theta-pinch, METS system is all the features necessary for the auto- characterized by the resonant circuit of matic operation of a large SMES unit on Fig. 7. Coil charging is accomplished the utility bus. with a shunt switch external to the dewar. Discharge is initiated by opening The total system was tested with dif- the shunt followed by opening of the HVDC ferent power demands. The transition time interrupter, B. The interrupter is coun- for rectifier-inverter and inverter- terpulsed to extinguish the arc by a cur- rectifier switching was measured to be 5 rent from the transfer capacitor, Ct, to 6 ms. Figure 6 shows the coil current, which has been back charged. Current coil voltage, and coil power for random then transfers to the compression coil power demand. Because of the fast time with a peak voltage across the circuit response of the converter, the coil power developing at one-half the transfer followed the power demand closely. The period. The energy is then trapped in response to sinusoidal power demands was the compression coil by closing the igni- measured at frequencies up to 30 Hz. tron crowbar, IGcg, for the fusion 10 20 30 40 50 60 70 TIME(s) Fig. 6. Coil voltage, current, and power response to a random power demand signal (L=33H). Mo control system instabilities were ob- burn cycle to be completed during the served during the experiments with a su- 250-ms L/R decay constant of the loop. perconducting coil. The parameters for the METS coils made by INDUCTIVE ENERGY STORAGE FOR FUSION LASL and Westinghouse are listed in Table 3. The LASL coil tested successfully to THETA-PINCH MAGNETIC ENERGY TRANSFER AND 12.5 kA and 386 kj of energy stored. It STORAGE (METST was pulse tested at 10 kA and 35 kV and with transfer times as short as 1 ms in The METS inductive energy storage sys- an L-C-L circuit as shown in Fig. 7. tem was developed to deliver 488 MJ in 0.7 26 ms to a 40-m radius toroidal theta-pinch The Westinghouse coil was system, called the Scyllac Fusion Test operated with pulsed energy transfer at 278 300 M SUPERCONDUCTING STORAGE COIl IGCP 1 AUAIATIC PtASMA SHUNT ' COMPRESSION COIl SWITCH m i COUNTER- PULSE J •OWBISUPPIY Fig. 7. Prototype SFTR-METS circuit. Table 3. Parameters of 300-kJ Energy TOKAMAK OHMIC-HEATING AND BURN CYCLE Storage Coils SIMULATION LAST WH~~ The Westinghouse-METS coil was also Inductance, mH 4.87 6.05 used to demonstrate the use of supercon- Resistance at 20 C, fi 0.0896 0.165 ducting magnetic energy storage in a sim- Stored energy at 10 KA, kJ 244 302 ulated tokamak ohmic-heating and burn Length, cm 73.0 79.1 cycle.18 This performance, although Mean radius, cm 28.7 25.5 much more demanding tha.i that required Winding thickness, cm 0.508 4.74 for the 30-MJ SMES unit intended for the Number of turns 122.5 159.5 BPA system, established the feasibility Number of layers 1 4 for the extension of the METS technology Central field at 10 kA, T 1.82 2.23 for transmission line stabilization units. Conductor support method self grooves Matrix ratio, Cu:NbT1 6:1 2.5:1 Figure 8 gives the circuit used for Wire diameter, mm - 0.813 bipolar operation of the coil in conjunc- Filament diameter, m 32.3 18 tion with a commutated dc generator act- Number of filaments/wire 2640 529 ing as a mechanical capacitor. Figure 9 Wire twist pitch, crrr1 0.13 1.42 shows the oscilloscope trace of the ex- Number of active wires perimental current. The storage coil was in cable 1 72 first charged to -12 kA with a continuous Type of transposition none Roebel duty dc homopolar and then oscillated Cable width, cm 1.016 1.69 through zero current to near +5 kA by Cable thickness, cm 0.508 0.84 connecting it in parallel with the dc generator. The damped half sinusoidal energy transfer corresponds to the energy the 300-kJ level for 50 cycles. This was change expected to occur in plasma heat- at the nominal design level of 10 kA and ing in a tokamak. After this bipolar 2.23-T central field on the winding. The operation, the rectifier bank (power sup- energy transfer period was 2.4 ms. The ply) was connected across the coil to coil was then charged to 13.4 kA and 2.99 T charge it to +12 kA. This corresponds to for a stored energy of 0.54 MJ. This energy the burn phase of the tokamak cycle. The was pulsed from the coil with transfer volt- entire cycle can then be repeated; how- ages near 40 kV. A fast transfer at reduced ever, the test of Fig. 9 was concluded in energy with a 1-ms time safely reached 58 kV. the damped oscillation caused by a dis- The coil has been run in a subsequent test sipative resistance of the coil, dc- to 14.1 kA and a field of 3.1 T with 0.6-MJ generator parallel loop. stored energy. 279 /-LOSS MEASUREMENT / APPARATUS e I »9 RECTIFIER BANK Fig. 8. Circuit for operation of the Westinghouse-METS coil in a tokamak ohmic-heating cycle. Fig. 9. Oscilloscope trace of current versus time of the experimental tokamak ohmic-heating cycle. CONCLUSIONS that of papers submitted to the Institute of Gas Technology and presented to the Superconducting magnetic energy stor- Instrument Society of America (Oct. age units should prove to be effective 16-18, 1978). components of electric power systems. These devices can be used for load level- REFERENCES ing and peak shaving, can satisfy spinning reserve requirements, and can improve sys- 1. IEEE Committee Report, May/June, 1976, tem stability. The fast time response of "Survey of Pumped Storage Projects in the control system will allow a fairly the United States and Canada to 1975," small SMES unit to damp oscillations on IEEE-PAS, Vol. 95, No. 3, pp. 851-858. power systems. 2. Ramakumar, R., 1976, "Survey of Energy Storage Techniques," Energy, IEEE Re- ACKNOWLEDGEMENT gion Six Conference Record, Tucson, Arizona, pp. 105-110. The author wishes to thank W. V. 3. Yao, N. P., Birk, J. R., Aug. 1975, Hassenzahl, W. S. Ranken, R. I. Schermer, "Battery Energy Storage for Utility W. D. Smith, R. D. Turner, P. Thullen, 0. Load Leveling and Electric Vehicles: A D. G. Lindsay, and D. M. Weidon for their Review of Advanced Secondary Batter- work contributing to this paper. The con- ies," Record of the Tenth IECEC, tent of this paper follows quite closely Newark, Delaware, pp. 1107-1119. 280 4. Mattick, W., Haddenhorst, H. G., Reactor Studies," Argonne National Weber, 0., Stys, Z. S., 1975, "Huntorf Laboratory Report ANL/CTR-75-2. the World's First 290 MW Gas Turbine 15. Ballou, J. K., Brown, R. L., Easter, Air Storage Peaking Plant," Proc. of R. B., Lawson, C. G., Stoddart, W. the American Power Conference, Vol. C. T., Yeh, H. T., February 1977, 37, pp. 322-330. "Oak Ridge Tokamak Experimental 5. Los Alamos Scientific Laboratory Su- Power Reactor Study - 1976, Part 3 perconducting Magnetic Energy Storage Magnet Systems," Oak Ridge National Program Progress Reports, Hassenzahl, Laboratory Report ORNL/TM-5574. W. V., editor, LAPR Reports Nos. 5258, 16. Baker, C. C, Project Manager, 5415, 5472, 5588, 5786, 5935, 6004, December 1976, "Experimental Fusion 6117, 6225, 6434, and 7132. Power Reactor Conceptual Design 6. Boom, R. W., Peterson, H. A., et al., Study, Volumes I, II, and III," 1974 and 1976, "Wisconsin Supercon- General Atomic Co., San Diego, CA, ductive Energy Storage Project," Vol. Report GA-A14000, July 1976; and I and II, Engineering Experiment Sta- Electric Power Research Institute, tion, College of Engineering, Univer- Palo Alto, CA, Report EPRI ER-289. sity of Wisconsin, Madison, Wisconsin. 17. Rogers, J. D., et al., "0.54-MJ Su- 7. Ribe, R. L., Krakowski, R. A., perconducting Magnetic Energy Trans- Thomassen, K. I., Coultas, T. A., fer and Storage," Adv. Cryog. Eng. 1974, "Engineering Design Study of a 23, 48 (1978). Reference Theta-Pinch Reactor (RTPR)," 18. Thullen, P., Lindsay, J. D. G., Special Supplement on "Fusion Reactor Vogel, H. F., Weldon, D. M., "Super- Design Problems: to Nuclear Fusion;" conducting Ohmic-Heating Coil Simu- International Atomic Energy Agency, lation," presented at the 1978 App. Vienna. Superconductivity Conf., Pittsburgh, 8. Kulcinski, G. L., Project Director, Pennsylvania, September 25-28, 1978. 1973, "UWMAK-I-A Wisconsin Toroidal 19. Forgey, H. L., 1974, "Feasibility Fusion Reactor Design," UWFDM-68. and System Planning Symposium on the 9. Kulcinski, G. L., Project Director, Ludington Pumped-Storage Hydroelec- 1974, "Major Design Features of the tric Generating Station," Proc. of Conceptual D-T-Tokamak Power Reactor, the American Power Conference, Vol. UWMAK II," IAEA-CN-33/G1-2. 36, pp. 797-806. 10. Arendt, F., Komerek, P., Herppich, G., 20. Cresap, R. L., Mittelstadt, W. A., Knobelach, A., Wermer, F., 1974, March/April, 1976, "Small Signal "Energetic and Economic Constraints on Modulation of the Pacific HVDC the Poloidal Windings in Conceptual Intertie," IEEE-PAS Vol. 95, No. 2, Tokamak Fusion Reactors," Proc. 8th pp. 536-541. Symp. on Fus. Techn., Noordwijkerhout, 21. Peterson, H. A., Mohan, N., Boom, Netherlands, June 17-21, 1974, R. W., July/August, 1975, "Supercon- Instituut Voor Plasmafysica, Jutphass, ducting Energy Storage Inductor Netherlands, EUR5182, p. 563. Units for Power Systems," IEEE-PAS, 11. Krakowski, R. A., Dudziak, D. J., Vol. 94, No. 4, pp. 1337-1348. Oliphant, T. A., Thomassen, K. I., 22. Boenig, H. J., Ranken, W. S., 1977, Bosler, G. E., Ribe, F. L., Dec. 3-4, "Design and Tests of a Control Sys- 1974, "Prospects for Converting tem for Thyristorized Power Supplies 232Th to 233u in a Linear Theta- for Superconducting Coils," Proc. Pinch Hybrid Reactor (LTPHR)," DCTR 7th Symp. Eng. Problems of Fusion Fusion-Fission Energy Systems Review Research, Knoxville, TN, Oct. 25-28, Meeting, ERDA-4; Germantown, MD. 1977, IEEE, Inc., Piscataway, NJ, 12. Roberts, M., Bettis, E. S., eds., IEEE No. 77CH1267-4-NPS, p. 484. November 1975, "Oak Ridge Tokamak Ex- 23. Thomassen, K. I., Editor, January perimental Power Reactor Study- 1976, "Conceptual Design Study of a Reference Design," Oak Ridge National Scyllac Fusion Test Reactor," Los Laboratory Report ORNL/TM-5042. Alamos Scientific Laboratory Report 13. Baker, C. C, Project Manager, July LA-6024; Los Alamos, New Mexico. 1975, "Experimental Power Reactor Con- 24. Rogers, J. D., Baker, B. L., Weldon, ceptual Design Study," General Atomics D. M., 1974, "Parameter Study of Corp. Eng. Staff, GA-A13534. Theta-Pinch Plasma Physics Reactor 14. Stacey, W. M., Project Manager, June Experiment," Proc. of the 5th Symp. 1975, "Tokamak Experimental Power on Engineering Problems of Fusion Research, Princeton, New Jersey, 281 Nov. 5-9, 1973; IEEE Inc., Piscataway, New Jersey, IEEE 73CH0843-3-NPS, p. 432. 25. Rogers, J. D., Williamson, K. D., April 1975, "Proposed METS-FTS Coupled Superconducting Prototype System," Los Alamos Scientific Laboratory Report LA-5918-P; Los Alamos, New Mexico. 26. Mole, J. D., Eckels, P. W., Haller, H. E., Janocko, M. A., Karpathy, S. A., Litz, D. C, Mull an, E., Reichner, P., Sanjana, Z. N., KA Superconducting 0.54-MJ Pulsed Energy Storage Coil," Adv. Cryog. Eng. 23, 57 (1978). 282 PROJECT SUMMARY Project Title: Superconducting Magnetic Energy Storage for Power System Stability Applications Principle Investigator: C. R. Chowaniec and P. H. Stiller Organization: Westinghouse Electric Corporation 700 Braddock Avenue East Pittsburgh, PA 15112 Project Goals: The objective of this investigation was to seek possible applications for small superconducting magnetic energy storage (SMES) devices as aids to maintaining power system stability. Project Status: The suitability of small SMES devices as an aid to main- taining power system stability is discussed. It was confirmed that small SMES units are not effective for keeping an electrical system in synchronism after a tran- sient disturbance because of the limited power and storage rating; however, they do increase the dynamic stability limit of interconnected systems and may be used for sub- synchronous resonance damping, provided power conditioning equipment different from a line-commutated converter is employed. A questionnaire was prepared and sent to about 25 utilities to obtain information about possible SMES unit application sites in utility systems. The remaining work within the project is the evaluation of the questionnaires. Contract Number: LP8-9415Ca Contract Period: July 1978 - Mar. 1979 Funding Level: $24,060 Funding Source: Los Alamos Scientific Laboratory 283 SUPERCONDUCTING MAGNETIC ENERGY STORAGE FOR POWER SYSTEM STABILITY APPLICATIONS C. R. Chowaniec and P. H. Stiller Westinghouse Electric Corporation 700 Braddock Avenue East Pittsburgh, PA 15112 ABSTRACT Superconducting magnetic energy storage (SMES) devices are characterized in terms of reasonable power and total energy ratings as well as other pertinent specifications. The objective of this investigation was to seek possible applications for small SMES devices in power systems, particularly as stability aids. Small SMES devices are those which are not large enough to be practical as bulk energy storage units. The suit- ability of the SMES device as an aid to maintaining powar system stability is discussed in light of the anticipated capabilities of the device. Possible application of the SMES device as a guard against subsynchronous resonance is also considered. INTRODUCTION Huge superconducting magnets have network, must operate in synchronism. been proposed in recent years as an Loss of synchronism occurs when the energy storage device for diurnal load angular difference between the rotors of levelling on electric power systems. The two generators or groups of generators much smaller 30-megajoule magnet; now in exceeds a certain value (nominally 90°). the conceptual design phase at Los Alamos Although the generating unit controls do Laboratories, represents a logical step influence the stability of the system, it in the development of superconducting is the transmission network which produces technology for power system applicat ions. the angular differences between machines. The 30-megajoule superconducting magnet For this reason, the main emphasis in the energy storage (SMES) system is intended study of stability has historically been for use in improving power system sta- on the transmission network. bility. Its characteristics are sum- marized in Table 1. System instability can generally be classified in one of three categories - Prior to the past ten years, sta- steady-state, transient, and dynamic. Of bility concerns for power systems were these three categories, stead/-state in- primarily related to maintaining the stability is the least probably to be generation of a system in synchronism. experienced. Thus, only transient and Acceptable stability performance was dynamic instability were considered in generally achieved without much diffi- attempting to find applications for the culty. As power systems expanded, the 30-MJ SMES unit. Another phenomenon which extensive use of interconnection between can occur on certain systems is undamped systems together with an increasing subharmonic osci1lation. Although it is dependence on firm power flow over these not a stability problem in the strictest lines has renewed the concern of power sense, it can damage equipment nonetheless. system stability. The consequences of Therefore, the potential application of instability were most dramatically ex- SMES units to eliminate subsynchronous hibited by the 1965 Northeast power fail- resonances was also considered. ure. The generators on a power system, which are connected by the transmission TABLE 1. CHARACTERISTICS OF A TYPICAL These oscillations often damp out SMALL SMES UNIT when excited., however it is sometimes possible for the oscillations to be nega- Maximum Power 10 MW tively damped. Negative damping arises Operating Frequency 0.35 Hz primarily from the induction generator Energy Interchange 9.1 MJ (2.53 KWH) effect of the generator at subsynchronous (based on 1/2 cycle) frequencies. Electromechanical inter- Maximum Stored Energy 30 MJ (8.4 KWH) actions will cause pulsating torques which Life 107 Cycles may excite torsional natural frequencies. The result can be severe damage to the Maximum \l^c 2.18 kV Maximum lc|c 5 KA turbine generator. Inductance 2.k H Diameter of Coi1 2.7 M A simple model of subsynchronous Height of Coil 0.86 M resonance is depicted in Figure 1. The subharmonic natural frequency is given by Losses the equation: Conductor 58.7 W Structure 50. W fssr - fc Cost of Prototype $3-5 Million where f is the power frequency, x is the reactance of the series capacitor, and SUBSYNCHRONOUS RESONANCE x) is the total series reactance of the When the synchronous machines of a system. This determines the slip of the power system are coupled together by very generator, s, which is negative since fssr long transmission lines, series capaci- GENERATOR INFINITE SOURCE FiGur:i. SIMPLE MODEL TO ANALYZE SUBSYNCHD. 'U~» RESONANCE 285 Subsynchronous resonance has occured in systems of the type described in Table 1. the West on several occasions. Two rotor shaft failures at the Mohave Plant in TRANSIENT STABILITY 1970 and 1971 were apparently caused by subsynchronous resonance. Transient instability represents the greatest concern, as it can occur on any The following approaches have been system as the result of a major disturb- taken in order to alleviate subsynch- ance such as a transmission line fault or ronous resonance problems: the tripping of a loaded generator. Tran- sient stability analysis is primarily 1) Blocking filters concerned with the effect of transmission 2) Temporary shunting of series line faults on generator synchronization. capaci tors If the remaining transmission is adequate 3) Reduce amortisseur resistance and the speed and angle variations at the time the fault is cleared are limited, the Blocking filters include static, system is likely to settle into a new high-Q. filters with supplemental exciter steady-state equilibrium. Loss of synch- control for additional torsional damping ronism as a result of transient disturb- and dynamic filters, applied either In ances will generally occur in the first series or shunt, which are modulated to few seconds following the initial disturb- resonate with the electrical circuit but ance, however system alternations result- 180° out of phase with the subsynchronous ing from the transient may precipitate sub- resonance. sequent steady-state or dynamic insta- bility which would also result in system By shorting out the series capacitors col lapse. as oscillations begin to grow, the reso- nating circuit is eliminated and the The time period involved in most oscillation will quickly diminish. The transient stability analyses is relatively series capacitors can then be reinserted. short (generally about 1 second or less). The obvious problem is that the series It is primarily concerned with the re- compensation may not be available when sponse of a single unit or plant to a the additional power transfer it affords contingency disruption. Because of the is required. short time period involved, governor re- sponse is usually not important although Reduced amortisseur winding resis- exciter response (especially with high tance reduces the tendency for self- response excitation) may play a signifi- excited osci 1 lat ion by reducing the magni- cant role in maintaining synchronism with tude of the negative rotor resistance, RR. the system. This will, however, reduce the damping under large siips. Figure 2 illustrates somewhat of a worst-case transient stability situation; Returning briefly to the dynamic that is, a generating unit connected to filters, a shunt inductive reactance, the system via a single, long transmission appropriately modulated, will create a line. Before a fault occurs, the system parallel resonant circuit which can actively cancel subsynchronous resonant conditions. The device is essentiaiiy a .PIN static VAR generator with additional con- trols and without capacitors. The induc- tance element and power switching equip- ment is sure to be less expensive than the superconducting magnet and its con- verter. A point to be made here is that a simple 1ine-commutated inverter is INFINITE BUS probably not usable in a SMES system for this application. Furthermore, the fre- quencies involved are much higher than -BREAKERS OPEN AND those considered in the conceptual design RECLOSE AFTER 4t SECONDS of the 30-MJ unit. In summary, it appears that subsynchronous resonant systems do FIGURE 2. not present a very viable application of superconducting magnet energy storage WORST-CASE TRANSIENT STABILITY SITUATION 286 is in equilibrium and the input power, Pin By driving the exciter to ceiling, it is transmitted over the line with a re- is possible to maximize the amount of power sulting phase angle of 6 between the"gen- transmitted over remaining transmission erator and infinite bus.° If the line is lines. One problem with the use of high- suddenly opened, the power generated can respo-ise excitation is its tendency to no longer be transmitted and is used to aggravate dynamic instability problems. accelerate the rotor instead. If the line is reclosed in time, the unit will de- Fast vaiving serves to reroute the flow of celerate and gain a stable operating much of the steam flow around the turbine, point after a few well-damped swings. If thereby eliminating accelerating power, too much time elapses, synchronism is lost. but resets slowly. Both high-speed break- The determination of ultimate system sta- ers and single-pole tripping minimize the bility after major disturbances can be disturbance, the first by re-establishing determined using the "equal-area crite- circuits as quickly as possible while the rion" depicted in Figure 3. other maintains power transfer (unbalanced) over the remaining phases. Braking resis- tors inserted during a fault create an artificial local load, thereby limiting the angular accelerator and increase in LINE POWER vs. PHASE ANGLE rotor angle. Series capacitors inserted into remaining circuits when a circuit is . LINE CLOSED taken out of service by a fault increase the power transfer capability in those lines, again limiting the acceleration of the rotor. The SMES unit, when used as a tran- sient stability aid, would operate in a way similar to the braking resistor. Ex- cess energy resulting from the sudden loss of transmission capacity would be absorbed by the magnet, thereby limiting the angu- lar excursion of the rotor from its steady- state operating point. The SMES unit would have the added advantage of provid- ing active damping of subsequent swings after the fault is cleared. In order to be effective as a transient stability aid, however, the SMES unit must first meet 30 60 90 120 150 180 certain energy storage and power ratings. Both of these ratings will typically be ANGLE - DEGREES well over an order of magnitude greater than those provided in the conceptual FIGURE 3. design of the 30-MJ SMES unit, with in- herent cost penalties. In view of these EQUAL - AREA" CRITERION requirements (particularly power rating) and compared to the alternatives, the application of SMES units for transient stability appears impractical. DYNAMIC STABILITY Several methods of ensuring transient stability have been applied. They include: Some interconnected systems exhibit a tendency toward dynamic instability in 1) Fast Response Excitation which negatively damped power oscillations 2) Fast tfalving can lead to the ultimate breakup of the 3) High-Speed Reclosing Breakers power system. Although there has been no k) Independent Pole Operation of evidence of dynamic instability in the Breakers eastern connections to date, the problem 5) Braking Resistors has been shown to exist in portions of the 6) Series capacitor Insertion West, Midwest, and Canada. 2B7 The time intervals involved in Utilities which contend with this pro- dynamic stability calculations are much blem are characterized by load centers larger than for transient stability. which are widely separated and are supplied Oscillat.on frequencies ranging from a by remote generation with longer angular fraction of a Hertz to a few cycles per displacements relative to local generation, minute are common. In studying the dy- long transmission lines connecting the namic stability of a system, the concern loads and generating stations are a common often relates to oscillation modes involv- feature. The large angular differences ing large groups of machines. The inter- in the generators which are required to action of automatic voltage regulator, transmit power over these long lines re- exciter, and speed governor controls is an sult in reduced line capacity and low nat- essential part of such analyses. ural frequencies which are poorly (or pos- sibly negatively) damped. Slight speed The most commonly employed measure variations which occur continuously under for combatting dynamic instability is the normal operation excite oscillations in power system stabilizer. This device is rotor angle and generator load. If the quite inexpensive and has been proven system damping of these oscillations is effective in eliminating or reducing the negative, the oscillation will increase in occurrence of spontaneous dynamic oscil- magnitude until it is limited by non- lations. It also improves the damping of linearities inherenet in equipment or con- oscillations subsequent to transient dis- tinue to the point that synchronism will turbances. The power system stabilizer be lost if no remedial action is taken. provides a supplemental leading frequency feedback signal to the exciter through The 30 MJ SMES unit described in Table the automatic voltage regulator which 1 is appropriately sized for dynamic oscil- compensates for lags inherent in other lation dampirg applications and would be portions of the control system. Although operated in a manner very similar to dc the power system stabilizer has been suf- transmission where power system stabilizers ficient in overcoming problems of dynamic provide insufficient damping in lieu of ac stability encountered in the past, it is or dc transmission line additions. There not as effective at very low frequencies are some problems, however. First, the of oscillation or for high system angle number of applications where a 30 MJ unit operation (heavy tieline loading). could be justified to increase dynamic stability limits will be quite limited (at Another approach is to simply reduce least at the present time). Second, the the gain of the automatic voltage regula- amount of energy passed through the device tor, however this tends to sacrifice the over a given time period will be relative- performance of the particular unit under ly small. When this is considered along transient conditions. with the relatively small inductance (2.k Henries), one must seriously question the The advantages of dc transmission need for a superconducting coil for this lines in improving ac system stability application. have been recognized and these capabili- ties have been incorporated into the con- CONCLUSIONS trols of existing dc lines. There, they have demonstrated the ability of the dc Superconducting magnet energy storage line to damp out oscillations by properly is not likely as an economic solution to modulating the power flow over the dc line power system stability problems except in and to improve transient stability by very special cases. The 30 MJ SMES units rapidly changing the dc line loading. To considered, when applied for power system date, however, the installation of dc stability, will give important operating lines has been justified on the basis of experience for superconducting bulk energy considerations other than improved sta- storage. bility characteristics. But the factors which contribute to the economic feasi- The primary applications of supercon- bility of dc lines (particularly long dis- ducting magnet energy storage will proba- tances) are often the same factors which bly be in the area of energy storage; cause a dynamic stability problem. The therefore, applications should be sought "fringe benefit" of improved stability for early demonstration of the technology might arise naturally with the expansion in this area. When SMES units are used of systems in which SMES units might be for the purpose of bulk energy storage employed for system stability applications. the stability improvements (transient and 288 dynamic) discussed in this paper will be fi. R. H. Millan, J. A. Mendoza, an additional asset realized by the SMES C. Cardoza, and A. de Lima, "Dynamic unit. Stability and Power System Stabiliz- ers," IEEE Trans. Power Apparatus ACKNOWLEDGEMENTS and Systems, Vol. PAS-96, pp. 855- 862, May/June 1977. The authors wish to acknowledge the support of the Los Alamos Scientific 9. P. Bingen, G. L. Landgren, F. W. Keay, Laboratory which provided funding for this and C. Raczkowski, "Dynamic Stability project. Tests on a 733 MVA Generator at Kincaid Station," IEEE Trans. Power REFERENCES Apparatus and Systems, Vol. PAS-93, pp. 1328-1331*, Sept./Oct. 197**. 1. H. A. Peterson, N. Mohan, and R. W. Boom, "Superconductive Energy 10. H. M. Ellis, J. E. Hardy, and Storage Inductor-Converter Units for A. L. Blythe, "Dynamic Stability of Power Systems," IEEE Trans. Power the Peace River Transmission System," Apparatus and Systems, Vol. PAS-9A, IEEE Trans. Power Apparatus and Sys- pp. 1337-13^8, July/August 1975- tems, Vol. PAS-85, pp. 586-600, June 1966. 2. R. L. Cresap, D. N. Scott7 W. A. Mittelstadt, and C. W. Taylor, "Opera- 11. W. H. Croft and R. H. Hartley, "Im- ting Experience With Modulation of the proving Transient Stability by Use of Pacific HVDC Intertie," IEEE Trans. Dynamic Braking," IEEE Trans. Power Power Apparatus and Systems, Vol. PAS- Apparatus and Systems. Vol. 81, pp. 97, pp. 1053-1059, July/August 1978. 17-26, April 1962. 3. W. A. Mittelstadt, "Four Methods of 12. M. L. Shelton, P. F. V.'inkelman, Power System Damping," IEEE Trans. W. A. Mittelstad, and W. J. Bellerby, Power Apparatus and Systems, Vol. PAS- "Bonnevllle Power Administration 87, pp. 1323-1329, May 1968. I'lOO MW Braking Resistor,1' IEEE Trans. Power Apparatus and Systems, Vol. PAS- 4. E. W. Kimbark, "Improvement of Power 9t, pp. 602-611, March/April 1975. System Stability by Changes in the Network," IEEE Trans. Power Apparatus 13. R. T. Byerly and E. W. Kimbark, and Systems, Vol. PAS-88, pp. 773-781, Stability of Large Electric Power May 1969. Systems, IEEE Press, New York, N.Y., 5. L. A. Kilgore, L. C. Elliott, and E. R. Taylor, "The Prediction and Con- trol of Self-Excited Oscillations Due to Series Capacitors in Power Systems," IEEE Trans. Power Apparatus and Sys- tems, Vol. PAS-90, pp. 1305-1313, May/ June 1971 6. E. W. Kimbark, "How to Improve System Stability Without Risking Subsynchro- nous Resonance," IEEE Trans. Power Apparatus and Systems, Vol. PAS-9~6~, pp. 1608-1619, September 1977. 7. L. A. Kilgore, D. G. Ramey, and W. H. South, "Dynamic Filter and'Ot^r Solutions to the Subsynchronous Reso- nance Problem," Proceedings of the Amer. Power Conf., Vol. 37, PP 923- 929, (1975): 289 SESSION VI: UNDERGROUND PUMPED HYDROELECTRIC STORAGE 291 PROJECT SUMMARY Project Title: Underground Pumped Hydro Storage Program Principle Investigator: Dr. G. T. Kartsounes Organization: Argonne National Laboratory 9700 South Cass Avenue T-12-3 Argonne, IL 60439 Project Goals: The primary objective of the Underground Pumped Hydro Storage Program is to enable the electric utility industry to supply power at the lowest possible cost, while elimi- nating the use of premium fossil, fuels for peak-power generation. Successful commercialization of this energy storage scheme requires research and development studies on turbomachinery, lower reservoir contruction, geology, motor/generators, and system arrangement and optimization. Project Status: This program was initiated during FY 1978. A contract was negotiated with Ailis-Chalmers Corporation, Hydro-Turbine Division, for preliminary design studies on single-and double stage reversible pump/turbines with wicket gates. These studies were completed and a report is in preparation. Contract Number: W-31-109-ENG-38 Contract Period: FY 1978 Funding Level: $200,000 B0 Funding Source: Department of Energy, Office of Conservation, Division of Energy Storage Systems 293 UNDERGROUND PUMPED HYDRO STORAGE - AN OVERVIEW S.W. Tain, C.A, Blomquist, G.T. Kartsounes Energy and Environmental Systems Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439 ABSTRACT This paper reviews the status of Underground Pumped Hydro Storage (UPHS) for elec- tric utility peaking and energy storage applications. The salient features of major re- cent studies are reviewed. Turbomachinery options and advances in high-head pump/turbines are discussed. The effect of head, capacity, turbomachinery unit size and type, and other performance variables on the cost of a UPHS plant are presented. Market potential, siting criteria, lower reservoir construction, and geological related issues are addressed. The environmental impact of a UPHS plant is reduced from comparable facilities, and these issues and other safety concerns are presented. UPHS is an economically viable scheme for energy storage and peaking applications, where considerable savings in premium fuels are achievable by the replacement of combustion gas turbines. The technology for UPHS is available, but additional research and development is required for high-head turbomach- inery, motor/generators, cavern geology, and system optimization. INTRODUCTION THE NEED FOR ENERGY STORAGE Electric utilities face large varia- community, or utility-size), there is a tions in the daily electric power demand. wide range of viable energy storage tech- During periods of relatively staady demand, nologies available.1* This storage is some- power is generally provided by nuclear or what different from the peaking application coal-fdred base-load plants, which are eco- discussed previously. nomical to operate. For peak load periods, gas turbine driven generators are used. UNDERGROUND PUMPED HYDRO STORAGE They are fast responding, easy to maintain and operate, but they have relatively low One energy storage scheme that has efficiencies and burn premium fuel (oil or long been utilized for peak power genera- natural gas). These undesirable character- tion by the electric utilities is surface istics provide the incentive to seek alter- pumped hydro storage (SPHS). In this nate energy technologies for peaking appli- scheme (Fig. 1) idle base plant capacity cation. is utilized to pump water from a lower reservoir to an upper reservoir during off- During periods of low demand (usually peak hours. During the peak demand period, at night), bar.,e-load plants operate at re- power is fed into the power grid by letting duced capacity. If they are operated at the water run down to the lower reservoir. full-capacity during these periods, the In this way, the stored potential energy extra energy generated could be stored pro- of the water is converted into kinetic vided viable energy storage schemes are energy which is utilized to run a hydro- available. The stored energy would then turbine coupled to a generator. be available to meet (at least partially) the peak demands, and thus, reduce the re- SPHS has been in the U.S. since 1953 liance on premium-fuel fired units. when the U.S. Bureau of Reclamation in- stalled a 9 MW unit at Flatiron.9 Today Solar energy (direct and indirect) many systems are operational. However, a requires storage because its availability SPHS system requires a suitable site for usually cannot be assured. Depending on the upper-level reservoir located at suf- the scale of power generation (residential, ficient elevation in the vicinity of the 294 lower reservoir and also at a reasonable through the penstock shaft, and the lower distance from the load center concerned. cavern is vented to the atmosphere through Sites with such characteristics are being the vent shaft. rapidly depleted today within the U.S. The concept of underground pumped hydro storage (UPHS) is designed to minimize UPPER RESERVOIR these siting and transmission difficulties while still fulfilling the primary mission GROUND LEVEL of generating peak power for the electric MNTAKE utility grid. ACCESS, VENTILATION - AND CABLE SHAFT CONVENTIONAL PUMPED HYDRO ^ PENSTOCK SHAFT - INTAKE VENTILATION'' SURGE TANK SHAFT LOWER RESERVOIR FOWERSTATION- Fig. 2. UPHS Single-Drop Scheme UNOERGROUND PUMPED HYDRO Fig. 1. Pumped Storage Schemes SERVICE BUILDING Conceptually, UPHS (Fig. 1) is very GROUND LEVEL similar to SPHS, and, similarly, does not require premium fuel. The major difference PENSTOCK - -ACCESS, VENTILATION from SPHS is that the upper reservoir for SHAFT AND CABLE SHAFT UPHS is at ground-level while the lower -VENTILATION SHAFT reservoir (an excavated cavern in general) POWER STATION ~ is located underground. This important -INTERMEDIATE feature results in the advantages of siting RESERVOIR flexibility and reduced transmission costs and retains the good system reliability and availability characteristics of SPHS. A typical proposed UPHS plant will have a capacity of 1000-2000 MW with 8-10 hours of storage, and an overall efficiency close to 80%. Presently, there are two basic plant configurations — the one-drop and tha two-drop schemes (Figs. 2 and 3). In the one-drop scheme, there is only one underground reservoir. The turbomach- Fig. 3. UPHS Two-Drop Scheme inery is housed in the powerhouse located below the underground reservoir to provide pump submergence. Access, cable and equip- The two-drop configuration utilizes ment shafts connect the powerhouse to the single-stage, reversible pump/turbines aboveground service building. The power- (RPT) with an average gross head of about house is linked to the upper reservoir half the depth at which the lowest reser- 295 voir is situated. A smaller intermediate and Harza Engineering Co.6 have conducted reservoir is located half way between studies of UPHS. Salient features of these ground level and the underground cavern. studies are summarized in Table 1, and A powerhouse is connected to each of the several points are worth noting. two underground reservoirs. The interme- diate cavern allows the two power plants (1) There is a considerable spread in the to operate in series without synchronizing range of head under consideration the turbomachinery. The shaft requirements (from 765 m to 1500 m). Generally, are similar to the one-diop scheme. the more recent studies tend to con- sider higher head. UNDERGROUND RESERVOIR (2) The estimated plant cost varies from For both configurations, the lower 273 to 405 $/kW. These costs have reservoir is usually constructed in a room- been updated to September 1978 dollars and-pillar layout with an intersecting grid by using a general inflation rate of of tunnels (Figs. 2 and 3). In the Chas. 8% per annum. In a more precise esti- T. Main report,2 which is the latest study mate, the inflation rate will be dif- on the subject, a typical UPHS lower reser- ferent for the various components of voir will cover an area of about 64 x 101* the plant. The more recent studies sq. m. ("VL/4 sq. mi.). The size of each tend to give higher cost estimates. tunnel is about 15-m wide by 25-m high with The higher costs are due to different a crown of 7.5 m radius. The pillars sep- system configurations and inflation- arating the tunnels are 45 m by 45 m. Such related costs. room-and-pillar construction is expected to give a stable configuration at depths (3) The turbomachinery considered is (>1000 m) considered in UPHS schemes. either conventional or slight exten- sions of present technology. Single- The three primary methods of shaft stage RPT units for heads up to about construction are conventional sinking 765 in are considered. For higher methods, machine boring, and raising by heads, either a two drop scheme is means of an elevating platform, followed, used or multistage RPT units or tan- if necessary, by downward enlarging. The dem units are employed. The state-of- time required for construction of a UPHS the-art of the multistage and tandem plant is 6-8 years. Total time in excess units is adequate for UPHS application, of 10 years may result if planning and but they are inherently more expensive. lead-time are considered. Economical operation requires some im- portant extensions to turbomachinery UPPER RESERVOIR technology. The upper reservoir for a UPHS plant can be a specially-constructed dedicated GEOLOGY-RELATED ISSUES reservoir, or natural lake, or one created by damming a river or stream. Because of The important factors governing UPHS high-head plant operation, the required siting opportunities are site geology, hy- upper reservoir volume will be significant- drology and the location of major load cen- ly less than a conventional pumped storage ters.'2'11 In general, the underground reservoir. Containment structures for the construction, particularly for the cavern, reservoir will be essentially the same as requires a medium with good competent rock those used for conventional pumped storage. with few discontinuities. Igneous intru- There does not appear to be any identified sive and crystalline metamorphic rock problems with reservoir construction other structure will in general satisfy UPHS con- than possible environmental aspects, dis- struction requirements. Favorable areas cussed later. A unique feature of the UPKS within the continental U.S. for UPHS de- concept is the availability of a large vol- ployment are shown in Fig. 4. Of course, ume of high-quality rock from the under- within the favorable areas, there may exist ground reservoir excavation that can be localized regions with unfavorable geology used for containment construction. and vice versa. Since transmission cost is about $0.34/kW/mi,2 incentives exist to MAJOR STUDIES keep plants relatively close to load cen- ters. It is worthwhile to note that within Within the last few years, Chas T. the favorable geological areas such as the Main, Inc.,2 PSEG,wAcres American, InC.,1'13 Midwest and the Northeast lie some of the 296 Table 1. Major UPHS Studies (adapted from Ref. 2) PREFERRED TYPE ESTIMATED PREFERRED SCHEME PRINCIPAL PURPOSE TITLE OF EQUIPMENT COST <$/kW) Underground Pumped Storage Indirectly suggested one 200 MW units: at 700 m 300 To identify particular as- Research Priorities (April or two drops, approxi- (2300 ft) single-stage for 1000 MW pects requiring detailed 1976) Prepared by Acres mately 900 m (3000 ft) runner reversible and at 900 m examination during a sub- American Incorporated head at 1000 m (3300 ft) (3000 ft) sequent comprehensive pre- multi-stage runner, and 10-yr llninary design phase reversible storage An Assessment of Energy Considered: Single-stage rever- <544 To provide the required data Storage Systems Suitable - Up to 765 m (2500 ft) sible units of about for 10-hr to establish research and for Use by Electric Utili- for single-stage re- 300 MW storage development priorities for ties (July 1976) Prepared versible units. energy storage technology by Public Service Electric - About 1070 m (3500 ft) and Gas Company, Newark, for two plants in series New Jersey Siting Opportunities in In economic computations Approxi- To categorize the geology the U.S. for Compressed assumed 200 MW plant, mately 273 of the continental United Air and Underground 10-hr storage States in accordance with Pumped Hydro Energy 1070 m (3500 ft) head the potential for the siting Storage Facilities of compressed air energy (December 1976) Pre- storage (CAES) and UPHS red by Acres American Incorporated Underground Pumped Two-drop scheme 250 MW single-stage 200 to 366To Identify and describe Hydro Storage and Com- Total head 1340 m (4400 reversible units at (2000 MW) regional markets for UPHS pressed Air Energy ft) 2000 MW, 6 to 10-hr 670 m (2200 ft) head 10-hr stor- aod CAES and perform geo- Storage. An Analysis storage age logic analysis to determine of Regional Markets regional development poten- and Development Poten- tial tial (March 1977) Prepared by Harza Engineering Company Underground Hydro- One-drop scheme 333 MW multi-stage 348 to 405 To Investigate state-of-the- electric Pumped Two-drop scheme One-drop: a. Multi- art of UPHS, to evaluate Storage. An Evalua- Total head -1200 i stage RPT; b. Tandem feasibility and relative tion of the Concept 10-hr storage unit economic viability and to (June 1978) Prepared Two-drop: two single- Identify needs for further >y Chas. T. Main Inc. stage RPT in series technological and economic research major load centers within the continental spacing of joints and their orienta- U.S. These areas are, therefore, prime tion and in situ stress state (verti- regions for UPHS applications. cal as well as horizontal stresses), One important issue is the stability (c) specific details of the water/pumping/ of the underground structure. This per- discharge cycle, and, finally, tains to the response and long-term beha- vior of the underground cavern when it is (d) specific configuration of the under- subjected to the cycling effect of charging ground reservoir. and discharging of water over a long period of time. During normal operation of a UPHS With these as inputs, the elastic/plastic plant, the underground structures are ex- response of the underground structure may posed alternately to water and air. The be analyzed. For site-specific stability effect of changing pressure, temperature analysis, of course, rock properties per- and erosion on the stability of the lower taining to that particular? site will be cavern is an issue that has to be investi- needed. gated. Necessary inputs for such an anal- ysis will be: Another important issue is the explor- atory technique used. Since rock mass in- (a) the thermo/physical/mechanical rock formation is a prerequisite for any stabi- properties of the region in question, lity analysis, the means of collecting such information is an important issue. However, (b) the state of the rock mass such as up to now the principal exploratory tech- 297 AREAS GENERALLY FAVORABLE FOR UPHS TAVERNS. Fig. 4. U.S. Sites Generally Favorable for UPHS Caverns nique still involves direct drilling of vary with the individual site, all loca- boreholes. Because rock mass properties tions have certain generic factors such as can vary substantially over the depth range disposal of excavated material, water con- involved in UPHS project (>1000 m), bore- tamination, and disruption of the natural holes have to be drilled to comparable habitat. Consideration must be given to depth in order that reliable information is the socio-economic impact of the area during obtained. For example, the in situ stress plant construction. ratio (horizontal to vertical stress) may vary significantly over the depth range It has been estimated that from 5.5 considered in UPHS schemes. In addition, to 7.5 million cubic meters of material all methods are accurate only over a short will be excavated for construction of the range from the borehole. However, this lower reservoir and powerhouse. Part of kind of approach, while expensive, seems to this material can be utilized for construc- have no substitute at the moment. Hence, a tion of the upper reservoir containment; UPHS project carries an element of risk the remainder requires either on- or off- that is characteristic of projects that re- site disposal. This is excellent quality quire geo-exploration to provide input in- construction material and its usage could formation. This risk can be reduced by be profitable. Off-site utilization or more extensive on-site geo-investigation disposal can produce air pollution or traf- within reasonable economic limits. fic congestion. Disposal by dumping can alter surface drainage, create a visual Other issues such as questions of hy- blight and limit land usage. drological and environmental concerns can be dealt with through judicious site selec- The contamination of reservoir surface tion. and subsurface water is a factor that must be considered. Construction of the lower ENVIRONMENTAL, SOCIO-ECONOMIC, reservoir will alter ground water charac- AND SAFETY ASPECTS OF UPHS teristics and local aquifers could be pene- trated, diverted or drained. Groundwater The selection of a UPHS plant will contamination may result from seepage around generally be determined by the availability access shafts or penstock leaks. Surface of competent geologic conditions, the prox- water drainage and quality will be affected imity of load centers and transmission by roads, buildings and the upper reservoir. lines, and a sufficient water supply. Cycling of water between the reservoirs Closely tied to these variables are envi- could result in mineralization, solids sus- ronmental considerations. Whereas the en- pension, and an increase in water tempera- vironmental effects of a UPHS plant will ture. Mineralization should be minimal due 298 to the high quality of rock required for turbines, and coal-fired cycling plants, cavern construction. Solids suspension are not as attractive environmentally as should approach the level of normal reser- UPHS. Conventional pumped storage requires voirs after initial operation and sedimen- additional land area for the lower reser- tation have occurred. voir and only limited sites are existent. Combustion turbines and coal-fired plants Water heating from the hotter under- introduce air pollutants and require stor- ground rock does not appear to be signifi- age and transportation of fossil fuels. cant, but additional investigations are Common to all storage schemes is the addi- required. If the surface reservoir is a tional pollution resulting from increased natural lake, then water level fluctuations, operation of the base-load plant. mineralization, temperature, groundwater seepage, oxygen depletion, and algae for- TURBOMACHINERY FOR UPHS PLANTS mation effects on the aquatic life in the lake must be evaluated. Water discharge As discussed previously, the two from the lake will in turn impact the re- schemes projected for UPHS are the single- gion downstream of the lake. A manmade drop and the two-drop systems. A schematic reservoir will reduce these effects but representation of the turbomachinery and will introduce concerns during the diver- present head limitations that will be used sion of water for the filling and replen- for these systems are shown in Fig. 5. ishment of the reservoir. Basically, the equipment is either a rever- sible pump/turbine (RPT) or a tandem unit It is assumed that site selection will consisting of an impulse turbine, usually exclude areas of historic or archaeologic a Pelton-type waterwheel, and a multistage significance, rare or endangered plants or pump. Single-stage and multistage units wildlife, planned parks, and valuable min- are available and two-stage turbine designs eral deposits. The disruption of the are being considered. natural habitat requires consideration of foliage elimination and displacement of HEAD TYPE wildlife. The addition or alteration of LIMIT terrain features could produce a visual SINGLE-STAGE impact. REVERSIBLE GATED Socio-economic impacts can be both TWO-STAGE beneficial and detrimental to the surroun- REVERSIBLE ding region. Some beneficial effects in- GATED clude increases in the tax base, employment and business activities. Detrimental ef- fects include traffic congestion and in- MULTI-STAGE REVERSIBLE creased demands on service such as police, NON-GATED schools, and medical. These socio-econo- mic factors are not unique to a UPHS plant but are generally similar to any large con- struction project. SEPARATE-UNITS Safety issues are similar for any 1400 m IMPULSE FOR plant construction, but a UPHS plant has TURBINE MODE the additional problems associated with underground mining and operation. Flooding MULTI-STAGE NON-GATED of the powerhouse is a concern because of PUMP FOR the serious equipment damage and danger to PUMP MODE operating personnel that could result. The possibility of having the underground sys- tem subjected to the full hydrostatic head could also be a problem. Fig. 5. Turbomachinery for UPHS Plants • Careful site selection, planning, ap- propriate design, and the use of a dedica- This single-stage RPT has been util- ted man-made upper reservoir will eliminate ized extensively for pumped storage pro- or mitigate the environmental impact of a jects since about 1954. These units pro- UPHS plant. Alternatives to UPHS such as vide the least complicated arrangement and conventional pumped storage, combustion are generally equipped with adjustable 299 wicket gates which provide regulation of gates and runner buckets, all of which power output and good control for starting could reduce turbine efficiency. To coun- and synchronizing the unit in turbine teract this potential efficiency reduction operation. In pumping operation, adjus- requires careful mechanical and hydraulic table wicket gates are relatively ineffec- design of the runner to operate at higher tive for regulating discharge but enable specific speeds than customarily used to- variable-head operation at optimum effi- day. To increase the specific speed, either ciency. One disadvantage of the single- the turbine speed or capacity must be in- stage turbine is its requirement of deeper creased. However, turbine speed is some- submergence to avoid cavitation; another what restricted by the available synchronous limitation is the head range. The maximum speeds of the motor/generator and manufac- design head for these machines has bean turers of this equipment are reluctant to steadily increasing. In the early sixties, go to very high speeds. Efficiency losses operating heads were less than 250 m. In associated with high-volume flow rates will the next ten-year period, the maximum head also limit the specific speed. In addition, reached 390 m (Robiei-Switzerland). Since some unidentified technological barriers that time, machinery has been ordered and may exist which will limit the achievable commercial operation started for a number maximum head. Resolving these concerns re- of plants with turbine heads in the 500-m quires conducting research studies and range. The current maximum head is 620 tn model tests. This activity is part of the for the Bajina Basta Power Company in ongoing Department of Energy-UPHS program. Yugoslavia. Therefore, utilization of The Hydro-Turbine Division of Allis- conventional single-stage RPT units for a Chalmers Corp., with a contract from Argonne 1000-1200-m UPHS plant necessitates a two- drop system. National Laboratory (ANL), has completed a preliminary design study of 500 MW, single- stage RPT units for operation at heads of Continual progress toward higher-head, 500, 750, and 1000 m. A contract report is single-stage RPT units is being made, but presently being written. Swiecicki estimates that it may take more than another decade before, if ever, a Multistage reversible pump/turbines machine is ordered and designed for a 900- without wicket gates are available for the 1000-m head. A conservative approach to head range proposed for UPHS. Some units head increases is attributed to the reluc- currently operational or under considera- tance of manufacturers to advance the tion are: state-of-the-art too rapidly and because utilities are not willing to accept mach- - Chictas-Italy, 1070 m, 4 stages, 150 MW ines which deviate significantly from es- - Edolo-Italy, 1290 m, 4 stages, 142 MW tablished practice. In addition, there - LaCache-France, 930 m, 5 stages, 79 MW. has not been a market for high-head tur- bines that would encourage manufacturers The multistage unit has the advantage to develop these machines. In the United of a developed pump cycle and requires less States, there are very few conventional submergence than a single-stage machine. pumped storage sites which could provide Its specific speed per stage can be higher heads in excess of 550 m. In Europe, high- which increases efficiency, but this is head pumped storage plants have used either offset by additional losses between stages. multistage RPT or tandem units. The prac- Since these machines do not have wicket tical upper limit on the design-head for a gates and therefore, cannot be regulated, single-stage RPT is estimated as 800-1200 m. they always generate a block-level of power. Being able to achieve the higher limit is Therefore, the system loading must be capa- mandatory for the single-drop UPHS scheme. ble of absorbing this ungoverned gneration without creating stability problems. In the pumping mode, these machines are all High-head design poses problems with started in the watered condition because rigidity of turbine parts, fatigue stress of technological problems in dewatering limits, seals, and the manufacturing com- and priming. The main obstacle in the de- ponents with narrow passages. Attention sign of multistage units appears to be the must be paid to surface profiles and fini- structural and metallurgical design of the shes to avoid cavitation and head losses. casing which tends to limit the size of the Serious stress and vibration problems of unit. the wicket gates and their operating mech- anism will occur with increasing head. Tandem units consisting of a Pelton- Higher heads also require longer and type waterwheel and a multistage pump thicker stay vanes, and thicker wicket 300 coupled to the same shaft are utilized in monly used in a number of European pumped Europe for conventional, high-head pumped storage plants. This scheme does not storage. Such units are: appear to be economically attractive and would require the development of a high- - San Fiorino-Italy, ]440 m, 140 MW head reaction turbine. Hartmann and Meier7 - Rottau-Austria, 1100 m, 200 MW. presented an arrangement called the "One + One." This idea basically combines two The Pelton turbine must operate in single-stage RPT units in series to form free air and is normally situated above a two-stage machine. Several advantages the highest tailwater. The pump has to be of this arrangement were cited, but addi- below the lowest tailwater level for cavi- tional study is required to assess the tation-free operation. For vertical shaft feasibility of this scheme. Consideration units, the shaft length is long, often in also needs to be given to the controlled- excess of 30 m, and requires considerable flow rate pump/turbine of Gokhman,5 which support bearings. The powerhouse cavern appears to be applicable for solar energy tends to be high and narrow which causes storage. This concept utilizes an adjus- stress problems in the walls. To overcome table hub and a supplementary adjustable these disadvantages, a booster pump can be cover instead of wicket gates to obtain a used to provide the necessary inlet pres- relatively flat efficiency curve over a sure to the main pump. This technique re- variable flow range. duces the vertical shaft length and is utilized for horizontal shaft units. UPHS ECONOMICS Operating the Pelton unit under an air back-pressure allows it to be located be- low the tailwater level. This technique Typical cost estimates for several has been sucessfully used at the Tysso IX UPHS plant configurations are listed in station in Norway. Tandem units have high Table 2. Included in the item designated efficiency over a wide head range, simpli- as reservoirs, dams, and waterways are the city of synchronous-condenser operation, upper and lower reservoirs and water con- ease of starting as a pump, rapid change- ductors with all the associated shafts ex- over from the pulping mode to the turbine cept those connected to the upper service mode, and vice versa. Its disadvantages building. Tlu lower reservoir cost in- include high cost and limited capacity of cludes excavation, disposal, rock-bolting the multi-stage pump section. and shotcrete. It comprises about 25-30% of the total plant cost, and is therefore the most important cost component. Items A two-stage RPT with wicket gates is such as power plant structures, pump/tur- being considered for UPHS application even bines, generator/motors, and miscellaneous though none of these units are operational equipment are each about 8-10% of the total or on order today. Escher-Wyss has com- cost. pleted model tests and performed mechanical design investigations for such a unit. The The approximate costs for the turbo- two-stage RPT with wicket gates is more machinery proposed for UPHS plants is shown flexible than a non-gated design, and re- in Fig. 6. These costs were obtained from quires less submergence than a single-stage Mitchell8 and include turbine governors and RFT for the same total head. It is capable penstock valves. The cost of a two-stage of being used for higher heads than a RPT with wicket gates will fall between the single-stage turbine, but requires initial single-stage and multistage cost. The development of a high-head single-stage single-stage RPT is the lowest cost unit unit. One disadvantage is the loss of ef- which provides the incentive for the de- ficiency between stages. Hartmann and velopment of this type of high-head turbo- Meier7 point out that considerable effort machinery. In addition, multistage and will be necessary in hydraulic research and tandem units are size-limited and will re- mechanical design to bring this type of quire more units for a given plant capacity machine to a stage of development where the than the single-stage RPT. Additional construction of a full-siae prototype will units require more waterways and gates be feasible. Allis-Chalmers has recently which increases plant cost. completed a study for ANL on the preliminary design apects of such a machine. A report is presently being written on this study. The effect of head on the cost of a 500 MW single-stage RPT is shown in Fig. 7. These data are obtained from the recent This discussion has excluded separate Allis-Chalmers study3 and show decreasing pump and turbine units that have been com- turbine costs with increasing head. This 301 trend is contrary to the multistage RPT cost data in the Chas. T. Main study2 which predicts increasing machinery cost with increasing head. 18 Table 3. Cost Estimate Summary on Several UPHS Systems ' 16 Item/System I II III Land and Land Rights 2 2 2 §14- Power Plant Structures and Improvements 71 89 75 SINGLE-STAGE RPT 12 - Reservoirs, Dams, and NOTE: SEPT, 1978 COSTS Waterways 293 320 297 Pump/Turbines 76 44 117 10 250 500 750 1000 1250 1500 Generators/Motors 52 49 54 HEAD, M 51 58 52 Miscellaneous Equipment Fig. 7. Effect of Head on Single- and Roads and Bridges 6 6 6 Two-Stage RPT Cost3 Contingencies, Engineer- ing, Supervision and Recent studies1'2> s> 8> 10 have shown Overhead 183 194 197 that with other factors (such as energy storage capacity and types of turbomach- 734 762 800 TOTAL inery equipment) being held constant, the $/kW 367 381 400 required underground reservoir volume de- creases as the net head increases. There- *1) Adapted from Ref. 3. I: Multistage fore, strong incentive exists to go to RPT; II: Single-Stage RPT (two-drop); higher heads to reduce reservoir cost. III: Tandem Unit. All three plants are However, higher heads (>1000 m) result in 2000 MW, 1200 m head with 10 hrs storage. 6 a cost penalty from increased construction 2) Units in 10 September 1978 dollars. time. An optimum head should be reached 3) $/kW in September 1978 dollars. where any further increase in head will only marginally reduce the overall system cost. The general trend of system cost 200 1 1 (in $/kW) versus net head is illustrated i i r in Fig. 8. Estimated costs are direct AND PUMP (TANDEM) costs and do not contain the effects of interest and escalation during construction; 100 inclusion of these effects can produce po- tentially significant changes to economic conclusions. * 50 RESEARCH AND DEVELOPMENT REQUIREMENTS £ 40 Even though the technology for UPHS g 30 is somewhat advanced, additional research o and development studies are required to 20 improve the economics, and eliminate any NOTE: technological uncertainties that would hin- NOV. 1975 COST INCLUDING der large-scale commercialization. Research GOVERNOR AND VALVES should be conducted to improve equipment 10 I I I 1 1 and construction techniques for excavating 30 40 50 100 200 300 500 large-diameter deep shafts and large under- UNIT SIZE, MW ground caverns. Geological studies are needed to define, evaluate and determine Fig. 6. Capacity Effect of Cost of Turbo- suitable lower reservoir design and sta- machinery for Single-Drop Schemes6 bility criteria with emphasis on stress 302 tally, a UPHS plant is more attractive than alternatives such as conventional pumped storage, combustion turbines, and 500 coal-fired cycling plants. Careful site selection, planning, appropriate design, and the use of a dedicated man-made upper 400 reservoir will eliminate or mitigate the environmental aspects of a UPHS plant. Recent studies have indicated that the optimum head for a UPHS plant is about 1000-1500 m. The highest head, single- 200 NOTE: stage RPT unit built today is 620 m. COSTS AS OF SEPT. 1978 PRICE LEVEL Therefore, to utilize available equipment TEN HOURS STORAGE, 2000 MW CAPACITY and operate at a high head requires a two- 100 - drop scheme, i.e., two single-stage tur- bines in series with an intermediate reser- voir. High-head multistage RPT units and 600 900 1200 ISOO tandem units (Pelton-type waterwheel plus HEAD, M a multistage pump) are available but are capacity-limited and expensive. The de- Fig. 8. Effect of Head on Single-Drop velopment of high-head, single- and two- UPHS Plant Cost (adapted from stage, regulated, RPT units is economically Ref. 2) attractive, but requires extensive tech- nological development. conditions of the surrounding rock. Tech- The head, unit size, plant size, plant niques should be developed to determine arrangement, type of hydraulic machinery, the underground geotechnical conditions and costs are not clearly defined. Thus, with surface detection methods. further system studies arj r1 quired to determine optimum conditions. Studies are Design, research, and testing programs also required to reduce plant construction are required for the development of large- time, improve mining techniques, and pro- capacity, high-head, single- and two-stage vide rock mechanics criteria for under- regulated RFT units. Research is needed ground reservoirs. to increase the unit size, efficiency, and regulation of multistage pump/turbines ACKNOWLEDGMENTS Coupled with the turbomachinery development for UPHS plants, investigations of the The research activities in underground motor/generator design requirements should pumped hydro storage, which formed the be conducted. System studies to determine basis of this paper, were funded by the the optimum head, equipment arrangement, Division of Energy Storage Systems, Office unit size, type of turbomachinery, and of Conservation, U.S. Department of Energy. plant size are required. REFERENCES In addition, specific site locations should be identified "for the near-term con- lSiting Opportunities in the U.S. for Com- struction of UPHS plants. pressed Air and Underground Pumped Hydro Storage Facilities, Acres American, Inc., CONCLUSION New York (1976) UPHS is a feasible energy storage 2Underground Hydroelectric Pumped Storage - scheme for utility peaking service. It An 'Evaluation of the Concept, DOE (Soli- offers potentially large benefits, espe- citation 6-07-DR-500), Chas. T. Main, Inc. cially in premium fuel savings. Siting of (June 1978). a UFHS plant will be determined by the availability of competent geologic condi- 3Degan, J.R., Allis-Chalmers Hydro-Turbine tions and the proximity of load centers Division, unpublished studies (1977-1978). and transmission lines. Construction tech- niques for the underground reservoir and powerhouse are well established, but im- provements should be pursued..' Environmen- 303 "Applied Research on Energy Storage and Conversion Photovoltaic and Wind Energy System, Vol. I, II, and III, General Electric Space Division under contract No. NSF C-75-22221 (Jan. 1978). 5Gokhman, A., University of Miami, unpub- lished work under ERDA contract No. EC- 77-S-05-5516 (1977-1978). 6'Underground Pumped Hydro Storage and Compressed Air Energy Storage: An Anal- ysis of Regional Markets and Development Potential, ANL-R-77-3485-1, Harza Engin- neering Co. (1977). 7Hartraann, 0., and W. Meier, Developments in High-Head Pumped Storage, Water Power 22(3):102-106 (March 1970). "Mitchell, W.S., Underground Pumped Hydro Storage, paper presented at the Engineering Foundation Energy Storage Conf., Pacific Grove, Cal. (Feb. 8-13, 1978). 9Pfaffin, G.E., Future Trends in Hydro Pumped Storage Equipment, presented at the American Power Conference, Chicago (April 29-May 1, 1974). 10An Assessment of Energy Storage. Systems Suitable for Use by Electric Utilities, EM-264, EPR1 Troject 225, ERDA E(ll-l)- 2501 Final Report, Public Service Elec- tric and Gas Company, New Jersey (1975). "Schmid, E.M., W.F. Adolfson, and C.K. Jee, Principal Geological Issues Related to Underground Pumped Hydro Storage, draft report contract No. EM-78-C-01-5114 (Sept. 1978). ESwiecichi, I., Trends in Pump-Turbine Design, International Water Power and Dam Construction, 29(1):45-47 (Jan. 1977); 39(2)::42-45 (Feb. 1977). 13Willet, D.C., Underground Pumped Storage Research Priorities, EPRI AF-182 TPS 75- 618 Planning Study (April 1976). 304 PROJECT SUMMARY Project Title: Underground Pumped Hydro Storage-Single and Double Stage Reversible Pump/Turbine Development. Principal Investigator: John Degnan Organization: Allis-Chalmers Corporation Hydro-Turbine Division P. 0. Box 712 York, PA 17405 717-792-3511 Project Goals: Extension of the present state-of-the-art technology for high head single and double stage pump/turbines to heads beyond 500 and 1000 meters respectively; Generation of preliminary machine designs and corresponding hydraulic performance which together will satisfy the operating characteristics specified in the contract made between Allis-Chalmers and the Department of Energy (DOE)j Projection of the costs for the various alternate machine designs developed through tne study; Evaluation of the feasibility of the single and double stage pump/turbine propositions as applied to Underground Punped Hydro Storage (UPHS). Project Status: It should be noted that this documentary addresses only the initial phase of a large development project extending into 1982. The total project is concerned not only with the con- ceptual development of mechanical and hydraulic designs but with actual model testing of prototype heads. The project's first phase has been completed and fully documented in a report issued to the Argonne National Laboratory. Completion of the project's next phase includes the preliminary design of multi-stage high head reversible machinery for the storage of hydro-energy. This multi-stage concept as well as the model testing for the single and double stage units which are addressed in this paper will be developed with funds appropriated for 1979. Contract Number: 31-109-38-4301 Contract Period: The first phase of the overall project was started June 1, 1978 and completed in late September 1978. Funding Level: $100,000.00 Funding Source: Argonne National Laboratory 305 EVALUATION OF ONE AND TWO STAGE HIGH HEAD PUMP/TURBINE DESIGN FOR UNDERGROUND POWER STATIONS John Degnan Allis-Chalmers Corporation Hydro-Turbine Division Box 712, York, PA 17405 ABSTRACT This paper attempts to summarize the significant design considerations found during development of one and two stage reversible hydro machinery for the high head under- ground pumped hydro storage (UPHS) project currently funded through the U.S. Department of Energy (D.O.E.). The immediate objectives of the initial phase of this D.O.E. investigation are as follows: " Extension of the present state-of-the-art technology for high head single and double stage pump/turbines to heads beyond 500 and 1000 meters respectively. To generate preliminary machine designs which will comply with the required operating characteristics for UPHS and also serve as a basis for development of model testing procedures and equipment. To project the manufacturing costs of any proposed hydro-turbine equipment resulting from the investigation. ' Evaluate the feasibility of the one and two stage pump/turbine propositions for operation at a rated power of 500 megawatts under heads of up to 1000 and 1500 meters respectively. Examples of the preliminary single and double stage machine configurations are included within. Discussion of the special analyses and results leading to these machine designs is given and finally, conclusions as to the feasibility and limitations of each type of machine are drawn according to their expected cost, mechancial and hydraulic performance. INTRODUCTION from current design and operation expe- rience. From the outset of the project, it was believed that the maximum head which Also, four separate double stage designs could be accommodated by a single stage were investigated. One of these was machine was approximately 1000 meters rated at 350 megawatts operating at a and 1500 meters for double stage rated 1250 meters head while the others machines. Also it seemed obvious that were designed for a rated power of 500 the economics would not justify a larg- megawatts operating at rated head of er two stage concept where head condi- 1000, 1250, and 1500 meters respective- tions allowed a single stage applica- ly. The 1000 meter machine will be used tion. For this reason, both concepts to exemplify significant design consid- were developed within their expected erations of the two stage pump/turbine aJlowable head ranges. concept. Three single stage pump/turbines were The feasibility of the mechanical designs developed with rated net heads of 500, for these machines was verified using 750, and 1000 meters respectively and very accurate computer models which pre- each with a rated power of 500 mega- dict structural behavior and stress con- watts. The 1000 meter machine will be ditions based on the finite element used to exemplify the points of discus- method of analysis. sion since this machine is most removed 306 The hydraulic shape of the water passages and the performance for the single stage machines were predicted from existing high head pump/turbine model configura- tions and test data. The hydraulic characteristics of the two stage machines required more rigorous studies. The per- formance estimates, for instance, involv- ed using known two stage pump performance curves or a combination of the perform- ance characteristics of two single stage units which were similar to the respec- tive stages of the pump/turbine under study. Any expected losses between stages were then accounted for. Of course, determination of the actual hydraulic performance can only be real- ized when the model tests are made. SINGLE STAGE PUMP/TURBINE DESIGN rig. 2 until »t«i« 1000 Mttr-41:trlbutor flu GENERAL SPIRAL CASE - STAY RJMG Figures 1 and 2 show the 1000 M single The most significant criteria used to stage pump/turbine preliminary design. design the configuration of these inte- The major component designs were gen- grated components is that their mechan- erated automatically using the IRIS ical interaction be related so any net program. The engineering design cal- moment generated in the stay ring crown culations performed in IRIS proved to during the design condition (maximum give a sensible optimized design. This waterhammer pressure outside the wicket is substantiated through the special gates and tailwater head inside the gates) computer studies whose results indicated and pump prime (full shut-off head inside only minor geometry changes would be closed wicket gates with maximum static necessary to afford a sound single stage pump/turbine design. Flu. 1 Sln*l. «t>||» 1000 F/T - dl.trlNKor nrcllon ne. Stress Distribution! 1 I Pump Shut-off * Fig. 3 Stay ring vane stresses 307 head outside) cause equivalent tensile stresses on the outside and inside diam- eter of the stay vanes respectively. Such a design criteria truly produces an optimized structure since this technique minimizes the stress amplitude section by section through constructive use of the residual ring moments. Figure 3 shows how these moments are controlled by the section's spiral case to stay ring attachment location. Note how the spiral case attachment point slides along the stay ring cone until the optimizing criteria are satisfied. DUcMi|« nai vittlcal lewd* HEADCOVER The mechanical design of this component It should be noted that cheaper, lighter is optimized based on the structure's discharge ring designs can be proposed rigidity or its resistance to distort so that the only compensation for the under load. Such parameters as maximum extreme head cover load is in the foun- hoop stress, maximum angular rotation dation reaction. This huge load, how- of the head cover and maximum deflection ever, becomes increasingly difficult or at the turbine bearing support are impossible to accommodate as pressure checked against preset limits in the heads increase and machines necessarily IRIS design system according to a log- become smaller. arithmic type head cover analysis. The structure is incremented in size accord- DISTRIBUTOR SECTION ANALYSIS ingly until all optimizing criteria are satisfied. The following boundary con- The 1000 meter single stage pump/turbine ditions were fovced on the IRIS design distributor components have been analyzed of the 1000 M single stage pump/turbine: using the axisymmetric-two dimensional representation including the discharge Maximum Hoop Stress 17000 PSI ring, spiral case, stay ring, head cover, Maximum Angular Rotation .0008 radians and connecting bolts (Figure 5). Turbine Bearing Radial .0500 inch Deflection This analysis will provide the following information: It will become apparent from the stress analysis of the distributor section why a. A general understanding of the these parameters characterize the general deflection and stress pattern in behavior of the head cover. the assembly. DISCHARGE RING b. A good evaluation of the interactive forces between components, for exam- This component draws its sizing from the ple, the calculation of the radial basic head cover configuration. The forces developed at the stay ring most important feature of the discharge to head cover and stay ring to dis- ring is that it is of heavy construction charge ring bolting flanges. and can support its own pressure loading without relying on a distributed founda- c. The calculation of the necessary tion reaction. This full strength dis- bolt prestress required to eliminate charge ring design has the advantage of fatigue. allowing only 40% of the vertical head cover load, F , to be realized at the d. Th-s foundation loading and required machines' foundation; the remaining. 60% prestress of the anchor bolts. of the head cover load is compensated by the discharge ring load. Figure 4 Ten basic loading conditions were run shows this concept graphically. simultaneously by the finite element pro- gram . These cases are factorized and 308 superimposed using a post processor pro- gram to represent the following relevant operation conditions: Mormal Running Condition. The machine operates at maximum static hend. The pressure distribution inside the runner periphery follows a forced vortex par- abolic law assuming a" mass of water rotating at half the speed of the runner. The maximum tailwater head is applied inside the runner seals. Pump Shutoff Condition. This loading case represents the condition occurring during a pump start after the release of the pressurized air. The pressure acting outside of the wicket gate is equal to the maximum static head. The pressure acting from the wicket gate to the runner periphery is equal to the maximum shutoff head. This pressure decreases parbolically from the periphery of the runner to the runner seals as described in the first case. The pres- sure inside of the seals is equal to the maximum tailwater head. Some of the important information and Pressure Rise, Wicket Gates Closed. This design criteria extracted from the com- loading case represents the normal tur- puter study are defined in Figure 5 and bine shutdown operation with the maximum tabulated in Table 1 and 2. They are: pressure rise and maximum tailwater head acting, respectively, outside and inside - "A", axial deflection of top deck at of the wicket gate. the inner radius. The value is cal- culated relative to the head cover to Runaway Condition. This loading case stay ring interface at the bolt center- represents the extreme condition of tur- line. bine load rejection with the wicket gates stuck open. The maximum pressure rise - "B", radial deflection of top deck at acts up to the runner periphery. This the inner radius. This value has to pressure decreases parabolically from be considered very carefully for this the runner periphery to the runner seal design where the turbine guide bearing following a forced vortex law with the is bolted directly to the top deck. mass of water rotating at 50% of the runner runaway speed. The maximum tail- - "C", axial deflection of top deck at water head is used inside of the seals. the outer radius. Results Interpretation. The use of - "D", radial movement of wicket gate deflection results for the purpose of top bearing. rigidity evaluation or field test com- parison has to be based on differential - "E", radial movement Of wicket gate values between two points of the struc- intermediate bearing. ture. This is necessary in order to establish a common fixed datum. For - "F", gap opening at head cover to stay example, the vertical deflections of a ring seal interface. This value is head cover can be defined as the dif- checked to assure that "0" ring extru- ferential vertical movement between any sion will not occur. point on the head cover and the stay ring bolting flange interface. 309 Table 1. HEAD COVER DEFLECTIONS AND ANGULAR ROTATIONS (cm & radians) Node Generating Pump Prime Pressure Rise Runaway A* .10018 .15990 .03608 .14961 B .04012 .06165 .01732 .05861 C* .01082 .02055 .00077 .01941 D .03379 .05196 .01459 .04958 E -.03069 -.04820 -.01348 -.04561 F .02388 .04053 .00853 .03988 G .06635 .10420 .02963 .10232 J* .11848 .18468 .04760 .17294 a .00062 .00096 .00024 .00089 B .00045 .00070 •00020 .00067 e .73116 .73310 .82353 .75144 Table 2. FOUNDATION LOADS AND COMPONENT INTERACTIVE FORCES , KG K 2,830,336 3,428,073 3,068,820 4,135,666 L 1,518,021 1,286,229 2,638,506 2,204,468 M 4,573,606 7,047,673 2, 005, 594 6,701,236 N 4,335,834 7,223,523 1,122,144 6,500,989 *referenced to datum - "G", distributor height growth. This 9 = 1.0 for an infinitely rigid value is particularly important at ribbing pump start and at shutdown when the additional clearance has to be filled by the movable gate end seal to limit leakage and wire drawing. Foundation Load and Components Inter- action. The component interactive forces • "J", axial deflection at the shaft and foundation loading are defined in seal bolting flange. This is usually Figure 5. These are the radial forces the largest axial deflection recorded tending to separate the stay ring from in a bead cover. the head cover "K" and also, from the discharge ring "L", plus the inner and "a", head cover top deck angular rota- outer foundation reactions "N" and "M". tion. This is one of the most impor- These values are tabulated in Table 2. tant head cover design criteria. Its magnitude is inversely proportional to WICKET GATE the head cover rigidity. Experience has shown that a should be lower than A critical balance exists between the 0.001 at the normal running condition hydraulic and mechanical requirements of to insure a satisfactory performance a guide vane system. It is essential to in a straight Francis turbine. A more assure that neither the hydraulic per- conservative value of 0.0009 radians formance nor the mechanical considera- is used for pump/turbines. tions of long life and reliability be compromised. "8", wicket gate bearings angular rotation. The same criteria used for The wicket gate system for the 1000 meter pump/turbine was also generated by the IRIS program. The preliminary design "9", is the ratio of gate bearing of the wicket, gate is achieved by incre- angular rotation to head cover top menting the gate geometry until all deck angular rotation. The magnitude static and dynamic criteria (stress, of 6 is an indication of the efficiency deflection and torsional frequency) are of the transfer, by the ribs, of the satisfied. A schematic representation shear load between the head cover top of this optimization process used by and bottom deck. IRIS is shown in Figure 6. 310 MCLINIHUT STRESS AKALtttl INITUL SUN * ICAF GEOMETRY •EARIHC REACTIONS: TORQUE! LINKAGE OPTIM1I*TION t SERVOMOTOR SUING I st>ESS, OEFLECTIONS I RCAC* INCRCKEHTKL GEO- DON CALCULATIONS FOR HCTRV CHANGt ON DIFFERENT CONDITIONS LEAF ON STEH ACCORDING TO UMSATW.tB OfSICl CRITERIA CHECK LIMITING DESIGN CRITERIA - STRESSES - DEFLECTIONS • SLOPE *f 9EMIWS • STEM ANGULAR ROTATION - lEARING PRESSURES j —-czml Gate Squeeze Condition. The gate stem is subjected to the maximum torque avail- \ FIHAj QESICH • OUTPUT able at closed position while the gate leaf is subjected to the maximum tran- g. 6 Icii pro^ru q«t optimisation sient head and the intergate seal reac- tions which are presented as contact pressures in Figure 8. Shear Pin Failure Condition. The gate stem is subjected to the maximum torque required to break the shear pin at or near the closed position. This condition -TIUUULATIQSAI. CM3TM1KI occurs when an obstruction becomes lodged between any two gates preventing further -THIUST COXSTUIVT closure of the gates. The highest stress- es result from the blockage which is located at the upper tail end of the leaf. Since full servomotor torque is available to these two gates alone, they must be protected by insertion of a shear pin in their linkage mechanisms. The pin was designed to break at 225% of normal TAIL SEAL COfTTACt LVtl _, —TtANStATKHAL CONSTRUCT servomotor torque. Fatigue Analysis. Once all the signifi- cant static loading conditions are analyzed, evaluation of the fatique life _SOSt SEAL COfTrXT LIKE is done. The Miner's cumulative damage ^ossnwwT rule is typically used which is based on the linear summation of the fractions of fatigue damage expressed in terms of the cycle ratio (n./N ) WICKET GATE STRUCTURAL ANALYSES f where n. = estimated if of cycles of Since the classical Beam theory used in each mode of loading IRIS does not adequately represent the behavior of the leaf and stem to leaf intersection, a three-dimensional finite element static analysis of the gate was conducted. Figure 7 shows the mathemat- ical model. The structural response to the following loading cases has been evaluated: 311 # of cycles to failure as gives solid evidence to the integrity of determined from material its conceptual design. specifications. This is a function of the mean and cyclic stresses occurring i: M u at potential sites of fail- ure. Z(n±/Nf) = 1 failure is hypoth- uoo esized 1050 LOW A conservative approach in using the analysis is to limit the accumulated damage to an order of magnitude less than the predicted level so that Z(n./N^) < 0.1 indicates a reliable design. HYDRAULIC PERFORMANCE In addition to the mechanical develop- ment, the hydraulic performance of the machines have been evaluated. Figure 9 shows the performance as a pump, CONCLUSION These studies indicate that both the tit. S fmf p«t[01MK. 1000 HUr but ll»|U ftM* « wicket gate and the components of the distributor section will, in fact, behave according to the criteria used to size them. The analyses show, for instance, that the 1000 meter machine will: * develop foundation loads which have magnitudes of only 40% of the respec- tive head cover loads (18,927,000 Kg at Pump prime) ' that the maximum hoop stress developed in the head cover is 17,970 psi ' the axial stress at the inside of the stay vane during pump prime and at the outside of the vane during the tran- sient design conditions are approx- imately equal at each spiral case sec-r tion. Also, it was evident that, despite the extreme conditions of operation assumed in the wicket gate analysis, the maximum accumulated damage was 0.000154 for a period of 50 years or less than 0.002% of the damage required for fatigue fail- ure. Subjected to normal design loads, a near infinite life can be expected for these gates. The general stress and deflection Fig. 19 D.O.E. two atage 1000 meter P/T - distributor response of this machine to its loading analysis finite clenent nodel 312 TWO STAGE PUMP/TURBINE DESIGN GENERAL Figures Hand 12 show the 1000 meter double-stage pump/turbine preliminary design. Some of the basic component designs such as the impellers, head cover, gate and spiral case configura- tions have been initially sized by the IRIS program. Other components such as the stay ring, intermediate water passage, and the shafting have been designed by careful layout coordinated with the IRIS designed components. At the outset of the development project, little was known as to the interactive behavior of the components making up this machine. In order to determine this, a static stress and deflection analysis was performed on an intermediate design sub- the calculation of the bolt prestress jected to its normal loading during gen- required to eliminate flange slippage eration mode. Although the study was and bolt fatigue. very preliminary in nature, it provided the required information to develop a foundation loading required to design feasible solution. the powerhouse. DISTRIBUTOR SECTION The proposed design as shown in Figures 11 and 12 incorporates many modifications An intermediate design of the 1000 meter based on the above analysis. two stage pump/turbine has been analyzed using an axisymmetric-two dimensional SHAFTING SYSTEM DESIGN representation. Figure 10 depicts the mathematical model developed for the The double stage machines were initially finite element structural analysis. The developed so that each stage generated analytical results provided for: equal amounts of power, however, a basic problem was brought to light concerning ' a general understanding of the deflec- this design: The critical speed of the tion and stress pattern in the assembly shafting system is greatly influenced by the amount of overhung runner mass and • a good evaluation of the interactive its distance from the turbine guide bear- forces and moments between components. ing. It was apparent that it would be This information is essential to prop- advantageous to design the second stage erly design connection of the compon- runner to carry as much of the impellers ents. mass as possible due to its nearness to the guide bearing. This was accomplished by designing the first and second stage to develop 40% and 60% of the power respectively. This concept was verified through critical speed dynamic analyses performed on the proposed pump/turbine impeller, shafting and bearing with assumed shafting and generator systems. The modification caused the fundamental critical speed of the shaft system (lateral mode) to be increased from 8.3 Hz to 13.0 Hz or to 73% above the normal speed. M». 11 Ito •teg* 1000 Mter r/T ~ distributer Mctloa 313 Another subject of notable consideration HYDRAULIC PERFORMANCE was the seal clearance of the first stage shafting. Again, due to the long over- In addition to the mechanical develop- hung length of the turbine shafting ment, the hydraulic performance of the downstream of the guide bearing, large machines has been evaluated. Figure 13 radial deflections could occur during shows the performance of the machine as transient conditions which must be a pump. accommodated without compromising proper sealing and thus performance. It is CONCLUSION proposed that the interstage seal be designed so that during transient loading Although further development work is conditions it will serve as a secondary indicated to design double stage machines guide bearing. A carbon impregnated which would have the same degree of metal material has been- developed which optimization as is built into the single resists corrosion and has varying degrees stage machine, the studies to date are of self-lubrication and load carrying more than sufficient to prove the fea- capacity. If contact should occur dur- sibility of the two-stage concept from ing short transient periods between this a mechanical standpoint for heads between stationary seal material and the rotating 1000 and 1500 meters. shaft, the seal can indeed function as a self-lubricated guide bearing. FEASIBILITY WICKET GATE The preliminary designs and hydraulic evaluation of these machines give witness The same design criteria considered on to their useful potential. the single stage machines apply to this wickfit gate design. Again, finite ele- The manufacturing cost analysis produced ment stress, deflection, and fatigue resultc showing that for the same applica- analyses were performed to verify the tion, the two stage machine costs are 55% initial sizing. According to these greater than the single stage unit. It studies, the wicket gate design presents should be pointed out that project con- no problems which would adversely effect struction, machine efficiency, submergence the feasibility of this two-stage con- and maintainability all must be weighed cept even when subjected to the limiting and evaluated to determine the feasi- head of the 1500 meter machine. bility of such a project. Current con- struction rates, for instance, probably have the greatest influence on the feasibility of a project. The amount of excavation for the project site is depend- ent on the required submergence of the hydraulic machine. This submergence requirement is inherently greater for a single stage unit than a two or multi- stage machine. REFERENCES 1. Chacour, S. and Graybill, J. D., "IRIS, a Computerized High Head Pump Turbine Design System," ASME Paper No. 76-WA/FE-12, December 1976. 2. Chacour, S., "A High Precision Axisymmetric Triangular Element Used in the Analysis of Hydraulic Turbine Components," ASME Paper No. 70-FE-19, Reprinted in the Transactions of the ASME, Journal of Basic Engineering, December 1970. 3. Chacour, S. and Degnan, J., "Structur- al Optimization of High Head Pump/ Turbines," CEA, March 1977. 314 PROJECT SUMMARY Project Title: Multistage Turbine-Pump with Controlled Flow Rate Principal Investigator: Alexander Gokhman Organization: Department of Mechanical Engineering School of Engineering and Architecture University of Miami Coral Cables, FL 3.3124 Telephone: (305) 284-4848 Project Goals: The goal of this project was to conduct a conceptual design of one and two stage turbine-pumps with controlled flow rate for use with pumped hydro storage power plants. Both the hydraulic and mechanical designs of the turbine-pumps were to be evaluated. Project Status: The project was completed on September 30, 1978,.and a final report was prepared and submitted to DOE/STOR in October 1978. The hydraulic analysis shows that this new machine has the unique ability to regulate the flow rate even in the pump mode of operation. This unique feature makes the new turbine-pump especially attractive for above or below ground pumped storage plant applications to store the energy produced by plants utilizing non-controllable natural energy sources such as solar energy, wind energy, etc. The new turbine-pump with controlled flow rate is more complicated and, consequently,, more expensive than conventional unit for the same operating parameters. However, the new machine does not require an operating valve since the outer cylinder of the movable upper band of the distributor acts as a cylindrical valve. Therefore, it is not certain that the new turbine-pump will cost more than a conventional unit which requires a costly operating valve. The comparison of the new machine and a conventional unit requires hydraulic experiments and technical design by leading firms in the field of hydraulic machinery. Even if the new machine proves to be more expensive than a conventional unit with a regulating valve, the additional cost will be offset by the energy savings due to flow regulation. Contract Number: EC-77-S-05-5517 Contract Period: June 1, 1977 -Sept. 30, 1978 Funding Level: $74,582 Funding Source: Department of Energy, Division of Energy Storage Systems 315 MULTISTAGE TURBINE-PUMP WITH CONTROLLED FLOW RATE Alexander Gokhman, Nail Ozboya University of Miami Department of Mechanical Engineering Coral Gables, Florida 33124 ABSTRACT The presented paper shows the results of the preliminary part of the development of the new hydraulic machine "Multistage Turbine-Pump with Controlled Flow Rate." The hy- draulic analysis of the turbine-pump shows that this new machine has the unique feature to regulate the flow rate even in the pump mode (the conventional turbine-pumps cannot control the flow rate in the pump mode). This unique feature makes the new turbine-pump especially attractive for the pumped-storage plant applications to store the energy pro- duced by the plants utilizing non-controllable natural energy sources such as solar energy, wind energy, etc. Obviously, during the process of storing energy, the conventional ma- chine will cause the waste of energy produced due to its lack of ability to regulate the flow rate. The new turbine-pump with controlled flow rate is more complicated and conse- quently more expensive than the conventional one for the same parameters (head, power) as can be concluded from the conceptual design of the new machine. However, the turbine-pump with controlled flow rate does not require an operating valve, since the outer cylinder of the movable upper band of the distributor (in case of multistage machine, the distributor of the first stage) acts as cylindrical valve. Therefore it is not certain that new tur- bine-pump which does not need the operating valve will cost more than conventional turbine- pump with the operating valve (It is well known that the high head valve for pov;er plants is very expensive). Even in the case that the new machine is more expensive than the con- ventional one with the operating valve, the additional cost of the new machine will be offset by the energy savings due to the unique feature of flow rate regulation. INTRODUCTION Modern turbine-pumps utilized for gidly fixed. Therefore all conventional pumped storage are divided into two groups. pumps, one-stage or multistage, have some The first one is one-stage turbine-pumps. disadvantages in addition to lacking the The largest head for such machines is 800 m ability to regulate power in pump mode. (Hitachi (Japan)). Allis-Chaltners (U.S.A.) Neyrpic (France) is trying to develop a one advertised the conceptual design of a one- stage turbine-pump, but their proposal does stage turbine with a 1000 m head, but other not look very promising. The absence of leading companies such as Escher-Wyss power regulation in pump mode is a big dis- (Switzerland), Hydroart (Italy) and Nyer- advantage in certain conditions, e.g. when pic (France) consider the limit for head the nominal power of a unit is comparable for a one-stage machine to be around 500 m. to the power of the system. Lack of abil- One-stage conventional turbine-pumps have ity of multi-stage turbine-pumps to reg- the ability to regulate power in turbine ulate power in turbine mode is even a big- mode, however in pump mode they can only ger disadvantage, because the machine works work with nominal power or not work at all. in turbine mode only during peak time with Multi-stage machines cannot regulate pow- sharp changes in power. er in either mode, but they seem to be more reliable especially for high heads owing The ability to regulate power is vi- to the absence of guide gate apparatus with tal to pumps working with non-controlled adjustable vanes. In a one-stage turbine- output electrical plants, e.g. solar plants, pump working as a pump the guide gate ap- because it leads to inevitable power losses, paratus become the sources of strong vibra- except for those few moments when the pow- tions, because the flow after the runner is er of the plant matches the sum of capaci- unsteady and causes pressure pulsations ties of several pump units. These losses around the guide vanes, which are not ri- can be as high as 95% of the capacity of a 316 - single unit, because the only available way the horizontal flange of the scroll casing of regulation in such a case is a complete by means of studs (26) . The outer cylinder shut down of one or more units. However of the lower ring is fastened to the verti- there exists a possibility to eliminate the cal inner cylinder of the scroll casing by abovementioned disadvantages of convention- means of studs (25). As seen from Fig. 1, al turbine-pumps in both turbine and pump the fastening of the distributor to the modes. Power regulation can be achieved in scroll casing by the horizontal flange and both modes by changing at will the height vertical outer cylinder gives the distri- of the water passage in tho runner and the butor absolutely fixed position and good distributor, the latter having rigidly se- rigidity. The outer cylinder of the mov- cured vanes. ' able upper band (18) is placed inside of the inner cylinder of the upper ring (21). This work accomplished under the De- The movable upper band is driven by the partment of Energy contract was devoted to vertical servomotors (37). The servomotor the conceptual design of such a machine. guide rods (19) are fastened to the flat According to this contract the researchers bottom part of the movable upper band (18) also had to develop all the necessary hy- on one end and to the servomotor rod on the drodynamic programs for this task, so the other end by the servomotor rod connector accomplished work consists of the proper (39). The rubber seals (20) are placed hydraulic analysis of the turbomachine with between the vertical inner lower cylinder variable water passage height in the dis- of the scroll casing and the outer cylin- tributor, the conceptual designs of one- der of the movable upper band. There are stage and two-stage turbine-pumps of this additional rubber seals (22) between the type accompanied by all the necessary hy- inner cylinder of the upper ring and in- draulic, force and stress calculations. ner cylinder of the movable upper band. The goal of this entire project is to The turbine cover (32) is fastened investigate the feasibility of machines rigidly to the upper horizontal part of the with this new principle. scroll casing by means of studs (33) and to the lug (45) of the upper ring by means DESCRIPTION OF THE DESIGN of studs (46). It is clear from Fig. 1 that the turbine cover, the horizontal ONE-STAGE TURBINE-PUMP flange of the upper ring (21) and the horizontal flange of the scroll casing (23) The one-stage turbine-pump was de- are pressed together. This design pro- signed for a 450 m head, 48 MW capacity and vides the turbine cover with fixed and 1.8 m diameter. rigid position. The one-stage turbine-pump with con- The servomotor guide rod (19) which trolled flow rate (TPCFR) comprises the moves the movable upper band (18) is scroll casing (23) fixed inside of the se- guided through the bronze bushing which is condary support concrete structure (24), tightly pressed in to the upper ring of the distributor placed inside of the scroll the distributor. casing and secured to the scroll casing body, the runner placed inside of the dis- The servomotors are mounted on the tributor and secured to the turbine shaft horizontal plate (44) which is welded to (9), and the draft tube (30) attached to the turbine cover (32) and the support the lower ring (17) of the distributor (38). The servomotor rod is also guided (see Fig. 1). by the bushing placed in this horizontal plate (44). There are 4 servomotors and The distributor comprises the upper the diameter of the servomotor piston is ring (21), lower ring (17), guide vanes 200 mm. The pressure of the oil in the (16) welded to the upper and lower rings, servomotor is 70 kgf/cm2. and movable upper band (18). The movable upper band (18) is made up of the outer The outer cylinder of the movable up- horizontal flange, outer cylindrical part, per band (18) has enough height to close flat bottom part and inner cylindrical the water passage completely. The rubber part, all forming a solid body. The lower seal (20) between the scroll casing and ring is formed by the outer cylindrical this cylinder will prevent the leakage of part, flat top part and inner cylindrical water between them. In order to prevent part. The horizontal flange of the upper the leakage of water between the horizon- ring (21) of the distributor is fastened to tal bottom of movable band (18) and the 317 I 318 lower ring (17), a rubber seal (47) is in- is attached to the guide rod (8) by means of stalled in the groove of the lower ring. studs (13). The streamlined hub extension This rubber seal is pressed by the adjust- (14) in turn is mounted to the streamlined able metal ring (48) which is fastened to hub (3) by means of bolts (15). The bolts the lower ring (17). It is obvious that (11) secure the upper ring (4) of the runner this seal is replaceable. This design to the hollow shaft (9). gives the upper movable band the ability to function as a cylindrical valve. During the regulation of the flow rate, the cross-sectional area of the water pas- The labyrinth seal (43) for the run- sage is changed by the vertical movement of ner is mounted to the inner cylinder of the the guide rod (8) inside the hollow shaft movable band (18). (9) and the subassembly of the streamlined hub (3), hub extension (14), connecting ring The runner comprises the lower ring (?) and the upper cover disk (6) of the run- (2), upper ring (4), runner blades (1) ner. The lubricating oil provides the welded to the upper and lower rings, smooth sliding of the guide rod (8) in the streamlined hub (3), hub extension (14), hollow shaft and the bushing (10) guides the connecting ring (5) and upper cover disk rod (8). The leakage of oil between the (6) of the runner. guide rod and the hollow shaft is prevented by the rubber seals (12) mounted to the up- There are 15 runner blades (1) and per ring (4) of the runner. each blade is designed for the minimum height of the water passage ((bo)opt), the The vertical movement of the guide rod upper part of the blade is a vertical ex- (8) is provided by regulating the pressure tension of the blade cross-section at of the oil in the servomotor on the upper bo = (bo)opt. The runner blades (1) are and lower sides of the piston (7) mounted welded to the lower ring (2) of the run- to the guide rod (8). The oil supply to the ner. The streamlined hub (3) of the run- servomotor in the hollow shaft is similar ner is then placed on the lower ring (2) of to that used for adjustable blade turbines. the runner, the runner blades passing through the slots of the streamlined hub. Guide bearing (41) is the conventional The upper ends of the runner blades (1) are oil-lubricated type bearing. welded to the upper ring (4) of the runner. The connecting ring (5) is welded to the The draft tube (30) is attached to the upper cover disk (6) of the runner and they support disk (28)' which is in turn mounted form a solid piece which is placed on the to the lower ring (17) of the distributor streamlined hub (3) of the runner and fas- by means of cap screws (29). The labyrinth tened to it by means of cap screws. seal (27) attached to the lower ring (2) of the runner and inserted into the grooves of The guide pivot (49) is secured to the draft tube flange reduces the water loss the upper cover disk (6) on the upper side from the upstream of runner to the draft by cap screws and screwed to the stream- tube. lined hub (3) on the lower side and it passes through the hole in the upper ring The governor of the machine is similar (4). While the guide rod guides the to the governors of conventional Francis streamlined hub (3) and the upper cover turbines. The only difference is that the disk (6) assembly during its vertical move- servomotors of the distributor have to be ment, it also gives additional strength to synchronized with the servomotor in the this assembly. The alternate guide pivots hollow shaft of the turbine. (49) are left with holes in them in order to direct the flow from the space between the turbine cover (32) and the upper cover TWO-STAGE TURBINE-PUMP disk (6) to the downstream of the runner blades (1), therefore reducing the excess The layout of two-stage TPCFR is shown pressure in that space. The leakage of in Fig. 2. The two-stage TPCFR is designed water in that space to outside is prevent- with the stages having the same hydraulic ed by the labyrinth seal (35) between the parameters, i.e., the same geometry of the hollow turbine shaft (9) and the seal sup- water passage and the same diameter of the porting disk (34) which is mounted to the runner. turbine cover (32). The main differences in the designs of The streamlined hub (3) of the runner the mechanism of flow rate regulation for 319 320 one-stage and two-stage TPCFR are in the The diameter of the guide rod below the se- runners, shaft, shaft support and second cond stage runner is 168 mm. It passes stage distributor. through the bushing (51) mounted to the hole of the shaft. This bushing directs the The first stage streamlined hub (3) motion of the guide rod (8). The vertical of the runner is not driven directly by the movement of the guide rod (8) is transferred guide rod (8) in the hollow shaft (9), but to the streamlined hub (63) of the second by means of a special connecting cylinder stage runner by means of two carrying arms (71). The connecting cylinder (71) is made (76 and 93) which are inserted into the from two halves and connected to the upper rectangular holes in the guide rod (8). The cover disk (66) of the second stage runner. upper arm (76) is placed in the groove of the streamlined hub (63) and secured to it The second stage runner is essential- by the screws (77). The lower arm is per- ly the same as the one-stage machine run- pendicular to this upper arm (76) and is al- ner, except for two differences. The first so fastened to the hub (63). Both arms pass difference is that the upper cover disk through the openings in the turbine shaft (66) has a longer cylindrical part and an wall so that they can move up and down with additional inner disk which is secured to the guide rod (8) in the range of regulation the connecting cylinder (71) by means of (in our case 90 mm). The arms (76, .93) are screws (73). The connecting cylinder also secured to the guide rod (8) by means transfers the vertical motion of the upper of screw (78). The rubber seals (12) on cover disk of the second stage runner to the turbine shaft above the opening for the the streamlined hub (3) of the first stage upper carrying arm (76) prevent the leakage runner. The second difference is that the of oil. These seals (12) are pressed by the streamlined hub (63) of the second stage ring (94) which has two partial flanges in runner is not connected directly to the the opening for the arms. The labyrinth guide rod (8) but to the carrying arms seal (72) between the upper cover disk (66) (76 and 93). of the second stage runner and the inner cylinder of the upper ring (52) of the se- The design of the upper part of the cond stage distributor is mounted on the turbine shaft (9) above the first stage inner cylinder of the upper cover disk (66). runner is similar to that of a one-stage The hollow shaft (9) has a solid extension machine. However, the diameters of the (79) which is secured to the turbine shaft piston inside the turbine shaft and the by the screws (80). Since there is almost flange of the shaft are larger in order to no torque acting on the shaft below the se- provide a larger force by the servomotor cond stage runner, the shaft extension (79) in the shaft. Also the flange for connect- is this and the connection to the shaft (9) ing the first stage runner to the turbine does not have to be very strong. The only shaft (9) is wider in the two-stage ma- torque is due to the friction in the water chine. The diameter of the shaft below lubricated lower guide bearing (81). This this flange is smaller, since the torque bearing is necessary in order to eliminate on the shaft between the two stages is the shaft beat (the shaft of the two-stage half of the torque on the upper part. The machine is approximately 3 tn. long compared lower flange of the turbine shaft is used to the 1.2 m. long shaft of the one-stage to attach the second stage runner to the machine). The lower guide bearing (81) is shaft and the diameter of this flange is mounted inside of the metal base (82) and a smaller because of assembly requirements. water pipe (85) supplies the clean water for The lower end of the turbine shaft is the lubrication and cooling of the bearing. tightly pressed into the opening of the upper ring (64) of the second stage runner. The second stage distributor comprises There are two bushings (10 and 50) on the the lower ring (60), upper ring (52), sta- turbine shaft located between the runners. tionary guide vanes (86), movable guide They direct the movement, of the connecting vanes (55) and movable upper band (54). The cylinder (71) and each bushing is made of stationary guide vanes (86) are welded to two half cylinders. The turbine shaft (9) the lower and upper rings (60 and 52) and is hollow like in the one-stage machine they form a rigid structure. The lower and the guide rod (8) secured to the piston ring (60) is pressed tightly in the cylin- (7) of the servomotor is placed inside the drical recess of the lower turbine cover shaft (9). The diameter of the shaft hole • (58) and secured by means of studs (87). and consequently the diameter of the guide The lower turbine cover (58) itself is rod (8) are 320 mm. above the first stage tightly pressed into the recess in the hor- runner and 240 mm. between the runners. izontal flange of turbine housing (23) and 321 secured to it by means of bolts (59), the In our analysis heads of which are in the secondary con- is the height of a guide vane crete structure (24). The lower ring (60) at any time, is also secured to the lower turbine cover ls b (58). This arrangement gives a fixed du- V 0min' is the radial coordinate in rable position to the second stage distri- the cylindrical system, butor. The movable upper band (54) has is the value of r for the slots through which the stationary guide point 0 of intersection of the vanes pass. In the second stage, the mov- outlet edge of a guide vane able upper band (54) does not function as with an arbitrary streamline a valve since it is sufficient to have one in the water passage, valve for the machine (the movable upper is the value of r for the band (18) of the first stage distributor). point 2 of intersection of the Therefore the outer cylinder of the movable outlet edge of a blade with an upper band of the second stage distributor arbitrary streamline in the is shorter than that of first stage and it water passage for a certain goes inside of the groove in the upper ring value of L_, (52). The inner cylinder of the movable is the length of a segment a- upper band (54) has similar design as the long the outlet edge of a first stage and at the top position it goes blade connecting point 2 and inside the other groove of the upper ring the point of intersection of (52). The second stage distributor has the outlet edge with the low- also four movable guide vanes (55). These er ring, movable guide vanes (55) are uniformly dis- tributed between eight stationary guide is the tangential component of vanes (86). The lower ring (60) has sever- the absolute velocity, al slots under each movable guide vane (55) is the meridional component of which permit it to go down inside the ring. the absolute velocity, The movable guide vanes (55) are casted to- is the radial component of the gether with upper (88) and lower (57) jour- absolute velocity, nals. The upper journal (88) passes W is the relative velocity, through the hole in the movable upper band n is the speed of the turbine (54) and a nut (89) presses the upper end (rpm), is the specific speed, face of the guide vane (55) to the movable '11 upper band (54). The lower journal (57) is is the efficiency of the tur- connected to the second stage distributor bine, servomotor rod (90) by means of connector H is the head of the turbine, (91). The lower journal (57) passes u=7r-n/3O is the angular velocity of the through the bronze bushing (62) in the low- runner (1/sec), er turbine cover (58) and the upper journal U=wr is the corresponding velocity, (88) is guided by the bushing (63) in the g is the acceleration of free upper ring (52) of the second stage dis- fall, tributor. There are four second stage (3Q is the angle between the vec- servomotors (56) mounted to the lower tur- tor of absolute velocity V. bine cover (58) by means of brackets (92). and the tangential unit These servomotors (56) provide the vertical vector § Q at point 0, motion of the movable upper band during the regulation period. By is the angle between the vec- tor of relative velocity $. and the tangential unit HYDRAULIC ANALYSIS vector e 2 at point 2 and is E2 the angle between the The flow rate and thus the power of streamline and a vector per- TPCFR can be regulated by moving the ad- pendicular to the outlet edge justable hub and band along the axis of the at point 2. machine (in case of a multi-stage machine this has to be done simultaneously at each The flow rate according to Fig.3 is stage). This can be shown in the turbine mode for a simultaneous movement of the adjustable hub and band as follows: Q=J 2-ir-r2-Vm2'cos(e2).dJl2. (1) 322 Vu0=Vr0/tan(V (7) (8) Fig. 3. Cross-sectional view of runner and distributor. Fig. 4 shows the development on the plane of cross-section of the outlet edge of the blade by a surface of revolution formed by an arbitrary streamline in the meridional flow. From Fig. 4 it can be seen tnat (2) where U,=wr . From (1) and (2) one ob- tains )2 Fig. 4. Kinematics of flow at runner = 2-ir- outlet. By combining equations 3, 4 and 8, one receives •tan(62)'d!>2, (3) and from Euler's equation one has cos(e2)'tan(62)«dJ!,2) where V ,'r, is the moment of absolute ve- ul 1 locity about the axis of the turbine before the inlet edge of a blade. In fact (l+(l/bo-tan(eo)))-j cos(e2) V .r =V «r (5) 0 •tan(62)'d«,2) (9) because hydraulic losses between the guide vanes and the inlet of the runner are neg- 3 ligible and the flow is axisymmetrical. If the generalized mean value theorem The flow in the distributor is uniform and is employed and both sides are divided by 2 S from Fig. 3 it follows that D1 -H°* , Equation 9 becomes (6) and since 323 (2) (l/((cos(e2)-tan(e2)) -L2)) bOmin=°- The analysis of the power losses, in +l/(bo-tan(3o))), (10) the runner, distributor, draft tube (suction pipe), scroll casing, during the transition - - from scroll casing to the distributor, and L =L /D and friction losses at the upper and lower where r^r ^ 2 2 i and superscripts (1) and (2) indicate the shrouds showed that TPCFR has almost flat mean values. The Equation 10 can be writ- curves n=n(N/Nmax) for both turbine and ten as pump modes in comparison with Neyrpic and conventional reversible machines (Figures 5 and 6). (2) /((l/((cos(e2)-tan(B2)) 100 {{•Constant o W Turbine Mode 90 • (l+a/bo))))+(l/tan(eo))), 80 A= b lt is clear from thls where ^2min~ 0min" 70 Equation that if X^O and n^is constant, 60 Q.1 increases with increasing bn, but at a rate slower than that for the increase of SO V 10 20 30 40 SO 60 70 30 90 100 Geometrical parameter A characterizes 100 i is)I R'Constant the rate of deviation of the water passage Furap Mode within the runner from strictly radial to 90 radial-axial. Only for a water passage purely radial within the runner, Q.. is 80 proportional to b-, because in this case 70 5 =i: (see 3)an d and Omin 2min "}' *2 60 (cos(e2)'tan(e_)) are two exact con- 50 stants. 0 10 20 30 40 SO M 70 80 90 100 Clearly the construction of a radial K/Nnux (70 machine allows a design with b\, . =0, in Umin The one-stage turbine-pump with controlled which case Q.... can be changed from 0% to discharge by changing the height of the runner and stay ring (N-U3 MW. H-450 m, N-750 rpm, 100%. The formula for Q1 for the pump D-1.80 m> mode differs from Equation 11, because in Neyrpic one-stage turbine-puop vith controlled discharge by distributor (N-3S MU. «>438 m. =b K-750 rpm'. D-1.70 m) pump mode V I'T^O and ^2 o" in the pump mode Fig. 5. Forecast topograms for one-stage TPCFR. (2) •(cos(e2)-tan(e2)) -b0 The drastic decrease of efficiency in Neyrpic turbine-pump for the powers smaller (2) than N can be explained easily. The re- where and (cos(e2)-tan(6 )) relate to the outlet edge in the pump mode, i.e., gulation of power in this machine is pro- vided by changing the vane angle in the dis- the inlet edge in the turbine mode. There- tributor. However, in the pump mode the fore in the pump mode Q_. is always propor- flow enters the distributor after passing tional to bQ, when it changes from b_. through the runner. Consequently, the change of vane angle cannot change the flow to £„_ . From Fig. 3 it is clear that for Umax character in the runner, but only increases a purely radial pump the design allows the head losses in the distributor. 324 vels. In the pump mode, both machines do have strong pulsations in the distributor, 100 since the flow after the runner is unsteady. H«Constane However in the conventional machine the Turbine Mode guide vanes are not rigidly fixed, since the upper and lower journals of the guide 80 vanes have to rotate freely in the bushings mounted to the upper and lower rings. 70 Therefore the pulsations in the distributor 60 of the conventional machine causes strong vibrations. This is the inherited disad- 50 vantage of the conventional turbine-pump "it 51$ SO 7i3 St 94 ffl which drastically increases with the in- crease of head. In the turbine-pump with controlled flow rate, the vanes are fixed 100 to the upper and lower rings (they are H-Constanc welded to these rings). Consequently the 1 (T.)90 Pump Mode pulsations of flow will not produce vibra- 80 tions, since the vanes and the upper and lower rings form a heavy rigid, monolith 70 structure. The pulsations cannot produce vibrations of the movable upper band either, «0 since they will be damped by this ring. Consequently, from this point of view, the 50 Is—ib—2for-~3fr—str—5b—6tr- rrt—at!—stn turbine-pump with controlled flow rate is more reliable. The weight calculations N/Nmox (7.) showed that the weight of the distributors of the conventional machine and TPCFR are approximately the same, (the new machine Two-atage turbine-pump with controlled discharge (N-96.MW, H-900 m, N»75O rpm. 0-1.80 n) does not have the gate ring with servomo- tors, but it has the upper movable band with servomotors). The runner with shaft Conventional two-atage turbine-pump (N-96 MH. H-900 m, U-750 ryn, D-1.80 m) for the new one-stage machine is approxi- mately 70% heavier than that of the conven- tional machine. However, TPCFR does not require the operating valve which might Fig. 6. Forecast topograms for two- weigh as much as four times the weight of stage TPCFR. the runner of the conventional machine. Consequently the conventional one-stage machine with the operating valve has to be RESULTS heavier than the one-stage TPCFR. It is clear that the cost of labor for the pro- The comparison of TPCFR with the con- duction of TPCFR is higher, therefore it is ventional machine gives the following re- difficult at this stage to predict which sults. machine will be more expensive. However, even if the cost of the new pump is accept- In the turbine mode, both the conven- ed to be higher than that of the convention- tional turbine-pump and TPCFR have approx- al pump by 20%, it is easy to show that the imately the same, ability to regulate the new pump is economically better than the power with high efficiencies as can be seen conventional for storage of solar electrical in Fig. 5. In the pump mode, although plant energy, because of its ability to re- TPCFR can effectively regulate the power, gulate the flow rate in the pump mode. In the conventional turbine-pump can only op- order to participate during peak time in erate at one point with high efficiency or the utility grid with certain power, the regulate the power only by drastically de- solar electrical plant with pumped storage creasing the efficiency' (Neyrpic scheme) as has to store a certain amount of energy seen in Fig. 5. The cavitation factor for during sunny days. TPCFR is approximately the same as that of the conventional machine since both are From Fig. 7 it is clear that new pumps radial-axial machines. From the pulsations can store energy of the solar plant contin- point of view, in the turbine mode, both uously without waste, but the conventional machines will have the same pulsation le- pumps working by steps inevitably waste en- 325 ergy. This waste causes the increase of of 1109 MW, the ratio of sunny days in this installed power in the solar electrical region is 80%, the efficiency of storage is plant in order to accomplish the same re- TI =0.8. The typical graph of insolation quirements of energy. for the latitude 25°N was used for compar- ison (Fig. 7). The result of the calcula- tions is that the solar plant equipped with two new pumps in combination with three conventional pumps (in order to have con- tinuous regulation, two of the new pumps were sufficient) will have the installed power of 500 MW. The solar plant equipped with five conventional machines after opti- mization has to have an installed power of 552 MW. If the cost of 1 kWe of solar e- lectrical plant is taken as $2000 and the cost of the conventional pump is $100/kWe, the cost of turbines is $100/kWe and the cost of construction is $100/kWe, the solar plant with storage equipped only with con- ventional machines will cost $1242.5/kWe. The cost of the solar plant with storage utilizing two new machines will be $1150.4 /kWe. Consequently the scheme with new ma- chines- costs 8% less in this particular S 9 10 11 i2 13 !'• 15 If. t fh> case. a. Storage Plane vlth Tuo I'uaps The analysis of feasibility and pros- wltb Controlled Flow Fate pects leads to the following conclusions. ? (MW) It is feasible to use the turbine-pump Output of Solar Electrical riant . with controlled flow rate at the high head conventional pumped-storage plants in com- bination with conventional machines. Two or three machines with controlled flow rate will provide the pumped-storage plant with the ability to regulate the power continu- ously and with high efficiency in both pump and turbine modes. The most attractive application of the new machine is the utilization of pump with controlled flow rate in storage systems of solar-electrical plants, wind-electrical plants, etc., since this pump eliminates the energy losses which are inevitable dur- ing operation of these plants. It is appropriate to begin the labora- (9 Iff 11 12 13/ 14 iy\ 16 t (h) tory experiments of the turbine-pump with >tj .J tj tij controlled flow rate, since the hydraulic b. ScoraRe Plant with Conventional Punpa analysis and conceptual design confirmed the positive features of this machine. Fig. 7. Pumped storage of solar elec- There is an immediate demand of solar trical plant energy. and wind electrical plant storage systems for rural areas of the U.S.A. In order to In the following comparison, these satisfy this urgent demand it is possible assumptions were accepted. to start the technical design and produc- tion of small machines of this new type. The solar electrical plant has to work The power of these machines should be no two hours during the peak time with power more than 1 MW. and the diameters no more 326 than lm. These machines will successfully operate under a head around 100m. ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial support of the Department of Energy and the continuous support and guid- ance of Dr. George C. Chang, Chief, Ad- vanced Physical Methods Branch, Division of Energy Storage Systems, D.O.E., during the accomplishment of the project. The authors also express their grati- tudes to Dr. George Kartsounes, Dr. Carl Bloomquist and Dr. Shiu-Wing Tarn of Ar- gonne National Laboratory for their contin- uous leadership and assistance during the project and especially the preparation of the final report, and to Dr. Selim Chacour of Allis-Chalmers, Hydroturbine Division, for his valuable suggestions concerning the design of the one and two-stage ma- chines. REFERENCES 1. Radial-axial Hydraulic Turbine with Double Governing, A. Gokhman, U.S. Patent No. 3 240 469, March 15, 1966. 2. Multi-Stage Turbine-Pump with Govern- ing, A. Gokhman, Patent Application No. 618 964, October 2, 1975. 3. Differential and Integral Calculus, R. Courant, Interscience Publishers, New York, 1964. 327 SESSION VH: COMPRESSED AIR ENERGY STORAGE 329 PROJECT SUMMARY Project Title: Compressed Air Energy Storage--Reservoir-Stability Criteria for Temperature and Pressure Cycling Principal Investigator: W. V. Loscutoff, J. A. Stottlemyre Organization: Pacific Njrthwest Laboratory PO Box 999 Richland, WA 99352 Telephone: (509) 946-2768 Project Objectives: Develop design and stability criteria for long-term operation of storage reservoirs subjected to temperature pressure and humidity fluctuations of a CAES olant. Project Status: This project is divided into the fell owing four phases: (1) Define preliminary design and stability criteria based on a comprehensive state-of-the-art survey; (2J Establish numerical models to study behavior of reservoirs and obtain additional design and stability criteria for reservoirs; (3) Perform laboratory experiments to obtain supplementary stability criteria; (4) Perform field tests to establish final design and stability criteria. The project status is as follows: (1) Preliminary design and stability criteria have been formulated for compressed air energy storaae in aquifers and salt cavities. Criteria for Hard rock caverns will be formulated by February 1979. (.2) Numerical analyses of the behavior of acquifers, salt cavities and hard rock caverns are proceeding according to the program clan. (3) Laboratory studies of aquifers are currently under way. A contract has been signed with Louisiana State University to perform laboratory studies of salt cavities. Proposals are being evaluated to select a contractor to perform laboratory studies of hard rock s-pectmens. (4) A program plan for field studies of salt cavities is being formulated. Program plans for field studies of aquifers and hard rock caverns are to be preoared during FY-1979. Contract Number: EY-76-C-06-1830 Contract Period: FY-1978, continuing Funding Level: $800,000 B0* Funding Source: Department of Energy, Division of Energy Storage Systems 331 •Included in this project are the following: L. W. Wiles, "Fluid Flow and Thermal Analysis for CAES in Porous Rock Reservoirs," page 337 J. R. Friley, T. J. Doherty, "Thermo Mechanical Stress Analysis of Porous Rock Reservoirs," page 349 J. A. Stottlemyre and R. P. Smith, "Potential Air/Water/Rock Interactions in a Porous Media CAES Reservoir," page 355 332 COMPRESSED AIR ENERGY STORAGE PROGRAM OVERVIEW Walter V. Loscutoff Pacific Northwest Laboratory Richland, Washington 99352 ABSTRACT The DOE compressed air energy storage (CAES) program consists of a group of inter- related studies directed at developing a new technology to improve the cost and efficiency of electrical power utilization. The program has two major thursts — Reservoir Stability Criteria Studies and Advanced Concepts Studies. The Reservoir Stability Criteria Studies are aimed at accelerating the near-term application of conventional systems. The Advanced Concepts Studies are directed at development of systems that require little or no firing of the turbines with, premium fuels. The Pacific Northwest Laboratory, as the lead laboratory in the CAES Program, assists DOE, Division of Energy Storage Systems in program planning, budgeting, contracting, managing, and reporting. In this overview, we summarize the CAES program, outline the specific tasks and identify the performers, indicate progress on current tasks, point out anticipated activ- ities, and examine significant milestones. INTRODUCTION considered include coal gasification, liquefaction, MHD systems, solar-thermal The Pacific Northwest Laboratory is augmentation, and nuclear waste heat of the lead laboratory for the Department of decay. Energy's Compressed Air Energy Storage (CAES) Technology Program. As such, it is In order to put the CAES Technology responsible for assisting the Department Program into the proper perspective, let of Energy in program responsibilities me briefly describe the overall DOE CAES for planning, budgeting, contracting, program. There are two distinct programs managing, reporting and disseminating in- on the CAES concept within DOE as shown formation. in Figure 1: the Demonstration Program, co-sponsored by EPRI and managed by three The CAES Technology Program is a series utility companies, and the Technology of interrelated studies being performed by Program with PNL as the lead laboratory. national laboratories, universities, and The Demonstration Program seeks to estab- industry to improve the cost and efficiency lish the feasibility and availability of of electrical power utilization. It has two CAES to a utility, leading to a demonstra- major thrusts — CAES Reservoir Stability tion plant. The Technology Program Criteria Studies and Advanced CAES Concepts objectives are to accelerate the commer- Studies. cialization of CAES and explore advance- ments that will make CAES totally indepen- The CAES Reservoir Stability Criteria dent of any premium fuels such as natural Studies are directed at accelerating near- gas and oil. It is responsive to the term application of conventional CAES questions raised by the Demonstration concepts by the utilities industry. Its Program and works with the Demonstration major tasks are to establish design and Program to maximize the efforts of the stability criteria for reservoirs to be two. used by CAES plants. The CAES Advanced Concepts Studies are directed at developing CAES TECHNOLOGY PROGRAM FOR FY-1978 advanced CAES systems for the future that will require little or no firing of the At this meeting, we shall present turbines with premium fuels. The efforts reports on progress in major tasks under- are concentrated upon CAES systems with taken during FY-1978. Because of space thermal energy storage (CAES/TES) and on and time limitations, several tasks will CAES systems with coal-fired fluidized bed not be presented here either because they combustion (CAES/FBC). Other concepts were relatively minor or because they have 333 CAES RESERVOIR STABILITY PROGRAM OUTLINE DOE CAES EPRI CAES PROQRAM PROGRAM TECHNOLOGY DEMONSTRATION MSS PSI PEPCO Fig. 1. CAES Technology and Demonstration Programs been initiated only a short while ago. The most important in the latter category are the tasks established to perform field studies in a salt cavity, to study the technical and economic feasibility of coal fired fluidized bed systems integrated with Fig. 2. CAES Reservoir Stability Criteria CAES, and to perform laboratory studies of Studies salt specimens under simulated CAES condi- tions. These tasks have been initiated in porous rock via state-of-the-art and we expect activity to begin during survey (PNL-Stottlemyre); FY-1979. The other major tasks will be presented by the principals involved. In Develop preliminary design and this overview, I would like to outline the stability criteria for air storage relationship that exists between the various in salt cavities via state-of-the-art tasks. survey (LSU-Thoms and Martinez): . Initiate study to develop preliminary A. Reservoir Stability Criteria Studies design and stability criteria for air storage in hard rock caverns The objective of this program is to (RE/Spec-Gnirk) establish design and stability criteria for CAES reservoirs. The program outlined . Start numerical studies of behavior in Figure 2 consists of four phases: of salt cavities subject to CAES conditions (SGI-Serata); . state-of-the-art studies . Perform numerical studies of the . numerical modeling behavior of porous rock reservoirs (PNL) . laboratory studies . Start numerical studies on behavior . field studies of hard rock caverns subject to CAES The three types of air storage reser- conditions (RE/Spec-Gnirk); voirs being considered for CAES are: . Perform laboratory studies on effects of CAES conditions on porous rock . salt cavities (PNL and Wisconsin-Pincus) porous rock reservoirs . Start laboratory studies on effects of CAES conditions on salt cavities . hard rock caverns (LSU-Thoms) The past year has seen a number of milestones reached in terms of accomplish- . Develop draft program plan for field ments and new starts. These include the tests in salt domes (PNL-McSpadden); following items: A summary of the foregoing items is shown in the milestone/accomplishment chart Develop preliminary design and shown in Table 1 together with anticipated stability criteria for air storage 334 Table 1. CAES Technology Program Mile- reservoir. This heat would then be re- stones/Accomplishments turned to the air before it is expanded in a turbine, thus eliminating any need for fuels. Another approach is to look —— . Fiscal *ear for other fuels to provide the required 1978 1980 heating of the air. One attractive con- *. Reservoir Design and Stability Crttert* 1. Preliminary design and stability cept is coal-fired fluidized bed combus- criteria (state-of-the-art-survey) a. Aquifers tion integrated with a CAES plant. These b. Salt cavities various options are outlined in Figure 3. c. Hard rock caverns 2. NuMrical modeling ». Aquifers b. Salt cavities c. Hard rock caverns 3. Laboratory studies ». fajuifers b. Salt cavities c. Hard rock caverns 4. Full scale field testing t a a. Aquifers b. Salt cavities rip V e. Hard rock caverns as B © complete test program plan 135 Future major activities in the Advanced Department of Energy, on risk insur- Concepts Studies project include the ance for CAES storage reservoir degra- following: dations. . Initiate technical and economic A joint study by the Central Elec- feasibility analysis of advanced CAES tricity Generating Board (CEGB) of for a utility; England and EPRI is looking at means of improving the performance of com- . Complete evaluation of technical and pressed air storage schemes by the economical feasibility of advanced use of thermal energy storage (TES). CAES equipment; Two approaches are being examined. In one approach, adiabatic storage, . Initiate demonstration program for the TES is used to eliminate any advanced CAES. need for fuel. In the other approach hybrid systems, TES is used to reduce RELATED CAES ACTIVITIES the premium fuel consumption by the turbines. This study is more compre- In addition to the DOE activities hensive arid complements the studies discussed above, there are several other funded by the Department of Energy. on-going projects in CAES. A brief summary of the major projects is given below. . The CAES plant at Huntorf was commis- sioned in August, 1978, and commercial operation of the plant began in October. Since the basic features of this plant, owned by Northwest deutsche Kraftwerke of West Germany,, have been described else- where, only highlights of the plant are given here. This is a 290 MW peaking plant with a two hour discharge and an eight hour charge time. The air is stored in two salt cavities having a total volume of 300,000 cubic meters with pressures up to 70 bars. The heat rate of this plant is 5500 Btu/ KWh. Plant construction began in May 1975. . The Kansas Utility Group, consisting of six private utilities, has com- pleted a statement of work to EPRI for a Phase-2 study that will look at the economics of CAES in Kansas. The study will examine a detailed load distribution model of a composite system for the utilities and evaluate the available options that include combustion turbines, small coal-fired units, and CAES. This study follows the Phase-1 study, performed by Black and Veatci . which was a geo- logic assessment of CAES in Kansas. If CAES is found to be economically feasible, Phase-2 is expected to be followed with preliminary CAES plant design. . The International Research and Technology Corporation, assisted by Dames and Moore, is performing a study, under a contract with the 336 FLUID FLOW AND THERMAL ANALYSIS FOR CAES IN POROUS ROCK RESERVOIRS L. E. Wiles Co-Author C. A. Oster Pacific Northwest Laboratory Richland, Washington 99352 ABSTRACT The analysis described in this report is a continuation of work initiated at PNL in FY-1977. A computer code was developed in FY-1977 to define the hydrodynamic and thermo- dynamic response to simulated mass cycling in a CAES dry porous media reservoir. The code was based on one-dimensional radial transport of mass and energy. The capability of the code has been extended in FY-1978 to include a one-dimensional, radial flow analysis of the effects of vapor and liquid phase water on reservoir performance. Parametric analysis included consideration of a range of injection temperatures, injec- tion humidities, and residual water levels. Potential consequences of the presence of water related to deliverability, thermal energy recovery, working fluid recovery, and storage volume are evaluated. Also, two-dimensional modelling has been developed for dry reservoirs to include the effects of gravity, vertical heat losses and the effects of stratified permeability. INTRODUCTION terized by the one-dimensional modelling, the delineation of the influences of the The analysis of the hydrodynamic and vertical boundaries, potential stratifica- thermodynamic response to mass cycling in tion of permeability, and gravity effects CAES porous rock reservoirs is intended to requires two-dimensional modelling. Code provide design guidelines for the efficient development and analysis to this end was and stable operation of the air storage also done at PNL in FY-1978. reservoir. The performance of the reser- voir depends on reservoir material proper- Summaries of these two efforts are pre- ties, reservoir geometry, and operating sented in this paper, which is divided into conditions. The influence of many of the two sections. The first section deals with important parameters was investigated at water in the reservoir. The second section PNL in FY-1977. That analysis made use covers the development of the two dimen- of a computer cede based on one-dimensional, sional code. radial transport of mass and energy in a dry porous rock reservoir.1 I. THE EFFECTS OF WATER ON RESERVOIR PERFORMANCE The capability of the code was extended in FY-1978 to permit the analysis of the THEORETICAL DEVELOPMENT effects of water on reservoir performance. Potential problems or areas of concern re- BASIC ASSUMPTIONS lated to the presence of water in the reservoir that could be evaluated with a In general, the hydrodynamic and one-dimensional model were identified. thermodynamic behavior of a CAES porous These include deliverability, thermal rock reservoir can be characterized by energy recovery, working fluid recovery a set of conservation equations that and storage volume. It was the objective describe the flow of an air-vapor mixture of this study to quantify the magnitudes and liquid water through a rigid porous of these effects. This was done by evalu- material. A number of assumptions must ating the influence of injection tempera- be made such that the resultant set of ture, injection humidity and residual water equations does not include terms irrele- content on reservoir performance. vant to CAES in porous rock reservoirs. These assumptions are: While the performance of the bulk of the reservoir and the influence of indivi- . the air-vapor mixture is homogeneous dual parameters can be reasonably charac- and behaves as an ideal gas 337 - . the rock porosity, density, and heat Conservation of energy in the capacity are constant; reservoir . the rock is immobile; . binary diffusion in the air-vapor mixture is negligible; . inertial effects are negligible; A definition of nomenclature is given in . the air-vapor mixture, liquid, and Table 1. rock are in thermal equilibrium; Table 1. Nomenclature . kinetic energy is negligible; and c heat capacity . viscous dissipation of energy is . the liquid water is immobile; and GREEK » porosity . Darcy's law can be used to write the L density fluid velocity in terms of the pres- U viscosity sure gradient to eliminate the percent of mass cycles momentum equation. SUBSCRIPTS With these assumptions the governing a air equations reduce to the following: s. liquid water m air-vapor mixture . Conservation of mass in the reservoir r radial coordinote s solid V water vapor z vertical coordinate . Conservation of mass for the vapor BOUNDARY CONDITIONS component The reservoir geometry adapted to this problem is shown in Fig. 1 where the •3 3P» . (2) region of influence of a single well is 3t vf r approximated as a cylindrical disc. At the well boundary the mass flow . Conservation of mass for the liquid rate is continuously specified. During component the reservoir charging cycle the temper- 1 3 ature and humidity are specified. During SP ) = -in ; and ( ) reservoir discharge, or when the reservoir is closed, the temperature and humidity at the well are obtained from local equilibrium conditions. 338 Table 2. Reservoir Parameters RADIAL FLOW OUTER IN POROUS MEDIA BOUNDARY Parameter Reference Value Geometry Reservoir Diameter 400 ft (122 m) Nell Diameter 7 In (18 cm) Properties Porosity 20* Permeability 500 md Rock Thermal Conductivity Operating Conditions Nominal Pressure 50 atm (S070 kPa) SINGLE WELL RESERVOIR Mass Flow Rate DIAMETER Injection "•"'seCTt <°1 Fig. 1. Geometry of the 1-D Reservoir Extraction <>•»<> Ie£t Model Initial Reservoir Temperature !00°F (38°C) The symmetrical arrangement of adjacent wells, which would be charged Table 3. Moisture Parameters and discharged at about equal rates, sug- gests that the outer boundary of the single well reservoir be insulated to Reference Value Range of Values the transport of mass and energy. Injection Temperature 450vF 100-450 (232°C) (38-232) Residual Water as Percent 205 0-40 SOLUTION of Pore Volume Injection Humidity 0 (dry air) dry air - saturated A detailed outline of the solution air at the Injection temperature and method of the governing equations is 8 provided in the FY-1978 Progress Report pressure to be published by PNL. The Injection humidity is limited by the maximum water vipor available In atmospheric air. This was considered to be 0.0429, the units being mass of water per unit mass of dry air; i.e., the saturation condition of lOOOF (38°C) air at atmospheric pressure. This limit applied to injection temperatures ANALYSIS OF THE MOISTURE PARAMETERS aboie 276°F (I36°C). The parameters that were analyzed with per foot (per meter) is used to imply the respect to dehydration and thermal develop- unit of vertical measurement. ment in the air storage zone are: A complete specification of operating . injection air temperature; conditions includes a schedule for the injection and extraction of mass to and . injection air humidity; and from the reservoir. To develop this schedule the following conditions were . residual water content. applied to the reference reservoir: The foundation for the analysis of . 40% of the cycled mass is stored in these parameters is a set of reference the reservoir over the weekend; values for the reservoir geometry, material properties and operating conditions. The . the weekdays are characterized by values given to these parameters are shown alternating injection and extrac- in Table 2. These values were fixed through- tion periods of 10-hours each, out this analysis. separated by 2-hour intervals when the reservoir is closed; The reference values and the range of values investigated for the moisture para- . the time averaged reservoir meters is given in Table 3. pressure is the discovery pres- sure; and The analysis was based on a reservoir having unit vertical thickness. Thus, . the percent of mass that is wherever appropriate, the specification of cycled is 14.8%. 339 The reservoir pressure is the spatial sultant condensation. Thus, a dehydrated average pressure. The percent of mass region surrounding the well will grow during that is cycled is calculated from mass injection. During reservoir discharge the mixture will gain heat as it nears M the well, thus, gaining moisture as it A _ max (5) moves radially inward. Net dehydration of the reservoir occurs by the eventual extrac- tion of this moisture. For these conditions, the weekly variation of reservoir pressure for the reference The process of dehydration for various reservoir is shown in Fig. 2. The discovery radial locations is shown in Fig. 3. The pressure was assumed to be 735 psi (5070 model predicts that the percent of pore kPa). volume filled with liquid water, which is initially uniform at 20%, never exceeds 25%. mm \ »m This result for the reference reservoir IMitlU satisfies the assumption of zero liquid mobility which required that the water / A content should never greatly exceed the residual level. More importantly, the model suggests that for dry air injection, / pore plugging should not adversely affect the deliverability. mm ;/ v 1 mmSul V\ m 25 l\l\f\«sa, \/ \ u \ j MM \ A | MIDWAY | SIMMY | iBHDAY | IIESMr |MgwsMr|iMMSMr 1 am* \ 47.frfl-RADlUS 1IK0FMDC.** > Fig. 2. Weekly Cycle of Reservoir Pressure, £ 15 - Beginning Friday at 8:00 p.m. * Changes to the moisture parameters will affect the dehydration rates and the rate 10 - - 3&2-ft-RA0IUS of thermal development of the reservoir. 1(1L6 ID) However, over the range of the values con- I \ sidered the characteristic behavior of the 2&.5fl-RADIUS processes of dehydration and thermal develop- 5 - 1(8.69 ml ment are identical. Thus, the description of these processes, as interpretted from i the results obtained from the computer model, are presented in this section. These 1 1 i \ descriptions will be applicable, in general, 0 12 3 4 5 to all of the results of the parametric TIME, yrs study of the moisture parameters. Fig. 3. Dehydration for Three Locations (5-Yr History) Dehydration Evaporation will occur near the well When heated dry air is injected into until complete dehydration of this region the reservoir it will exchange heat with is achieved. Gradually, the dry region the surrounding material by sensible heat- expands. The model implies that a sharp ing of the rock and liquid water and by interface exist between the dehydrated evaporating some of the water. The end zone and the region still containing result is that the air-vapor mixture will liquid water. The "dry front" is a term be saturated and will be in thermal equili- used to define the location of this hypo- brium with the rock and liquid water. When thetical interface. When the water content the air-vapor mixture moves radially out- reaches zero for a given radial location ward it will continue to loose heat to the the dry front is defined to exist at that surroundings. The mixture will maintain a location. By this definition the growth state of saturation by virtue of the re- of the dry front is shown in Fig. 4. 340 n H r. MOPM TIKOFKSK. On Fig. 4. Radius of the Dry Front for the Fig. 5. Thermal Cycling During the First Reference Reservoir Week of Reservoir Operation By the process of dehydration, water 500 may be removed from the well bore region (260) although it is not necessarily removed MAXIMUM WEEKLY from the reservoir. After 5 years of EXTRACTION TEMPERATURE continuous reservoir operation the model shows that the total water removed from A the reservoir represents only-5.5% of the original mass of water. The net increase in reservoir storage volume is a mere 1.1%. a- 4oo MINIMUM WEEKLY Thus, the dehydration process is quite EXTRACTION TEMPERATURE slow and should not, in general, be expect- sr ed to significantly increase the potential storage volume. Thermal Cycling and Thermal Growth RESIDUAL WATER, When the simulation of reservoir PERCENT OF PORE VOLUME operation begins the reservoir is at a 0* uniform temperature of 100°F (38°C). Air is injected into the reservoir at 450 F 20% (232 C) and is extracted at the equilibrium temperature adjacent to the well. Figure 200 5 shows the temperature variations (93) occurring in the near wellbore region during 2 3 the first week of reservoir operation. TIME, yrs While large temperature swings occur at the well boundary, thermal cycling is observed Fig. 6. Effect of Thermal Growth on Extrac- to be nearly non-existent within 14.0 ft tion Temperatures (4.28 m) of the well center. The volumetric thermal capacity of the rock mass is large By examination of Fig. 6 it can be enough to contain the injected thermal concluded that a significant percentage of energy very near the well. Net heating of thermal energy injected into a reservoir the reservoir beyond 14.0 ft (4.28 m) can be recovered during reservoir discharge. occurs primarily by conduction. The average thermal energy recovery over the first 5 years of operation of the refer- The development of thermal cycling ence reservoir is about 81%. The dashed predicted by the model, is shown in Fig. 6, lines in Fig. 6 represent the result for a where the maximum and minimum weekly extrac- dry reservoir. The average thermal energy tion temperatures are plotted as functions recovery for the dry reservoir is about of time for a dry reservoir and for the 82%. This comparison suggests that the reference reservoir. A significant reduc- presence of residual water will have only tion in thermal cycling at the well occurs a small effect on temperature cycling arid in the first year of operation. Beyond one thermal growth. year the changes occur more gradually. 341 INJECTION TEMPERATURE This portion of the analysis estab- lishes tha difference in the dehydration rates associated with the spectrum of INJECTI ON TEMPERATURES, °F (°C) possible injection temperatures. Also, the coexistence of high temperature and liquid water can dramatically increase the potential for adverse geochemical reactions. Thus, it is important to quantify when such conditions might occur. A reasonable lower limit of injection temperature was considered to be 100°F (38OC). An upper limit of 450°F (232°C) was chosen as this represents the approxi- mate maximum outlet temperature that can be tolerated by currently available cen- trifugal compressors. In Fig. 7 the radius to the dry 100(381 front is plotted as a function of time for each injection temperature. 10 20 30 50 TIME, wks Fig. 8. Effect of Injection Temperature INJECTION TEMPERATURES, °F(°C» on the Net Reservoir Dehydration Rate front moves beyond the region of thermal cycling the dehydration rates slow as the heat necessary for evaporation is avail- able primarily by conduction. Another cause of the decrease in dehydration rate with time is that an increasing portion of the injected air-vapor mixture is stored within the expanding dry region. That portion of the air is subsequently extracted without having encountered any liquid water. INJECTION HUMIDITY Unless it is otherwise defined the reference to humidity in this report is 20 30 absolute humidity where the units are mass TIME, wks of water vapor per unit mass of dry air. Fig. 7. Effect of Injection Temperature The moisture content of the injected on Growth of the Dry Front air is limited by the humidity of the ambient air that is to be compressed. In Fig. 8 the rates of net reservoir Saturated air at 100°F (38°C) and 1 atm dehydration are shown. A time dependent (101 kPa) of pressure will have an ab- decrease in the dehydration rate, which is solute humidity of 0.0429. For a pressure most apparent for high temperature injec- of 50 atm (5070 kPa) an absolute humidity tion, characterizes the results. During of 0.0429 can exist only if the tempera- initial operation the dehydration in the ture of the mixture is above 276°F (136°C). near well region is influenced by the Below this temperature the humidity of the periodically high temperatures present injected air-vapor mixture cannot exceed due to thermal cycling. When the dry the saturation conditions dictated by 342 the injection temperature and the reser- 200 voir pressure. This is shown in Fig.9. (298) -______^^ 0.06 DRY AIR INJECTION 5 0 if if -200 _ ^______£ 0.04 - (-298) ^ INJECTION HUMIDITY - 0.0429 LIMIT OF 0.0429, MI t SATURATED AIR AT -400 1 ATM (101 kPa) AND RISER ! (-595) { in 0.02 - 100°F(38°C) i i i I 1 C 10 20 30 40 50 TIME, weeks Fig . 11 Effect of Injection Humidity on Dehydration 100 200 300 400 500 (38) (93) (149) 1204) (260) tion rates shown in Fig. 11 suggest that TEMPERATURE, °F(°C) the injection of moist air into the reser- voir can result in the increase of the mass of liquid water in the reservoir, the Fig. 9. Saturated Absolute Humidity at region near the well bore can still be 50 atm (5070 kPa) dehydrated. The progression of the dry front with time is shown in Fig. 12. The effect of injected moisture on Although the dry fronts appear to move the dehydration at a radius of 14.0 ft outward rapidly, conservative reservoir (4.27 m) is shown in Fig. 10. The dehy- design will likely dictate that considerable dration is obviously slowed by the inclu- effort will be given to minimizing the sion of humidity in the injected air stream. injection humidity. Also, local pore volume water content can significantly exceed the residual levels, which indicates that pore plugging may be a potential problem. 60 - 1 40 - 50 INJECTION \ HUMIDITY 20 Fig. 12. Effect of Injection Humidity on \ 0.0 0.0429 \ Growth of the Dry Front Vl\ 1 I 1 \ The results obtained from the model 10 20 30 40 50 suggest that for injection humidities above those for which the net dehydra- TIME, wks tion is negative, the temperature at the dry front is characterized by the dew Fig. 10. Effect of Injection Humidity on point temperature of the injected air- the Dehydration of the Region vapor mixtu1"5. This is an important ob- Centered at a Radius of 14.0 ft servation because, if the existence of (4.27 m) liquid water at this temperature was known to produce adverse geochemical Reservoir dehydration rates are shown reactions, then some dehydration of in Fig. 11. Although the negative dehydra- the injected air stream would be necessary. 343 The effect of injected moisture on thermal growth in the reservoir is shown in Fig. 13. When dry air enters the reservoir the subsequent evaporation will inhibit the thermal growth. In RESIDUAL WATER, contrast, the injection of moist air PERCENT Of PORE VOLUME results in a net increase of liquid water in the reservoir due to conden- sation, and the thermal growth will be enhanced because of the deposition of latent heat. 50 io 20 30 40 50 TIME, wfcs Fig. 14. Effect of Residual Water on Dehydration Rate 30 (9) RESIDUAL WATER, • PERCENT OF PORE VOLUME^^—-—• 20 20* ^^-"^^ . . — —~~' (61 e IUS . 10 •SCKVOIItllUIIJS.lt M (31 A' Fig. 13. Effect of Injection Humidity on RAD I / Thermal Growth 0 1 1 1 1 1 RESIDUAL WATER CONTENT 20 30 40 50 TIME, wks For reasonable values of residual Fig. 15. Effect of Residual Water on water content the thermal capacity of the Growth of the Dry Front mass of water is not so large that it can significantly affect the thermal behavior CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER and dehydration in the reservoir. Its STUDY greatest impact would be to reduce the storage volume. To obtain given mass flow The results of the computer analysis rates to a given size reservoir, a greater provide insight as to how the presence of percent of pore volume initially filled residual water in an aquifer CAES reservoir with water will result in proportionally may affect the performance of the reservoir. higher pressure cycling. More importantly, perhaps, the results can be used to suggest how the presence of the The primary effect of residual water residual water will affect the areas of on the dehydration can be deduced by con- concern outlined in the introduction. sidering Figs. 14 and 15. Although the thermal development occurs at nearly iden- Deliverability. It appears that high tem- tical rates, the reservoir having more perature injection may not be necessary in residual water has a higher dehydration order to avoid reduced deliverability due rate. Because the dry front moves radially to reduced relative permeability in the outward at a slower pace for the case hav- vicinity of the wellbore. For a broad ing more residual water, the dry front is range of conditions the computer model in a region of the reservoir where the predicts that comparatively rapid dehy- temperature is higher. Since the tempera- dration of the wellbore region is achieved. ture at the dry front is greater, the extracted humidity will be greater. Thus, Injection humidities may prove the net reservoir dehydration rate is to be very important with regard to the greater even though the temperature of the integrity of reservoir deliverability. extracted mixtures would be about the same. Calculations show that pore plugging can 344 result. Although a liquid mobility BASIC ASSUMPTIONS model was not included in the analysis the potential plugging problem is prob- The basic assumptions applicable ably real. Conservative reservoir design to this problem are given in Section I would, therefore, dictate that relatively under the same heading. dry air be injected to the reservoir to decrease the potential for pore plugging. REDUCED EQUATIONS Thermal Energy Recovery. In general, the The additional assumptions applied to thermal capacity of the residual liquid this problem include: water in an aquifer CAES reservoir should be small comparad to the total thermal . the transport of mass and energy occurs capacity of the rock in the reservoir. in the radial and vertical directions; Thus, variations in the residual water con- tent will not have a large effect upon the . the reservoir and the air are dry; overall thermal development of the reservoir. and Thermal energy recovery will be affected by the humidity of the injected air stream, . Darcy's law applies. although the effect will be small. With these assumptions the governing Working Fluid Recovery. The injection of equations reduce to the following: dry air resulting in net evaporation will increase the working fluid available in Conservation of mass the reservoir. Net condensation of mois- ture injected into the reservoir results kr in a loss in the amount of working fluid. Thus, it may be necessary to weigh the importance of thermal energy recovery against working fluid recovery. However, (6) it is suggested by the model that there is relatively little to be gained or lost in Conservation of energy the balance between the two. It would seem to be more important to consider the potential pore plugging problem. s s Storage Volume. The presence of residual 1 3 / u water in the pore structure of the reser- F 3? (paha voir rock reduces the potential storage ••- 3r' 3z PaV" volume. The process of dehydration should , 1 3 not, in general, be expected to signifi- (7) cantly improve this situation. In fact, r zr 3? if the injected air is not relatively dry, the deposition of moisture in the reservoir BOUNDARY CONDITIONS could reduce the storage volume. The geometry adapted to this two- Further Study. These conclusions regarding dimensional analysis, is shown in Fig. 16. the potential problems associated with The radial boundary conditions are similar residual water are subject to the assumptions to those specified for the one-dimensional made in the numerical model. The two most analysis. At the well boundary, however, limiting assumptions of the model are zero pressures are specified that were computed liquid mobility and instantaneous saturation by the one-dimensional dry model of FY-1977 of the air-vapor mixture. Adjustments of to provide approximately constant mass the governing equations can be made to flow rates. include liquid mobility. Improvements to the prediction of the rate of evapora- The boundaries between the porous tion will require inputs from an experi- zone and the caprock and basement rock are mental program scheduled for FY-1979. closed to mass transfer. Heat is lost to these regions by conduction. II. TWO-DIMENSIONAL ANALYSIS OF DRY RESERVOIRS SOLUTION THEORETICAL DEVELOPMENT Details of the solution method are 345 GEOMETRY Table 4. Additional Reference Conditions Applied to the 2-0 Analysis WELLBORE DIAMETER Parameter Reference Value Geometry 3^ OVER- =£3 Caprock Thickness 37.5 ft {11 m) Storage Zone Thickness ]00 ft (30 m) Basement Rock Thickness 50 ft (15 m) Properties (Basement Rock and Caprock) CAPROCK Permeability 0 Porosity 0 Operating Conditions POROUS ROCK Injection Temperature 450°F (232°C) Well Pressures Computed by code developed in FY-1977 to approximate flow rates given in Table 1 BASEMENT ROCK -RESERVOIR DIAMETER' Fig. 16. Geometry of the 2-D Reservoir Model l.OP-ll- RADIUS available in the FY-1978 Progress Report to be published by PNL. ANALYSIS OF VERTICAL EFFECTS 5.38-H- RADIUS i ]}-lt RADIUS 1 B0-I1-RADIUS \ \ The intent of this analysis is to quantify the influence of vertical bound- aries, stratification of permeability, and S gravity on the performance of a dry 300 400 reservoir. 11491 TOI ItMPlR'T^t. "F I°CI The reference conditions given in Table 2 were used in this analysis. Fig. 17.Temperature Profiles in the Those conditions apply to the storage Homogeneous Reservoir zone. Additional parameters are given in Table 4. Material properties such as week of simulated reservoir operation. permeability and thermal conductivity were The temperature profiles indicate that assumed to be isotropic. The weekly very little thermal energy is lost to the charge and discharge cycle as described in caprock or basement rock. Since there is Section 1, was applied. The overburden no convection in the caprock or basement was assumed to be a thermal insulator. rock then the model dictates that thermal The geothermal gradient was taken as energy transport depends entirely on con- 8 duction, which is extremely slow. The 0.02°F/ft (0.12 C/m). result also indicates that buoyancy effects are not significant in the homogeneous ANALYSIS OF THE REFERENCE RESERVOIR reservoir. Heating of the rock occurs somewhat uniformly in the radial direction. The analysis of the reference reser- Thus, there are no significant vertical voir permits an evaluation of the influence temperature differentials to create a of the vertical boundaries and gravity on noticable buoyancy component. the performance of the reservoir. This analysis can also be used to judge the usefulness of the one-dimensional modeling. The thermal development and pressure cycling in the bulk of the reservoir agrees Temperature profiles are shown in with that predicted by the one-dimensional Fig. 17 for a homobeneous, dry reservoir. model. Thus, the one-dimensional model can This result was obtained after completion be used to characterize the performance of of the final charging cycle of the first the bulk of a homogeneous reservoir with equal accuracy but at a much lower cost. 346 STRATIFIED PERMEABILITY tures will follow successful completion of that effort. A horizontal layer (i.e., a cylindri- cal disc) having a permeability of 50 md In addition, temperature and pressure was defined to exist between zones having profiles computed by the two-dimensional permeabilities of 500 md. The layer was code are necessary for evaluation of the 15 ft (4.6 m) thick and was centered half- potential for failure of reservoir inte- way between the caprock and basement rock. grity due to thermo-mechanical stress behavior. Thus, data from the code will In Fig. 18 vertical temperature pro- be used as input to the stress analysis. files are shown after completion of the final charging cycle of the first week of REFERENCES simulated reservoir operation. The results show that thermal development of the low 1. G.C. Smith, J.A. Stottlemyre, L.E. permeability zone is significantly dimin- Wiles, W.V. Loscutoff, H.J. Pincus, ished. However, comparison of Figs. 17 FY-1977 Progress Report: Stability and 18 show that thermal development of and Design Criteria Studies for Com- the 500 md zones, and the caprock and base- pressed Air Energy Storage Reservoirs. ment rock is not noticeably affected by PNL-2443, Battelle, Pacific North- the presence of the low permeability layer. west Laboratory, Richland, WA 99352, March 1978. i.oen-RADius V V. \ 5.»H»«OIUS^J ).1)HR»OIOS J I. W-n RADIUS-. J 1) Fig. 18. Temperature Profiles tn the Stratified Reservoir CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER STUDY A two-dimensional model was developed to analyze the performance of dry, nonhomo- geneous reservoirs. Initial analyses were performed to evaluate vertical heat losses and gravitational effects, both of whfch appear to be small. Also, the results show that the one-dimensional model developed at PNL in FY-1977 is an equally valid tool for describing the performance of the bulk of a homogeneous reservoir. The capability to analyze nonhomo- geneous reservoirs and other vertical transport effects has been demonstrated. Continued code development is planned for FY-1979 to improve the efficiency of the calculations. A parametric evaluation of possible nonhomogeneous reservoir struc- 347 THERMO MECHANICAL STRESS ANALYSIS OF POROUS ROCK RESERVOIRS J.R. Friley T.J. Doherty Pacific Northwest Laboratory PO Box 999 Richland, Washington 99352 ABSTRACT Stress levels in a generic porous rock compressed air energy storage site were investigated. Site loadings considered thermal, pore pressure, and in situ effects. Loading conditions simulated site operation for approximately one year of reservoir charging and discharging. Structural concerns focused primarily on tensile stress and fatigue stress magnitudes. In addition, the reservoir structural response due to tensile fracture of the cap rock region was investigated. For the conditions studied, fatigue stress levels were observed to be greatest in the porous material. These stresses decreased with age of reservoir operation and with distance from the well bore. Tensile stresses were observed to be greatest in the cap rock. While the tensile stress magnitude will most likely depend on site dependent overburden characteristics, mechanisms thought to cause this tensile behavior were postulated. While lack of site dependent information precluded interpretation of localized fatigue and fracture stress levels, it is felt that several salient features of cap rock and porous rock structural behavior in a porous rock reservoir were illustrated. INTRODUCTION Structural behavior of a porous rock - •?.••« reservoir subjected to loading conditions of compressed air energy storage operation 0 ° o. OVERBURDEN at elevated temperatures is, as yet, not • fully known. No such operational facility 0 exists today. In addition, scale model mCt CAP ROCK testing of such a concept has not been TYPICAL '.;ffi carried out at this time. In order to ; gain some insight regarding these unknowns ISOTHERMVT-;; f. l' POROUS ROCK and perhaps to aid in performing testing in a more enlightened fashion, structural ''SOT analyses with assumed site geometry and lf%f/ BASE ROCK material properties have been carried out. This paper discusses these initial investi- Fig. 1. Cross Sectional View of a Typical gations. Porous Rock Site BACKGROUND the porous body. During peak demand, air is drawn from the porous zone to produce Structural features of a porous rock power. reservoir can best be described with reference to Figure 1. Air is compressed Structural loading of the cap, porous, and forced down the well bore during and base rock masses is caused by several periods where excess generating capacity sources. In situ loading is caused by is not needed. Non permeable cap rock horizontal and vertical stress fields ori- and base rock provide for vertical con- ginally existing in the rock forms. The tainment of the air while radial contain- vertical stress level is approximately ment is provided by means of interstitial equal to the stress resulting from the water contained in the outer regions of overburden loading. The horizontal compo- nent depends largely on the tectonic acti- . the flexural stiffness of the cap vity of the site region. Both vertical and rock horizontal stress components are likely to be compressive. . the compliance of the overburden above the cap rock The structural behavior due to air pressure in the porous zone will depend on Loading conditions discussed thus far several factors. These factors include: are essentially independent of daily cyclic the rock porosity, the spatial dependency fluctuations of charging and discharging. of the pressure, and the micro elastic Cyclic stress levels induced by alternating characteristics of the porous rock fabric. periods of charge/discharge could have significant structural effects if fatigue Another significant structural loading stress limits of the materials involved method is that of non-uniform temperature are exceeded. distributions within the rock masses affected. Heat from the compressed air PROJECT DESCRIPTION flows primarily radially in the cap rock and the porous rock. In the cap rock, this In order to gain insight into struc- heat flow is by conduction. In the porous tural behavior, a generic site description material, however, heat is transported by was hypothized and analyzed numerically. conduction and by flow effects as well. As Since structural effects are likely to a result of this difference, thermal effects change from initial charging to steady can be expected to decay more rapidly in state operation, various times during the cap rock than in the porous rock, and the site operation were studied. These isothermal lines will appear/some what like times included initial start up through one a bottle as shown in Figure 1. year of simulated operation. The radial temperature gradient in The structural analysis was preceded both the porous and nonporous rock has a by an independent thermal/flow analysis pronounced effect on rock stresses. In discussed in Reference 1. The theory the cap rock, hot areas near the well bore on which the structural analysis was based location tend to expand and thus stretch the assumed that the thermal and flow responses cooler outer regions. This characteristic were uncoupled from the mechanical behavior. tends to induce compressive radial and loop This assumption tends to be valid for prob- stresses near the well bore. The cooler lems in which the pore structure is stiff outer rock resisting the expansion will when compared to the flow medium which, in incur tensile loop or circumferential this case, is air. Further discussion of stress levels. structural behavior of porous bodies sub- jected to a flow environment may be found In the porous rock, radial temperature in References 2, 3, and 4. gradients will tend to cause the same struc- tural behavior as in the cap rock. Since Both analysis phases of this study thermal gradients in the porous material are (the thermal/flow as well as the struc- less severe, however, the stress levels will tural) assumed axisymmetric models. The likely be less pronounced. geometry of this generic site is shown in Figure 2. In addition to radial expansion and compression, another thermal response mode The analysis procedure on which the exists. As the large central region of structural behavior was based made use of porous rock heats up, vertical expansion is the poro-elastic formulation^ and the attempted. As a result, upward loading on finite element method^. The elastic stress the central portion of the cap rock will strain relationship in this case takes occur. This will tend to produce flexure the following form when using indicia! or bending stress behavior. As a result, notation. tensile stress will form on the top of the cap rock and compressive stresses on the bottom. The magnitude of this flexural (1) stress trend will depend upon: (1-B)P+<*T (2) . the axial stiffness of the porous zone 350 •- MODEL AXIt Table 1. Rock Mechanical Properties too* Ml M> OVtMUNDEN /. " < Cap Rock UMDWia H H and Base Rock Porous Rock hull. 4.65 x 106 «.35 x 10* CAP E (psD HOCK 0.25 0.25 tin m 5.6 x lO"6 5.6 x lO"6 20% POROUS HOCK doom 1 burden loading of 562 psi which is equivalent to a depth of about 560 ft. Horizontal stress will occur due to the radial res- MtE ROCK traints at the outer portion of the model. 1 i AXIAL (VCftnCAU RCSTttAlttr These features are illustrated in Fig. 2. Fig. 2. Model Geometry and Finite Element RESULTS Mesh Used for Structural Analysis Due to space limitations, only struc- tural results corresponding to the first a.. stress tensor and last weeks of site simulation will be P pore pressure presented here. Figures 3 and 4 show f rock porosity isothermal plots of these two conditions. &•• 6. Kronecker delta These figures illustrate the radial growth of thermal effects with increased simulation elasticity tensor time. Pore pressure loadings for both of total strain tensor these cases were about 690 psig and quite hi uniform throughout the porous media with free strain a variation of only about 25 psig. v Poison's ratio E Young's modulus a coefficient of thermal expansion -«-250 T temperature bulk modulus ratio of porous rock to interpore material •160 (.25 assumed after ref. 1) -.1-260 Equilibrium equations take the form: 1 , = 0,0. (3) Where (, j) denotes partial defferentiation with respect to the j*" coordinate direc- tion. The above equations were formulated by using a modified version of the finite element computer code AX1S0L5. Fig. 3. Thermal Distribution After One Material properties used for the Week of Simulation (°F) structural analysis were chosen to repre- sent typical values of shale for the cap Plots of maximum tensile stress and and base rock forms and typical values for maximum shear stress for the two cases sandstone for the porous rock. Numerical considered are shown in Figures 5 though values used were taken from Reference 6 8. These results show the effect of the and are shown in Table 1. increased thermal zones in all three rock forms. In particular, the differences in Structural analyses were performed tensile stress results between week 1 and for various loading conditions of temperature week 53 indicates the degree of cap rock and pore pressure encountered during the flexure caused by vertical thermal growth first year of site simulation. In situ of the heated porous zone. effects were simulated by using an over- 351 Fig: 4. Thermal Distribution After 53 Fig. 7. Contour Plot of Maximum Shear Weeks of Simulation (°F) Stress (psi) After One Week of Simulation Fig. 5. Contour Plot of Maximum Tensile Stress (psi) After One Week of Fig. 8. Contour Plot of Maximum Shear Simulation Stress (psi) After 53 Weeks of Simulation .200 The magnitude of daily fatigue or cyclic stress level was investigated by 1400 1000 analyzing the model for stress differences at the minimum and maximum loading condi- tions occurring during the daily cycle. For these conditions, temperatures at the well bore were 250°F and 450°F. Correspond- ing input pressure values were 690 psig and 735 psig, respectively. Thermal response to daily cycling was observed to occur in the vicinity of the well bore. Pressure changes in the 400 200 porous zone, however, were somewhat uniform and followed the well bore pressure very closely. Fig. 6. Contour Plot of Maximum Tensile Stress (psi) After 53 Weeks of Fatigue stresses (shear stress) are Simulation shown in Figures 9 and 10. As can be 352 This incremental procedure was carried out for several time steps during the one year simulation. As a cracking threshold, 780 300 psi was assumed to be representative 1 of shale tensile strength. BOO Patterns of damaged rock resulting "—— __ ^ from the analysis procedure mentioned above R=12' R=3' are shown in Figures 11 and 12 for the 260 first and last week of site simulation, respectively. • i i i i 0 10 20 30 40 60 60 WEEKS - RESERVOIR AGE Fig. 9. Fatigue Stress Trends on the Bottom Region of the Cap Rock 1600 1000 - 600 - Fig. 11. Region Affected by Tensile Fracture (300 psi) After One 10 20 30 40 SO 60 Week of Simulation RESERVOIR TIME - (WEEKS OF AGE) Fig. 10. Fatigue Stress Trends in the Central Region of the Porous Rock seen, fatigue stresses tend to be greatest near the well bore and diminish considerably with increasing radius. In addition to this trend, it can be observed that fatigue stresses tend to be greatest during initial periods of operation and tend to diminish with time. The analyses discussed thus far assumed that material behavior was elastic; that is, no allowance of material cracking was considered. Rock is notorious for its lack of tensile strength and it is likely that the tensile stress shown in Fig. 6 will result in some localized cracking. Fig. 12. Region Affected by Tensile Frac- ture (300 psi) After 53 Weeks of In order to gain some insight into Simulation the effects of cracking, another sequence of analyses was performed. In this sequence, CONCLUSIONS temperature and pressure loading was applied and material stiffness in the direction of The stress results presented here maximum tensile stress was assumed to vanish were arrived at by analyses based upon if that stress exceeded some threshold several important assumptions. The degree value. These new material properties were to which these results depict reality then used with later load conditions. depends of course on the adequacy of 353 these assumptions. CRC Handbook of Tables for Applied Engineering Science, 2nd Edition, CRC Particularly important in this regard Press, Cleveland, OH, 1973. is the assumption dealing with cap rock shape and overburden stiffness. The assumptions dealing with flat cap rock and no overburden stiffness will most likely yield results which show large cap rock flexure when compared to a model which assumes a domed cap rock with stiff overburden. Thus, the severity of tensile stress levels and the degree of cracking is probably conservatively estimated by the analyses performed. The stress levels in the immediate vicinity of the well bore should be inter- preted with the knowledge that localized effects due to the well casing and its attachment to the rock forms were not modeled. Global structural trends of site behavior were the objective here, not localized effects. The consequence of fatigue stresses in the well bore vicinity will be influenced not only on the fatigue characteristics of intact rock samples but also on the extent of fracturing which occurs during the drilling process. REFERENCES 1. Wiles, Larry E., "Analysis of CAES Porous Rock Reservoirs", Proceedings of the First Annual DOE Contractors Review Meeting for Mechanical and Magnetic Energy Storage, Luray, VA, 1978. 2. Lubinski, Arthur, "The Theory of Elasticity for Porous Bodies Display- ing a Strong Pore structure", Proceedings of the Second U.S. Congress of Applied Mechanics, 1954. 3. Mordgren, R.P., "Strength of Well Completions", 18th U.S. Symposium on Rock Mechanics, Johnson Publishing Co., Boulder CO, 1977. 4. Stagg, K.G., and Zienkiewicz, O.C., Rock Mechanics in Engineering Prac- tice, John Wiley & Sons, N.Y., 1968. 5. Wilson, E.L. and Jones, R. M., "Finite Element Stress Analysis of Axisymmetric Solids with Orthotropic Temperature Dependent Material Properties" Air Force Report No. BSD-TR-67-222, September 1967. 354 POTENTIAL AIR/WATER/ROCK INTERACTIONS IN A POROUS MEDIA CAES RESERVOIR J.A. Stottlemyre and R.P. Smith Pacific Northwest Laboratory Richland, Washington 99352 ABSTRACT There appears to be motivation for storing elevated temperature (50-300°C) compressed air in underground porous media reservoirs. The feasibility of this concept depends to a major extent, on the potential physical and chemical response of the reservoir to any new environmental conditions imposed by Compressed Air Energy Storage (CAES) operations. For example, changes in absolute and relative permeabilities due to elevated and time varying temperatures, interstitial fluid pressures, humidity condi- tions and free oxygen would be of specific interest. This paper briefly discusses ten potential reservoir damage mechanisms: 1) dis- aggregation and particulate plugging, 2) thermo-mechanical plugging, 3) corrosion, 4) clay swelling and dispersion, 5) matrix consolidation, 6) water evaporation and mineral precipitation, 7) oxidation reactions, 8) hydrolytic reactions, 9) mineral solutioning and precipitation, and 10) fluids incompatibility. These topics are discussed in what is considered their order of importance with respect to the CAES concept. Fatigue failure of well casings and cement materials is also considered on exceedingly important item, but is not elaborated on in this paper. Finally, some explanation cf current and planned experimental and theoretical investigations is presented. INTRODUCTION In addressing how CAES operating conditions might alter the reservoir over The economic and technical feasibility 30 to 40 years, the following properties of Compressed Air Energy Storage (CAES) (dependent variables) should be stressed: technology depends to a major extent on 1) absolute and relative permeabilities, locating, developing,.and maintaining an 2.) effective porosities, 3) bulk compres- adequate subsurface reservoir. The manner sibilities, and 4) the corrosiveness of in which such a reservoir may respond, any interstitial fluids. In the study physically and chemically, to new environ- at PNL, a series of assumptions have been mental conditions associated with CAES made to facilitate a description of the operation is one of the fundamental unknowns system. These assumptions are: 1) all facing the implementation of this promis- phases will attain rapid thermal equilib- ing new technology. rium, 2) a heat pillar will be formed around each wellbore and will change only The operating conditions of primary slowly with time via thermal conduction interest include: maximum mean pressure and vapor transport, 3) a certain percen- which is controlled by reservoir depth tage of water (the residual content) will (0.11 bars/meter), 2) maximum temperature remain in the air storage zone subsequent which is apparently restricted to 300-350 C to initial bubble development, 4) the by standard well bore casing and grouting residual water will be gradually but material limits, 3) maximum loading and steadily displaced by evaporation with unloading rates which are as yet undefined, warm undersaturated air, 5) the residual and 4) maximum inlet air humidity and water will be immobile in the liquid phase, suspended solids concentrations. Other 6) there will exist a high impedance to independent variables, which become con- the flow of condensate back towards the stant once a site is selected, include: well bore, 7) the reservoir will be rapidly 1) the physical, chemical, and mineral- transformed from a reducing environment to ogical characteristics of the reservoir one abundant in free oxygen and carbon rocks, 2) groundwater constitutent concen- dioxide, and 8) reservoir heterogeneity trations, and 3) the physical and chemical and anisotropy will dictate that generalized characteristics of the confining strata. and simplified solutions should not be 3 (p 3k considered as absolute. This paper briefly discusses ten potential reservoir damage mechanisms. These mechanisms are ordered in terms of expected importance and where warranted, arguments are presented to support or reduce the applicability to CAES reservoirs. The potential damage mechanisms are pre- sented in three categories: Category I includes: 1) rock disaggregation and particulate plugging, 2) thermo-mechanical plugging, 3) corrosion, and 4) clay swel- ling and dispersion. Category II includes: 1) rock consolidation and subsidence, 2) residual water evaporation and mineral precipitation, 3) oxidation reactions, and 4) hydrolytic reactions. Category III includes: 1) mineral solutioning and precipitation and 2) fluids incompatibility. Figures 1 and 2 are presented to lend perspective to the discussion. The micro- graphs are of Berea sandstone from an outcrop in Ohio. •HE* MNOSTONI {30tfC.I2«lors.*73Hrll ' 0.3pm ' Fig. 2. Micrograph of Altered Berea Sandstone Particulate Plugging Particulate plugging refers to the formation of a "filter cake" of small micron size particles at the sandface or the formation of barriers to flow due to bridging of particles in channel restric- tions within the reservoir rock itself. Figure 3 conceptually portrays this bridg- ing phenomenon. There are a multitude of potential sources for such particles. The inlet air stream can contain suspended inorganic (dust) particles or micro- organisms. Any mineral scale or corrosion products on the inner walls of pipes or well casings can result in sloughed particles. The rock itself is a good potential source. Dislodged grains or fragments of grains, frayed clay plates, interstitial cement materials and precipitates deposited during evaporation of the residual water serve as examples. nit MNOSTONt One conclusion, based on very limited IUNW.TUUOI experimental data, is that for at least elevated temperature CAES concepts, parti- Fig. 1. Micrograph of Unaltered Berea culate plugging may be the most important Sandstone phenomenon investigated. Figure 4 presents 356 MimcutMc nuoomo fully controlled pressure, temperature, and humidity conditions. Figure 5 shows tMmiTMt a distribution of particles that were MUMIDMCItHIMMNn pumped into a Berea sandstone. In this mmcun W§ CQMOMON MOOUCTS case, the flowing fluid was distilled water. Figure 6 shows the incurred pressure drop across the core and the effluent and particle discharge downstream from the core.u' Nearly total plugging was the result. Backflushing was only of moderate assistance due to the near random distribution of pore and channel sizes in the sample. Fig. 3. Conceptual Representation of Particulate Plugging 8EKCA NUMBER 1 •EREA SANDSTONE IN A WATER VAPOR-AIR ENVIRONMENT CONCENTRATION- 112,000 countslnl (436 Hre. IMBanl MEAN-4.25|i MODE-4.35M MEDIAN-4.BM I V" 1 25 - I ** -' H sot 300*C PARTICLE DIAMETER, microns Fig. 5. Particle Size Distribution for Fig. 4. Disaggregation in a Water Vapor- Air Environment Berea SandstoneO) two initially identical cores of Galesville It is concluded that: 1) investiga- sandstone (an actual natural gas reservoir tion of particulate plugging may be very rock from Illinois). Both samples we-e important for potential CAES reservoirs, exposed to an atmosphere of air and water 2) filtering of the inlet air may be vapor at 134 bars for a period of two weeks. necessary, 3) sand screens may be necessary, The only difference was that the core on 4) filter cake formation is not likely, but the left was reacted at 50°C and that on bridging may occur, and 5) elevated tem- the right at 300°C. Although the test perature and cycling may aggravate the conditions were such that definitive conclu- problem, but low temperature CAES concepts sions cannot be drawn, it is apparent should not be considered immune. that substantial grain dislocation and superficial erosion occurred at the higher Thermo-Mechanical Plugging temperature. In a recent study on the effects of To study this phenomena in more de- temperature or relative and absolute per- tail, a fluid flow system is required meabilities of Berea and Boise sandstone, whereby gas and/or liquid can be vented it was concluded that the permeability to through a large diameter core under care- water was reduced substantially and the 357 THERMO-MECHANICAL PLUGGING (REVERSIBLE COMPONENT) 1000 N. BOISE SANDSTONE —o 500 SO 100 150 200 250 300 350 400 450 500 BEREA SANDSTONE 0 i 9 1 1OO TEMPERATURE (XI Fig. 7. Effect of Temperature on Absolute Liquid Permeability of Berea Sandstone(2) permeabilities were not measured and no intuitive conclusions should be drawn based on a comparison of densities and viscosities, SO 100 150 200 250 300 350 400 ISO 500 4) the permeability reductions were rever- 700 sible, i.e. recovered as the temperature EEREA CORE NUMBER 1 was decreased, 5) the cause is probably 600 thermally-induced increases in rock 500 tortuosity and is not due to increases in compressibility or decreases in poro- L, *m sityw), and 6) similar experiments should i be conducted on an CAES candidate rock. ! 300 Note that since the effects are reversible, rm - only measurements under elevated tempera- ture and pressure will reveal the decreased wo - i—>—-i r i t permeability. Post-test measurements at r--r-i T iT-rrri m room temperature are insufficient. 100 150 200 250 300 350 400 450 500 TIME iminutesl Corrosion Fig. 6. Pressure and Effluent Particle In a CAES reservoir, corrosion could Concentration Versus Time for be a serious problem if the near-wellbore Berea Cores'1) region is not maintained totally free of liquid water. Table I presents the re- irreducible water content increased sub- lationship between corrosion rates and stantially in going from room temperature various reservoir conditions for geothermal to 80°C. "For t!ie five Boise sandstone conditions.^ ' Free oxygen will aggravate cores on v/liicli displacement experiments were the problem. The potential results of run, average (liquid) permeability decreased corrosion include: damage of casing mate- from 2,050 md at room temperature to 884 md rial and surface pipes, blowouts or rapid at 80°C. This is an important new observa- depressurization and the formation of tion not made previously, to our knowledge, particles that may subsequently plug the in the petroleum literature."(2) reservoir. Figure 7 graphically presents the re- Clay Swelling and Dispersion sults of the tests. The main points are: 1) these results are for liquid water, 2) Many sandstones are known to be water permeability to oil showed substantially sensitive, i.e. permeability to water less temperature sensitivity, 3) air decreases as the water salinity decreases. 358 CLAY PRODUCTION, SWELLING AND DISPERSION Table 1. Effect of Various Parameters on SOURCE: ORMMALRESERVOM the Corrosion Rates of Iron Based CONTENT HVDROLYTK Alloys in Geothermal BrinesW REACTIONS l-actor Direction Effect on Corrosion Water Content ' t Oxygen Concentration f Salinity t AcWty Time Temperature* Silica Concentration •related to scaling phenomena CAUSE: DECREASED SALINITY DEWATERINO TREATMENT: HVOROXY-AWMINUM In tfie CAES concept, water vapor will be SWELLING CLAY PARTICLE PLUGGING injected with the airstream and/or will be produced from residual water in the warmer near well bore zones. As the air and vapor move farther out into the reservoir where the temperatures are decreasing, condensa- tion and dilution of existing formation waters may occur. This could result in the swelling of residence clays. This swelling, in of itself, should not plug the reser- voir since the effected zones will be a reasonable distance from the wellbore. However, clay plate fraying and particle dispersion may result in particulate plugging much closer to the wells since air would sweep particles in that direc- tion during the production phase of the CAES operating cycle. fresh water fresh water Figure 8 conceptually portrays this phenomenon and figure 9 shows the signifi- Fig. 8. Conceptual Representation of the cant amount of interstitual clay that can Clay Plugging Phenomena exist even in natural gas reservoir rocks. Only bentonite clays tend to cause severe dous scatter in compressibility magnitudes problems. Berea sandstone has less than from sandstone to sandstone. Although the VI by weight of bentonite clay. Figure 10 potential for matrix consolidation and shows how Berea sandstone permeability to subsidence cannot be arbitrarily dismissed water decreases with decreasing water at this point for elevated temperature salinity. Air permeability should not be storage, it is considered a lower priority as dramatically affected, but particle issue. dispersion from zones containing residual water ma., be a general particulate plugging It is being suggested that rock sof- problem. Fortunately,pretreating a reser- tening due to the production of clays from voir with hydroxy-aluminum compound tends other silicates and carbonates may be the to significantly inhibit the problem.C5/ cause of observed reduction in penatrometric Testing of potential CAES candidate rocks strength with temperature. Berea sandstone and treatment when necessary should be exhibits as much as a 30 percent decrease in considered for CAES projects. point load strength after five days of steam treatment at 300°C.(°) Fortunately ammonia pretreatment of the reservoirs seems Matrix Consolidation to totally eliminate the rock softening For Berea sandstone, the dry bulk compres- problem. " Such a treatment might be consi- sibility decreases by approximately 20 dered for CAES projects. percent in going from 20°C to 200 C at 4000 psi effective stress.(3) However, this decrease is much less then the tremen- 359 20 30 10 50 THROUGHPUT (ml/cm?) Fig. 10. Water Sensitivity (Clay Plugging of Berea Sandstone in the Pre- sence of Progressively Decreasing Salinity Waters(5)) Salts. The quantity of iron and manganese is such that precipitation-type plugging should not be a problem especially since SAMPLE -6 precipitation will primarily occur in zones already occupied by residual water and therefore never available to the air anyway. However, once again, the solid residue may be a source of mobile particles. It is Fig. 9. Micrograph Showing Interstitial suggested that reservoir rich in iron or Clay in Mt. Simon Sandstone from manganese be studied carefully and that the Media Natural Gas Storage •• dewatering of near well bore zone be expedited. Field in Illinois. Hydrolytic Reactions Evaporation and Precipitation Hydrolytic reactions refer, in this Residual water may coexist with ele- case, to the production of clay forming vated temperatures for a period of time. minerals from feldspars and carbonates. The dissolved solids and adjacent rock These reactions require liquid phase water mineralogy may be altered. Subsequent and in general reaction rates can be roughly evaporation of the water will result in considered to double for a 10°C temperature deposition of the dissolved solids. Know- increment. Potential consequences include ledge of the composition and friability those already mentioned for clays, e.g. of the residue is necessary since this may swelling, dispersion and rock softening. be a potential source of mobile particles. It is being suggested that hydrolytic reac- Naturally, this will be a function of the tions may occur for a short period of time initial salinity of the aquifer. in a CAES reservoir but that evaporation of the necessary liquid phase water may out- Oxidation Reactions pace the slowly advancing temperature front and therefore hydrolytic reactions should be Subsurface reservoirs tend to be re- of secondary concern. Reservoir hetero- ducing in nature. The CAES concept calls geneity and anisotropy dictate that the for the introduction of free oxygen and problem not be totally dismissed from con- carbon dioxide. Iron and manganese minerals sideration, however. will oxidize and precipitate as solid residues. In general, metal oxides are less dense and less soluble than metallic 360 Mineral Solutioning and Precipitation Table 2. Screening Autoclave Experiments for Site Specific Sandstone Silica, carbonate, and sulfate scaling are severe problems for the geothermal industry. Changes in temperature or » FOR OAUSVHU ROCK AND OROUNOWATER acidity can drastically affect the solu- bility of these minerals and the response to temperature is diametrically opposed for silica and calcium carbonate. The problem is being considered secondary for CAES, however. This is because the only available liquid phase water that may be heated to any degree is the residual con- tent left after air bubble development. This residual or irreducible water is es- sentially immobile in the liquid state. diameter (10-30 cm) rock cores (Figure 11). Furthermore scaling, if any, should occur The facility will be designed to permit: in zones initially occuppied by water 1) axial flow of gases and/or liquids and not available to the air phase anyway. through the core, 2) introduction of The only possible problem appears to be suspended solids, 3) temperature from 50- the sloughing of scale deposits and the 500°C, 4) confining pressures from 60 to associated production of mobile particles 400 bars, 5) uniform temperatures or im- (formation fines). posed temperature gradients, and 6) various inclinations of the core barrel to investi- PNL Experimental Program gate drainage, imbibition, and advective heat transport. Naturally it would be Current experiments to investigate ideal to have the system triaxially loaded air/water/rock interactions consists of instead of hydrostatically loaded but this autoclave tests and thermal properties does not appear to be within the project tests. A flow system capable of venting scope. liquids and/or gases and suspended solids through large diameter cores under con- trol red. temperature, pressure, and humidity PRESSURE VESSEL condition? is in the design phase. AW HEATER Table 2 presents the preliminary ex- commiiON AFFLUENT COLLECTION perimental strategy for autoclave tests on VALVE AND MONITORING a site specific candidate rock and compatible •O I groundwater. Analytical techniques include: GAS SOURCE (T) UNIFORM TEMPERATURE CIRCULATING t optical petrography, x-ray fluorescence, © rEMFWATURt PROFILE OPTION TANDP x-ray diffraction, electron microscopy, ® CONFINING PRESSURE quantitative shade differentation, ion ® UNIFORM FLUID PRESSURE chromatography, plasma-arc spectrometry, © FLUID PRESSURE PROFILE atomic absorption, and standard wet © HORIZONTAL TO *> ANGLE chemistry. Permeability measurements are made with standard permeameters. Auto- clave tests at 300°C have resulted in some Fig. 11. Conceptual Representation of the reduction in liquid permeability, and a Fluid-Flow Test Apparatus more moderate reduction in air permea- bility. General quantitative conclusions The following tests are examples of cannot be made at this time based on the the planned flexibility of the facility: limited number of experiments and the 1) porous media evaporation rates and site specific nature of the problem, how- processes, 2) thermomechanical permeabil- ever. Tests to date have been on Berea ity degradation, 3) particulate plugging, sandstone as a control sample and on 4) rock disaggregation, 5) geochemical Galesville sandstone (Media natural gas plugging, 6) clay swciiing and dispersion, storage field) as an example reservoir 7) corrosion, 8) bulk compressibility, rock. 9) geochemical reactions (hydrolytic and redox), 10) mineral solutioning and pre- The next phase of the experimental cipitation, 11) fluids incompatibility, program will include design and fabrication 12) liquid mobility, drainage and of a fluid flow test facility for large imbibition, 13) advective heat transfer, 361 and 14) chemical pretreatment. Critical formation properties . absolute and relative permeabilities Theoretical Efforts corrosion potential . caprock threshold pressure (macro) A computer code based on thermodynamic equilibrium states has been used to predict Most likely damage mechanisms changes in rock mineralogy and liquid . disaggregation constituent concentrations. Reactions . particulate plugging rates are not considered. The results . thermo-mechanical plugging for Berea sandstone reacted with capatible . thermal fatigue of well casings groundwater at 100, 200, and 300 C are and cement grouting presented in Table 3. Primary theoretical and computer study Table 3. Total Dissolved Solids Concen- areas trations and Mineral Weight . thermo-mechanical stress analysis Percents for Berea Sandstone at . heat, and mass transfer analysis 100, 200, 300°C (Equilibrium . particle transport in porous media Simulation) geochemical equilibrium reactions Concentration (Holts/Kg H:>0) Primary laboratory scale study areas Groundwater Constituent Original 100°C iOO'C 3D0°C . autoclave air/water/rock tests A) Z.3E-06 9.3E-04 2.BE-03 thermal properties identification K 6.3E-4 5.4E-03 1.3E-04 3.0E-O4 «1 7.1E-4 2.1E-02 1.4E-03 4.5E-03 . thermal fatigue analyses Ca J.JE-3 2.2E-O5 1.5E-06 l.lt-OB particle transport in porous media "9 5.7E-4 1. IE-OS 2. )£-O7 2.7E-07 . desaturation of porous media 1*1.2 - 2.9E-I7 2.1E-16 3.7E-I5 1*1+3 1.9E-07 3.U-0S 1.3E-03 . thermo-mechanical plugging tests in SI S.OE-04 4.7E-03 1.1E-02 a flowing system C 9.6E-03 1.4E-O3 1.6E-04 - 1.5E-O2 l.U-02 . geochemical alterations in a flowing % 2.3E-02 H, 3.8E-3S 7.4E-2J I.7E-21 system Mineral Original 10QOC zxfic 300°C Calcltt 8.0E-03 1.5E-0? REFERENCES Quartz 8.5E-01 1.0E-00 9.4E-01 9.2E-01 Corundun - 7.1E-0? - 1. Donaldson, E.C. and Byron A. Baker, Kyanite - 1.5E-0! S.iE-02 7.6E-02 Pyrope 1.2E-04 3.4E-03 - "Particle Transport in Sandstones" Alternate TA 4.8E-04 _ - Society of Petroleum Engineers Conf., Pyrolusite I.OE-02 I.OE-02 1.0E-02 8.7E-O3 Treaollte 1.5E-04 8.3E-06 Denver, CO 1977. Chlorite 1.4E-02 3.1E-04 - Pnlogopite 3.0E-02 7.1E-03 2.7E-O2 3.1E-02 2. Weinbrandt, R.M., H.J. Ramey, F. J. Adulerta - 2.8E-02 1.4E-02 9.SE-03 Zoisite 8.5E-O3 1.6E-02 1.6E-02 Casse, "The Effect of Temperature on Eoldote 1.0E-02 1.0E-0! 1.0E-02 I.OE-02 Relative and Absolute Permeability rajc - 1.4E-04 - Paragonlte - 1.7E-02 - of Sandstones", Society of Petroleum Ca-Hontwrillinite - K5E-07 - Engineers Journal, SPE 4142, 1975. Low Alfaite 2.OE-O2 1.6E-02 Ootontte 2.OE-03 - - Kaolinite 4.0E-02 - - 3. Somerton, W.H. and A. K. Mathur, "Effects AnorthUe 2.0E-02 - - of Temperature and Stress on Fluid Microcline I.0E-02 Flow and Storage Capacity of Porous Rocks", Proceedings of the 17th Sympo- sium on Rock Mechanics, 1977. Some important points noted in the 4. Shannon, D.W., "Corrosion of Iron-Based CAES Reservoir Stability work thus far Alloys Versus Alternate Materials in include: Geothermal Brines", PNL-2456, 1977. Critical CAES parameters 5. Reed, M.G., "Stabilization of Forma- . initial rock physical and chemical tion Clays with Hydroxy-Aluminum Solu- properties tions", J of Petroleum Technology, July . initial groundwater constituent 1972. concentrations irreducible water content 6. Day, J.J., B.B. McGlothlin, and J.L. . inlet air temperature, pressure, Hewitt, "Study of Rock Softening and Means humidity, and suspended solids of Prevention During Steam or Hot Water concentration Injection:, Society of Petroleum Engineers Journal, SPE 1561, Feb, 1967. 362 PROJECT SUMMARY Project Title: State-of-the-Art Review and Formulation of Stability Criteria for Underground Caverns Used for Compressed Air Energy Storage Principal Investigator: Dr. Paul F. Gnirk Organization: RE/SPEC Inc. P. 0. Box 725 Rapid City, SD 57709 605/343-7868 Project Goals: (1) Survey, evaluate, compile, and document those, physical chemical, petrophysical, thermal, and mechanical charac- teristics and properties of rock types and geological formations for which the CAES concept is applicable, for conditions of elevated temperatures and pressure and moisture presence; (2) Formulate a preliminary set of design and stability criteria for underground CAES caverns with expected lifetimes of 30 years; (3) Identify areas of analytical, laboratory, and field scale research that are necessary for establishing a consistent and applicable set of final stability and design criteria. Project Status: The survey of literature has been effectively completed, and the rock properties appropriate to the CAES concept are being evaluated and documented. The formulation of a preliminary set of design and stability criteria for underground CAES caverns is currently in progress. Finally, specific areas of research need are being formally identified and described. Contract Number: Sepcial Agreement No. B-51225-A-L/Prime Contract EY-76-C- 06-1830. Contract Period: June 1, 1978 - Nov. 30, 1978 Funding Level: $24,900.00 Funding Source: Battelle Pacific Northwest Laboratories 363 PRELIMINARY DESIGN AND STABILITY CRITERIA FOR CAES HARD ROCK CAVERNS Paul F. Gnirk and Debra S. Port-Keller RE/SPEC Inc. P. 0. Box 725 Rapid City, SD 57709 ABSTRACT The primary objectives of this study have been to compile rock properties data from the literature; to formulate a preliminary set of design and stability criteria for underground CAES hard rock caverns; and to identify areas of required research. The rock properties compilation has concentrated on assembling and evaluating those physical, chemical, petrophysical, thermal and mechanical characteristics and properties of igneous, metamorphic, and sedimentary rock types for which the CAES concept is applicable. This survey has indicated that the data base for hydrological and thermomechanical properties is generally inadequate for jointed rock under conditions of elevated temperature and increased moisture presence. In general, the formulation of preliminary design and stability criteria for CAES caverns has been based on information available for large underground caverns utilized for mining purposes. INTRODUCTION give rise to cavern air temperatures as high as several hundred °C. The cyclic Conceptual design studies have been nature of the pressure-temperature/air- recently conducted in the U.S.A. to iden- water fluctuations during CAES cavern tify the potential for using mined caverns operation could conceivably perturb the in hard rock to store compressed air for stress field/strength characteristics in use in electric utility load-leveling the rock structure and lead to unacceptable operations 1 3. The concept involves com- global and/or local instabilities. pressing air during off-peak demand periods, storing it in underground caverns, and The specification of CAES caverns in heating and expanding it through turbines hard rock immediately categorizes the to generate power. The cavern storage situation to one involving igneous concept is generally categorized as (granitic), possibly metamorphic, and (1) Compensated: Constant pressure with certain sedimentary (limestones, marbles, varying volume (wet system with hydraulic dolomites) rocks. The categorization compensation), and (2) Uncompensated: "hard rock" is an implication of relatively Varying pressure with constant-volume great strength and resistance to large (dry system). The useful life of a CAES ductile and creep deformation. Thus, we (Compressed Air Energy Storage) cavern is may eliminate from consideration such rock tentatively set as 30 years, over types as salt, shale, and the like. Con- approximately 7,500 cycles. versely, however, this categorization must necessarily include the influence of joints Milne, et. al.1 consider a reference (or planes of weakness), permeability (and cavern design with a volume of 385,000 m groundwater presence), and elastic/brittle at a depth of 690 m, a 20 hour storage mechanical behavior (or fracture initi- capacity, and a storage pressure of about tion) on cavern stability. Vhe cyclic 66 atm. Mailhe, et. al. describe a cavern temperature and pressure fluctuations in design with a volume of 300,000 m , com- the CAES concept also introduce the possi- pression/production durations of 7-8 hours/ bility of elastic/ductile/brittle rock 6 hours, and a working pressure of about behavior with "fatigue" and possibly 25 atm. Wittke, et. al. 5 discuss a con- transient "creep rupture" contributions. ceptual CAES design with a volume of 100,000 m^ and a storage pressure of about With regard to establishing stability 50 atm. Depending upon the temperature of and design criteria for underground hard the inlet air and the heat exchange between rock caverns used for CAES, the goals of the air and rock, the above situations may this current study have been to: 364 - (1) Survey, evaluate, compile and pressure/temperature and moisture environ- document those physical, chemical, petro- ments of a CAES cavern. In addition, physical, thermal and mechanical charac- contacts (such as bedding and pluton teristics and properties of rock types and boundaries) between various rock types in geological formations for which the CAES the rock mass (whether gradational or concept is applicable, for conditions of abrupt) must be carefully defined to elevated temperature and pressure and determine the extent of competent rock moisture presence; and to indicate possible zones of weakness. (2) Formulate a preliminary set of The nature and extent of structural design and stability criteria for under- features of a potential host rock mass ground CAES caverns with expected life- must also be defined - both primary times of 30 years; features (those formed at the time of rock (3) Identify areas of analytical, origin) and secondary features (those laboratory, and field scale research that superiir-osed on original features). It are necessary for establishing a consistent should be emphasized that such features and applicable set of final stability and can appear in all rock terranes (a geologic design criteria. area considered in relation to its suit- ability for a specific purpose) and include This paper summarizes the work accomplish- major and minor faults, folds, and joint ed to date in satisfying the above goals. systems, to name a few. These features may define the major strength, elastic, GENERAL GEOLOGIC CONSIDERATIONS thermal, and hydrological properties of the rock mass as a whole. Joint systems, INTRODUCTORY REMARKS faults, and fracturing associated with folding, in particular, may produce crit- CAES in mined hard rock caverns will ical zones of weakness and may also require extensive geological and geotech- indicate an active seismotectonic situation. nical assessments for any potential siting Likewise, such features, particularly area. Although each particular site will intense folding, may indicate regions of be a unique geologic environment with its extremely high horizontal in situ stress. own peculiarities, certain characteristics must be evaluated for each of the several IGNEOUS ROCK geologic environments that may be suitable for CAES, including: (1) regional and local petrography (mineralogy and fabric Igneous rocks are those rocks formed of the cavern host rock) and stratigraphy by solidification of molten or partially or relations of various rock formations to molten rock material either at some depth one another, and (2) primary and secondary in the earth (plutonic or intrusive rocks) structural features and/or those of or on the earth's surface (volcanic or regional as opposed to local extent. These extrusive rocks). Compositional classi- two general geologic characteristics and fication of igneous rocks is generally their significance in potential cavern based on the percentages of quartz and the stability will be briefly discussed for percentages and types of feldspar minerals situations that could be common to igneous, in the rock. Textural classification, metamorphic, and sedimentary host rocks. although somewhat complicated, is based mainly on predominate grain size. Intrus- ive igneous rocks are generally phaneritic COMMON GEOLOGIC CHARACTERISTICS - coarse to medium grained (grains greater than 1 mm in size or easily viewed with the The complete petrographic description unaided eye), while extrusive igneous of a rock includes its primary mineralogic rocks are generally aphanitic - fine composition and that of associated mineral- grained (grains less than 1 mm in size). ization or discontinuities. Also included Figure 1 includes a topical igneous rock are its macroscopic and microscopic compositional classification scheme with textural and fabric features, including accepted names for phaneritic and aphan- grain size, grain shape, grain orientation itic compositional equivalents. or lineation, foliation, schistosity, and microfracturing. These factors essentially Intrusive igneous rocks (of all determine the strength, elastic, thermal, compositions) formed at depth occur in and hydrological properties of the intact masses (plutons) of all sizes and shapes. rock and also its weathering characteristics They are generally considered to be the - all critical factors in the cyclic most favorable geologic terrane for CAES 365 caverns for several reasons: (1) they may probably one formed by massive, nonviolent be found in masses sufficiently large to flow of lava onto the earth's surface. If accommodate the CAES system; (2) the such a flow occurs in sufficient mass to intact rock is generally homogeneous and be a cavern host, other characteristics isotropic; (3) the petrography, particu- cited as favorable in igneous intrusives larly the interlocking crystalline texture, would also generally apply. imparts generally high strength and favor- able elastic, thermal, and weathering METAMORPHIC ROCK properties to the rock; and (4) porosity and permeability are generally low, Metamorphic rocks are igneous, sedi- obviously favorable characteristics when mentary, or metamorphic rocks which have considering cavern air and water leakage. been altered by temperature and/or pressure, Metamorphic petrography is extremely compli- ICHEOUS HOCK CLASSIFICATION cated; however, primary classification can MI IB1 FttDOniUMElY SB J WDOniUMELt IRON ESSENTIAL POIUSIW PlAfilKLASE] riMIOCLASC MB be based on the presence of laminated FClDSPM I FELDSPAR fHSHESttff MtfMLS CALCIIH structures in the rock resulting from DIUH PLAGIOCLASt'*!*—INHftntDIAtt* ••—FLAC—» niMCMLS segregation of different minerals into DUUTZ bUMifi - GRANODIORITC (jttaOUTE - DMtTC) layers (foliation). Foliated metamorphic >1QX 6R0U* DlMtTC rocks include slates, phyllites, schists, (AHDEIHE) SUNITC SAIMO PfllDOTlTE OuARTI 6iour and gneisses. Nonfoliated metamorphic (TRACHYTE) (IAIALO GROUP GROUP rocks include quartzites and marbles. FlLOSMlHOIDS FfLDSPATHOIOAL S«E«1ti Fl LDJPA1 HOI DAL RlPLACt (lILOSPATNOIOAL tUCHill) C*if*o OllARTI 5ROI# (FELDIPAIHOIOAL Nonfoliated rocks, such as quartzites SttALI CROUP) and marbles, when they occur in suffi- KMmmic IIKK CLASSIFICATION ciently large masses, may be as suitable XOCKKAHE cawsniw TEXTURE for caverns as igneous intrusives. Such Sun VERY FIMI C«*ttti rock is generally homogeneous and isotropic o PHYUITI Fi-t GKAIKI and has good strength characteristics, s particularly the quartzites. Marble, SCHIST TfT ! | GMIIII however, may occur less frequently in large 4 CO*IJE GRAINS DcroitncD ru&nEHis OF AM* masses, have somewhat lower strength, and Pkl*COM6LOMUT[ COAKI GRAINI 1 •«• vtn be subject to undesirable weathering and £ OuARTIITt 0AIARTZ FINE TO Count CALCITE OK DoLonitt solution effects since it is primarily calcium carbonate. The desirability of SEDIHEKIin ROCK CIASSIF1CMIBI marble for other commercial purposes would ROCK I1AIC CCHPOSITIOM TF.XTORE also tend to eliminate it as a potential host rock. CONGLOHtRATC FnAWtCNTS OF ANV IDCI COARSt ROUNDtD t«AI« > 2 m TtPt COAKSE ANCULM (RAIMS » 2 MM VA>IMS OuANTI, VARIOUS MOUNTS Of REDIUH CRAINI U/1G TO 2 wi> The major difficulties in CAES cavern 'Ufii'M, •«• »ACMfMT» Ty»E) stability in foliated rocks would probably SLITSTOKC OtMitrz JIKD CLAY HIWEIULS FlHt GRAINS (1/256 ID 1/16) arise from the foliation itself, or from SHALC schistosity (parallel grain alignment). VANIOUS fllCkD TO COARSE CRYSTALS, fOSSIL llMCSTME Count (CACOJ) Foliation and schistosity planes may cause AND FOSSIL fflAGMENTS Um the strength, elasticity, and thermal DOLONUI DOLOHITE CUV; ICOjlj SIMILAR TO LIMESTONE properties of the rock to be anisotropic i CHEIT CKKLCEDONV (SlO2> CRVfTOCRVSTALLINE, OENSC and highly variable. In addition, weather- CIPSUN (CAS0 • 2H 0) 4 ? fl«t TO COARSE CRt5T*tJ ing of such rocks may be accelerated along RDC< SALT HAL lit (HACL) planes of anisotropy and by the presence FIG. 1. GENERIC ROCK CLASSIFICATION (AFTER JACKSON, of easily altered metamorphic minerals TRAVIS', HAMBLIN i HOHAHD^, AND MCKENZIE, ET.AL.^) (chlorite, talc, etc.). This present study has included all types of metamorphic Extrusive igneous rocks (of all rocks, but concentrated on competent compositions), formed at the surface or at quartzites. very shallow depths in a variety of ways (such as nonviolent lava flows and violent SUMMARY REMARKS volcanic eruptions), are generally less favorable for CAES caverns for several The most favorable and most likely reasons: (1) they are often found in masses host rock types for CAES in mined hard of limited extent; (2) the rocks may be rock caverns appear to be igneous intrusive highly vesicular, with high porosity and rocks. Other possible hosts include permeability and generally low strength. igneous extrusive rocks, dolomites and The most favorable igneous extrusive is limestone, and massive metamorphic rocks, 366 particularly quartzites. Possibilities groundwater inflow during construction and also exist for locally favorable conditions operation of the cavern, and (2) to in other rock types. minimize air leakage from the cavern during operation. In the general sense, considerable geologic literature exists concerning Groundwater behavior is also a prim- petrography and structure of all favorable ary consideration, particularly in regard (and unfavorable) geologic terranes dis- to stability and consistency of the cussed here for CAES. We emphasize, how- saturation zone depth, or that depth at ever, that each individual site considered which voids and fissures are filled with would require its own detailed geological water under hydrostatic pressure. The investigation. optimum condition for successful CAES cavern operation is that the cavern be GEOTECHNICAL PROPERTIES located entirely in the saturated zone and that after initial construction The geotechnical properties of a rock disturbances, the upper boundary of the mass include the hydrogeological charac- saturated zone (phreatic surface) remain teristics (porosity, permeability), geo- stable. Water-filled voids and discontin- thermal gradient, in situ stress state, uities should effectively retard cavern and joint characteristics. The joint air leakage, while low permeability characteristics are generally quite site (although under saturated conditions) specific and depth dependent as regards would prevent excessive water inflow to spacing and orientation. the cavern. HYDROGEOLOGY Finally, groundwater chemistry and its stability may or may not be important, The literature survey of hydrogeology depending on rock type, in successful and case studies of various underground cavern operation. Unusual chemistry or facilities, including underground stores chemical changes, perhaps caused by tem- and mines, suggests important hydrogeolog- perature fluctuations within the cavern or ical considerations and criteria to be met by other external factors, could cause for CAES hard rock cavern siting. Primary accelerated degradation of the rock mass consideration must be given to such factors and subsequent cavern instability - either as: (1) the hydraulic characteristics of undesirable surficial wall effects or gross the intact rock and of the rock mass as a structural instability. In addition, whole; (2) groundwater behavior; and (3) undesirable chemical constituents occurring groundwater chemistry. naturally in the water or later due to cavern operation, could pose difficulties General Hydrogeological Considerations. in successful equipment operation. The important hydraulic characteristics of a rock or rock mass include primary and Particular Hydrogeolosical Considerations. secondary porosity and permeability (syn- Table 1 summarizes the magnitude of per- onymous here with hydraulic conductivity). meability generally found in various rock Porosity naturally determines how much free types, both for intact rock and rock masses. volume is available to contain fluids such It is evident that igneous intrusives, some as air and water. Porosity of intact rock, igneous extrusives (particularly flow due only to pores or voids within the basalts, rhyolites, or trachytes), high rock matrix, is said to be primary, whereas grade metamorphics, and some limestones that porosity due to fractures, fissures, and dolomites (those devoid of karst faults, or joints in a rock mass as a features) have the most favorable perme- whole, is said to be secondary. Likewise abilities for optimum CAES cavern oper- permeability, or the ability of a rock to ation. Most extrusive vesicular igneous transmit air or water, may be either primary rocks, some quartzites and marbles, and or secondary. Primary permeability depends most clastic rocks have generally on effective porosity, or on the assemblage unacceptable permeabilities. of connected void spaces in intact rock that allow fluid flow. Secondary permeabil- Generalization of the stability of ity depends on the transmission of fluids groundwater conditions is much more through discontinuities in a rock mass. difficult, owing to the very site specific Low porosity, and more importantly, low nature of the problem. Considering avail- permeability are essential in CAES host able literature, the prediction of stabil- rock masses in order to (1) minimize ity on the basis of rock type is virtually 367 impossible. Other factors external to of discontinuities under pressure) with rock type and to the cavern situation, depth, and with confining pressure in such as precipitation, runoff, and with- laboratory experiments, but few data exist. drawal for consumption, are also signifi- The permeability-depth relation is not yet cant. well defined for variations of rock types. TMU 1 Finally, literature concerning groundwater UKB OF HTMAUUC MC« KOrtltlES chemistry changes and subsequent rock tCKMt «wos m muwnic poncsm CMEAMUn (f.T.WLIC stability, particularly as it relates to ROCK TYPE CONDUCTIVITY murinn (X) (c?> (TVS) temperature changes, is also inadequate. MUM**: UWl (CCOMWlr: SCMULLV ION 0.01 • M-« ftUT VAIIEt MEATLT MHNOIIK ON 0ISCOM TO IWEOUI tlNTftUIIYt) TO TO ROCK MASS PERMEABILITV (M/SEC) TINUIT* OIHtHIIOMS MO MttvCHCV 11.2 9 8 7 7 66 5 FtlHARV: LOO, IXCCPT tO* SOME VEIICULAI 0.5 l(f(f lO" 10'10' 10"" 10"10 MCKIi lECOWUtr: LOW TO NJCM, DtUWB- 10-12 lo-H ISNEDUS It is also evident that the over- riding permeability factors in rocks generally considered favorable for CAES are the dimensions and frequency of dis- continuities of all types. As indicated 0 100 200 300 400 in Figs. 2 and 3, the literature suggests EFFECTIVE CONFINING PRESSURE (IIPA) general reductions of permeability (due FIG. 3. LABORATORY DETERMINATION OF PERMEABILITY to a combination of frequency and closing AS A FUNCTION OF CONFINING PRESSURE 368 GEOTHERMAL GRADIENT AMD HEAT FLOW duced by the weight of the overburden, but may be perturbed by regional or local The possible major sources of heat in tectonic features. As indicated in Fig. the upper few km of the earth's crust 4, the average vertical stress to a depth include: (1) the outward flow of heat from of 3 km is of the order of 0.025 MPa/m the central core of the earth; (2) the (after Haimson11*) to 0.027 MPa/m (after presence of cooling magmas; (3) the dis- brown and Hoek15), which corresponds to integration of radioactive elements; and, an average bulk density of 2,550 to (4) subcrustal convection currents 2,755 kg/m . The approximate limits (Levorsen11). The discussion of these given in Fig. 4, as deduced from the sources and the concept of global heat compilation of worldwide published data flow is most relevant to the field of plate by Brown and Hoek15, are indicative of tectonics and is not particularly signifi- variations in the overburden density for cant in the CAES situation. What is most different rock types and of the influence significant here is simply that rock of tectonic features. temperature does increase with depth, and that the temperatures likely to be VERTICAL IN SITU STRESS, O"z encountered at the relatively shallow 10 20 30 HO 50 60 70 depths of a CAES cavern (600 to 700 m) are dependent on such factors as atmospher- 1 1 1 ic temperature changes and groundwater circulation, and also on material thermal conductivity and local geologic structure 500 - APPROX. LIMITS - 1l ^- (AFTER WORLD-WIDE (Levorsen ). Elevated temperature at depth \ \c \^^ DATA COMPILATION may be significant in the initial mining BY BROWN S H0EK15) phases, where it could pose problems in 1000 - cooling and ventilation. However, it is crz = 0.027 Z - doubtful, considering available published -j ''••<' (AFTER BROWN & S '•-.. H0EK15) data and again, depending greatly on "" specific site and cavern depth, that the _, 1500 geothermal gradient of an area could pose g serious difficulties in the development Q and operation of mined CAES caverns. ?. 2500 D increases in one km of approximately 40°C HETAM0RPH1C > (AFTER COMPILATION --s\ ^ and 30°C, respectively. Such temperature ° SEDIMENTARY S BY BROWN 8 HOEKlS) --..\ increases probably would not be significant, 0 *•. considering that CAES operation may cause 3000 1 1 1 1 1 1 cavern temperatures to increase to as much FI3. i). VERTICAL IN SITU STRESS AS A FUNCTION OF DEPTH. as several hundred °C. A fair amount of such data exists for specific localities and for larger regions; however, the nature The ratio of the in situ horizontal of the factors are site specific and could stress to the in situ vertical stress is be obtained for a particular area, under known as the coefficient of lateral earth usual circumstances, using presently stress. Fig. 5 illustrates that this available technology. coefficient varies from less than one to greater than three at depths of a few IN SITU STRESS STATE hundred meters. For depths of several kilometers, the coefficient ranges from The in situ stress state in a rock about one-third to one. As indicated by mass is defined as that state of stress the limiting curves due to Haimson11*, as which exists prior to disturbance of the obtained from hydrofracturing data, the rock by excavation,. This natural state of two orthogonal stresses in the horizontal stress, in conjunction with the strength plane are not necessarily equal. The and structural characteristics of the rock, curves by Brown and Hoek'15 are upper and are important in determining the geometrical lower "average" limits as deduced from a shape and dimensions of an underground worldwide compilation of published data. excavation. Generally, the in situ We note that the in situ principal stresses vertical stress is taken to be that in- may not be aligned with the vertical and 369 horizontal directions due to the influence The survey of thermal fatigue, spall- of tectonic features. ing, and cracking literature for rock has not yielded abundant applicable informat- COEFFICIENT OF LATERAL EARTH STRESS. K,, -O"H0R/crz ion, and much of that available (partic- ularly regarding thermal fatigue) is of ,0 1 2^ 3 a qualitative nature. However, enough study has been conducted on the three processes to indicate that they would merit consideration in the design of hard rock CAES caverns. Rock thermal fatigue is essentially a type of weathering process. The effect of cyclic heating of a rock, with or with- out the effects of cyclic moisture con- ditions, generally causes some degree of rock disintegration, particularly leaching of compounds from the rock and subsequent weight loss. Tne significance of this phenomena is evident when the cyclic temperature and moisture conditions of a CAES cavern are considered. Likewise, the processes of thermal spalling and thermal cracking are also significant, and their disintegration effects (along with METANORPHIC > (AFTER COHPILAT those resulting from fatigue) may occur SEDIMENTARY S BY BROWN 8 HOEK simultaneously. Thermal spalling and/or i ' i -^ i I 3000 cracking of a rock surface may occur when heating induces thermoelastic stresses FIS. 5. COEFFICIENT OF LATERAL EARTH STRESS AS A FUNCTION OF DEPTH which exceed the fracture strength of the rock. Spalls are relatively thin and are THERMAL/MECHANICAL ROCK PROPERTIES usually curved pieces of rock broken off from the rock surface during or after On the basis of many articles in the heating, and may be several meters in published literature, a brief compilation length or microscopic in size. Thermal of thermal/mechanical rock properties is cracking results in the production of given in Table 2. The properties are fractures in the rock, usually microscopic, listed 'in terms of ranges of values, and as indicated in Fig. 6. are indicative of the relative strength and thermal characteristics of the various The causes and degree of suscept- generic rock types, without differentiation ability to fatigue, spalling, and cracking for competency. In general, the data for are many. They includ e rock type (com- jointed rock under conditions, of elevated t position and fabric), maximum temperature temperature and confinement stress are change, rate of temperature change, pres- inadequate for cavern stability evaluations!. ence of fluids, composition of fluids, and frequency of cyclic temperature and mois- MKK OF TKHmCHUICN. HOCK NOPRTIES ture changes. Thermal fatigue effects in a CAES cavern could conceivably cause WCWFIPCI rocais COEF. OF WML now. LISM TOBltf COIVCS* OF •oissw-s anucT- SKCIFIC severe leaching of compounds which could mx WE sire T1CHW. KAT moon usncm MTI0 IHTT 3TWSin STWWH 1 alter groundwater chemistry, make cavern »t) «f»> ao-V/s dO-'/ *) water unsuitable for effective equipment IOKVS 1S.0 n IM n.n i.n? 1.751 ».n not NU 1 operation, and cause elastic and strength inmivt mn }.< f.j 1.5 o.os 0.76 0.X1 «.s 7» properties of the rock mass to change. xatm mz M.S JSJ Q.« 2.71 i.m u.o ffi Major spalling effects could also produce nnwtvt • IK 1.5 17.2 IT? 0.0> 1.17 O.(K $i0 «t stimmun lD.1 Ml •• 370 and thermal expansion and diffusivity rock types from possible CAES host rock (Friedman16). mass, must be undertaken. 160 STABILITY CRITERIA FOR CAES CAVERNS DEFINITION OF INSTABILITY The notion of instability of CAES caverns may be categorized as follows: (1) Global Rock Instability. Identified by massive roof falls, wall slabbing, and floor heave, leading to the loss of structural integrity of the cavern and/or its entrance. (2) Local Rock Instability. Identified by localized thermo- mechanical spalling and thermo- chemical disintegration of the rock over the cavern periphery, leading to participate transport 100 100 300 500 700 during compressed air withdrawal to the turbine system. TEMPERATURE (°C) (3) Global Air Leakage Instability. • SLOWLY CYCLED, THIN SECTION Identified by unacceptable air O QUENCHED, THIN SECTION leakage from the cavern during Q SLOWLY CYCLED,, POLISHED compressed air injection and SURFACES storage (due to greatly enhanced hydraulic conductivity as a FIG 6. MICROFRACTURE INDEX AS A result of induced fracturing or FUNCTION OF TEMPERATURE joint dilation). LEVEL (AFTER FRIEDMAN16). In practice, we may define the time More study is needed on thermal periods of instability concern for a system effects in rocks, particularly with regard of CAES caverns as: (1) Excavation, (2) to the contribution of each factor listed Operation, and (3) Decommissioning. The above to total rock mass instability. notion of global rock instability applies Some general conclusions drawn from present to the excavation and decommissioning literature are that thermal effects combin- periods, while all three instability ed with moisture effects generally are concerns apply to the operational period. more degenerative than thermal effects Clearly, the concept of stability criteria alone, and that abundant microfissurization involves the specification of limits on the promotes more disintegration (Aires-Barros, thermal/rock mechanics/hydrological behavior et al17) as does higher porosity and of the rock mass, wherein instability is higher permeability. prevented when the limits are not exceeded. In addition, carbonates, such as EXCAVATION STABILITY limestones and dolomites may be much more susceptible to thermal and moisture effects We shall assume that the geotechnical than igneous rocks (Mailhe, et al"*). and thermal/mechanical properties of a Igneous rocks with abundant mica minerals rock mass for a potential CAES site can be and calcium plagioclase feldspar may be properly characterized and quantitatively more susceptible than other igneous rocks defined. For a choice of cavern depth, (Aires-Barros, et al17). At present, shape, dimensions, and spacing, we may however, the literature on the cyclic compute, by say the finite-element method, thermal and moisture effects of intact the state of stress in the rock mass during rock are only marginally adequate and that a simulated excavation. The appropriate on jointed rock totally inadequate. There- criterion for evaluation of the global fore, more general laboratory testing of stability of the excavation is the Mohr- such effects, and also testing of specific Coulomb failure condition. This criterion, 371 which mathematically relates the tensile The local rock instability of the and compressive strengths of a rock to a cavern periphery is related to the spalllng state of applied stress, permits evaluation and microfracturing characteristics of the of the potential for incipient rock failure. rock under cyclic pressure/temperature Specifically, if the state of stress around loading and air/water intera. ion. The the excavation violates the Mohr-Coulomb limit of acceptable rock disintegration criterion, failure of the rock mass is must be established from the viewpoint of indicated. The criterion is applicable to allowable particulate transport to the both the intact rock and the joints, with turbine system during compressed air appropriate characterization of the withdrawal. strength properties in each case, and to a failed rock mass in the sense of DECOMMISSIONING STABILITY residual strength. By altering the cavern geometry and dimensions, and the sequence After cessation of the operational of excavation, the stability of the cavern phase of a CAES cavern, consideration may be effectively optimized for a given must be given to the eventual collapse of rock mass and state of in situ stress. the cavity, leading to possible surface subsidence. It would appear that an appro- An actual measure of instability priate evaluation of "long-term" stability must be quantified in terms of loss of would involve consideration of a creep cavern serviceability. Regions of failure rupture criterion in conjunction with the within the rock mass around the periphery stress state around the cavern. of the cavern, as indicated by finite- element modeling, do not necessarily imply DESIGN CRITERIA FOR CAES CAVERNS global instability if the regions are localized and reasonably disconnected, or Due to operational and/or economic can be "hardened" by use of artificial considerations, the choice of a CAES support. Thus, a certain degree of cavern site may be somewhat restricted. "contained" rock failure may be acceptable Apart from some lateral variation in in the sense that the future serviceability selection in a given area, the depth and of the cavern is not impaired in a rock quality may be the only significant detrimental fashion. variables of choice. An important real- ization is that in practice it may be OPERATIONAL STABILITY necessary to encompass the design of a particular CAES cavern system within the During the operation of a CAES thermal/rock mechanics/hydrological cavern, the rock is subjected to cyclic capabilities of a given underground rock variations in applied pressure and mass. As a consequence, "hardening" of temperature. These conditions induce the caverns by artificial support means additional stress perturbations in the may be required in order to obtain a surrounding mass, and the global insta- technically and economically feasible CAES bility of the cavern must be evaluated by system. In fact, the initial start-up of use of the Mohr-Coulomb condition of rock the operational phase may require a special failure with temperature-dependent pro- sequence of compressed air cycling in order perties. However, in this situation, to promote hardening or shakedown of the the strength of the rock will probably be rock by psuedo-plastic deformation within progressively reduced with the number of the limits of global stability. loading cycles. The design criteria for a cavern must The hydraulic conductivity of the be established from the viewpoint of (1) rock, which is a function of stress and the type of CAES system, (2) the desired temperature, will be perturbed by the air volume and pressure, and (3) the initial excavation, and subsequently thermal/rock mechanics/hydrological perturbed by the cyclic pressure and constraints appropriate to lue rock mass. temperature loadings. Failure of the The constraints must be utilized in the intact rock and/or joints will also give determination of optimum cavern shape, rise to conductivity perturbations. From dimensions and spacing, and excavation the viewpoint of global air instability, sequence. the criterion will be related to the loss of air from the cavern in an economic or operational sense. 372 ACKNOWLEDGMENTS (12) Stille, H., Burgess, A., and Lindblom, U. E.: "Groundwater Movements The authors are indebted to Mr. Henry Around a Repository", KBS Tech. Rept. No. Waldman for the computer processing of the 54:01 (1977). rock properties data, and to Mr. Joe L. (13) Brace, W. F., Walsh, J. B., and Ratigan and Dr. Arlo F. Fossum for their Frangos, W. T.: Permeability of Granite constructive suggestions during the course under High Pressure", J. Geophys. Res., of the study. V. 73 (1968), pp. 2225-2236. LIST OF REFERENCES (14) Haimson, B. C: "The Hydro fracturing Stress Measuring Method and Recent Field (1) Milne, I. A., Giramonti, A. J., and Results", Int. J. Rock Mech. Min. Sci., Lessard, R. D.: "Compressed Air Storage Vol. 15 (1978), pp. 167-168. in Hard Rock for Use in Power Application", (15) Brown, E. T. and Hoek, E.: "Trends Rockstore 77, Stockholm (1977), Vol. 2, in Relationships between Measured In-Situ pp. 199-205. Stresses and Depth", Int. J. Rock Mech. (2) Willett, D. C. & Lawrence, J. D.: Min. Sci., Vol. 15 (1978), pp. 211-215. "The Design of Reservoir Caverns for (16) Friedman, M.: "Thermal Cracks in Underground Pumped Storage", Rockstore 77, Unconfined Sioux Quartzite", Preprint - Stockholm (1977), Vol. 2, pp. 145-148. Proc. 19th Symp. on Rock Mechanics (3) "Preliminary Feasibility Evaluation (1978), pp. 423-430. of Compressed Air Storage Power Systems", Final Technical Report (R76-952161-5) to (17) Aires-Barros, L., Graca, R. C, and ERDA & NSF by United Technologies Research Belez, A.: "Dry and Wet Laboratory Tests Center (1976); Vols. I & II. and Thermal Fatigue of Rocks", Eng. Geol., V. 9 (1975), pp. 249-265. (4) Mailhe, P., Comes, G., Perami, R.: "Geological and Geotechnical Process for the Siting of a Hydropneumatic Pumped Storage Plant in Brittany (France)", Rockstore 77, Stockholm (1977, Vol. 2, pp. 495-500. (5) Wittke, W., Pierau, B., & Schetelig, K.: "Planning of a Compressed-Air Pumped- Storage Scheme at Vianden/Luxembourg", Rockstore 77, Stockholm (1977), Vol. 2, pp. 149-158. (6) Jackson, K. C.: Textbook of Lithology, McGraw-Hill, Inc. (1970). (7) Travis, R. B.: "Classification of Rocks", Quart. Colo. Sch. of Mines, V. 50 (1955). (8) Hamblin, W. K. and Howard, J. D.: Exercises in Physical Geology, Burgess Pub. Co. (1975). (9) McKenzie, G. C, Pettyjohn, W. A., and Utgard, R. 0.: Investigations in Environmental Geoscience. Burgess Pub. Co. (1975). (10) Walia, M. and McCreath, D. R.s "Siting Potential for Compressed Air and Underground Pumped Hydro Energy Storage Facilities in the United States", Rockstore 77, (1977), V. 1, pp. 117-123. (11) Levorsen, A. I.: Geology of Petroleum, W. H. Freeman and Co. (1967). 373 PROJECT SUMMARY Project Title: Long-Term Stability of Compressed Air Energy Storage Caverns Principal Investigator: R. L. Thorns Joseph D. Martinez, Coprincipal Investigator Organization: Institute for Environmental Studies Atkinson Hall, Louisiana State University Baton Rouge, LA 70803 (504) 388-8521 Project Goals: To conduct a state-of-the-art survey on storage caverns in salt domes relative to compressed air energy storage (CAES) and to formulate preliminary lonq-term stability criteria for salt dome CAES caverns. Project Status: Caverns in salt domes offer perhaps the most promising type of geostorage space for compressed air energy. Storage of hydrocarbons has been practiced in salt domes of the U. S. Gulf Coast region for approximately twenty-seven years; however, state-of-the-art techniques relative to geostorage have been considered largely proprietary until only recently. Many early storage caverns were not designed specifically for storage purposes, but were a result of primary solution mining (brining) operations. Little effort was made to monitor storage caverns except when serious problems obviously had already developed, e.g., collapse into caverns that had penetrated through the dome surface by uncontrolled dissolution. Generally storage of hydrocarbons, particularly in liquid form, in salt dome caverns has been highly successful. However, little is known about the long-term effects of cyclic loads on salt rocks. Compressed air energy storage (CAES) will involve continuous daily cycling of pressure, temperature, and humidity. Possible deleterious effects due to cyclic loadings must be controlled so that the CAES caverns remain stable and functional over time periods of approximately thirty-five years. Long-term stability criteria are presented for CAES caverns in salt domes. The criteria, which are relatively general in nature because of the unknown effects of cyclic loadings, are based on a review of relevant technical literature and information from persons knowledgeable about storage in salt dome caverns. Finally, a methodology is presented for development and imple- mentation of quantitative, long-term criteria applicable to CAES caverns in site specific salt domes. Contract Number: Special Agreement B-54804-A-L Contract Period: Oct. 1977 - Aug. 1978 Funding Level: $17,000 Funding Source: Battelle Pacific Northwest Laboratories 375 PRELIMINARY LONG-TERM STABILITY CRITERIA FOR COMPRESSED AIR ENERGY STORAGE CAVERNS IN SALT DOMES Robert L. Thorns Institute for Environmental Studies Louisiana State University, Baton Rouge, LA 70803 ABSTRACT Caverns in salt domes offer perhaps the most attractive type of geostorage for com- pressed air energy. Storage of hydrocarbons in salt dome caverns has been practiced in the U.S. Gulf Coast region for approximately twenty-seven years, however this kind of storage has involved mainly liquids and relatively slowly varying pressure loadings. Compressed air energy storage (CAES) in salt caverns will involve continuous daily cy- cling of pressure, temperature, and humidity. Possible deleterious effects due to cy- clic lo. dings must be controlled such that the CAES caverns remain stable and functional over time periods of approximately thirty-five years. Long-term stability criteria are discussed relative to CAES caverns in salt domes. The criteria are based on a review of the technical literature and interviews with persons knowledgeable about storage in salt domes. The methodology for development and implementati-on of quantitative criteria spe- cific to any particular potential CAES dome also is presented. INTRODUCTION tenn "stability" of air storage caverns must be considered as a primary concern This, study is directed to the long- in projecting the satisfactory operation term stability of compressed air energy of CAES facilities.2 storage (CAES) caverns in salt domes or salt anticlines. The Gulf Coast Basin, As used in this report, "stability" as depicted in Fig. 1., is the only con- of a storage cavern implies the extent to firmed salt dome region in the U.S.A.; which an acceptable amount of cavern however, the salt anticlines of the Para- storage volume can be utilized with rou- dox Basin also would possess many geo- 1 tine maintenance for a specified time in- logic features similar to domes. terval , e.g., 35 years. In this context, cavern stability is relative to both planned utilization and time interval of operation. Although the storage of liquid hydro- carbons has been practiced in United States (U.S.) Gulf Coast salt domes for around twenty-seven years, the cyclic pressure and temperature variations which are an integral part of CAES operations intro- duce new considerations relative to the long-term stability of the associated storage caverns. Fig. 2 is a schematic of a "leached-out" storage cavern in a salt Ml•>!>•« ItIM JMHM 111 IHIIHI, IfTI dome along with symbols for significant FIG I SALT DEPOSITS IN THE UNITEO STATES spatial dimensions. Fig. 3 depicts a typ- ical weekly cycle for pressure and temper- ature within a CAES air reservoir cavern. Projections of a desirable useful "life" for CAES facilities include a time Many hydrocarbon storage caverns now period of around 35 years of cyclic oper- in operation in salt domes are a secon- ation. Thus the air storage caverns, which dary benefit of an original brine solu- are an essential and integral component tion-mining operation and were not de- of a CAES plant, should be designed and signed with stability as a primary con- operated so as to perform satisfactorily cern. More recently however, energy util- over the intended life of the overall fa- ization concerns have gained such promi- cility. It follows that the long- nence that optimizing cavern designs in 376 "383 • Site Specific Utilization History of Domes • Geology of Salt Domes Including Caprock and Neighboring Formations • Material Properties of Salt and Adjoining Materials • Configurations of Dome and CAES Caverns • Loading Cycles for CAES Operations FIG. 2. SCHEMATIC OF Fig. 4. Principal Factors Affecting CAES CAVERN IN SALT DOME Long-Term Stability of CAES Caverns in Salt Domes •0 / through the salt that could open under TO (MM I CAES operations; and a good record of J-A A' ability to maintain well casings through •0 • \ \ A // caprock (if any) into the salt. Any pre- 90 '—V vious brining and/or sulphur mining opera- / \ 40 / \ tions in the caprock should be checked to T H 1 PCI */ » / insure that associated effects such as 90 " " "'' surface subsidence have essentially ceased and will not otherwise significantly 10 affect a CAES facility. 10 The geology of potential air-reser- WON TOE S ' WED THUR Ft I SAT SUN voir salt domes, including the caprock and neighboring formations, should be studied and reported in the early stages of plan- FIG. 3. TYPICAL CAES CAVERN CYCLE ning for CAES facilities. For example, the presence of liquids or gases in the salt .domes for storage purposes has be- salt stock itself may not disqualify a po- come a primary, rather than a secondary, tential storage dome. However, the pres- consideration. Thus a great deal of in- ence of a highly permeable anhydrite sand terest now exists in further developing at the caprock-salt contact could permit the technology currently associated with circulation of brines with H2S gas that design of salt dome storage caverns, and could cause rapid corrosion of well cas- extending this technology to implement ings, unless preventive measures were relatively new storage concepts, such as 3 taken by using appropriate cement around compressed air energy storage. the casings. Leaks in casings could lead to an abrupt depressurization of CAES GENERAL STABILITY CRITERIA caverns with resulting damage to the sur- rounding salt ("wall slabbing" and "roof General criteria for long-term sta- falls"). Furthermore, the geology of the bility of CAES caverns in salt structures salt, caprock, and adjoining formations must account for a number of factors. The is directly related to the mechanical principal factors are listed in Fig. 4. properties of the associated materials. Criteria related to these factors follow in a summary format. Salt domes display anomalous zones in which weaker salt, gas pockets, and hydro- The site-specific utilization history carbons are present in the salt stock. of potential air storage salt structures Megascopic features associated with these should be researched to insure man-made zones would directly affect CAES caverns. effects will not threaten the integrity of The geology of salt domes in the U.S. Gulf the containing salt. For example, there Coast generally varies in character, and must be an absence of man-made hydrofrac- should not be assumed to be the same even turing connections or solution channels for domes separated only by a few miles. 377 The state of tectonic stress within gram methodology for developing data spe- the salt stock of U.S. Gulf Coast domes cific to potential CAES operations. is usually assumed to be hydrostatic in character. However, this assumption The configuration of a potential should be checked; and, it is suspect in CAES dome must be capable of accommodat- the Paradox Basin salt anticlines. In ing a system of caverns suitable for a both cases, the study of tectonic stres- compressed air reservoir. Depths of CAES ses within the salt stock deserves care- caverns apparently can ranqe from approx- ful attention, because creep closure rates imately 2000 to 5000 ft .(609.6-1524.0 rO . of caverns in evaporitic rock are strongly Minimum dimensions for salt "wall" thick- affected by the state of initial geostatic ness between caverns and "roof" thickness stress. over caverns (as depicted in Fig. 2.) will require analyses utilizing site- An ideal CAES dome would include the specific material properties response following geological features: (1) mass- associated with CAES loading environments. ive anhydrite caprock (if any) free of Examples of minimum wall thickness be- vugs and lost circulation zones, (2) solid tween storage caverns include 100 ft caprock-salt contact, (3) relatively homo- (30.5m) for pseudostatic hydrocarbon geneous salt of uniform character and free storage5 and 7?1.8 ft (220m) for the two of liquid and/or gas. CAES caverns at Huntorf.6 Examples of cavern roof thickness include 200 ft The material properties of the salt (61.Om) for hydrocarbon storage and ap- and adjoining geologic units, e.g., cap- proximately 328 ft (100m) at Huntorf. rock, should be established so that sta- These recommended and existing dimensions bility analyses will have a basis for pre- apparently are based mainly on a great cluding cavern failure under environments amount of first hand experience in hydro- associated with cyclic CAES operations. carbon storage and brining operations in Currently there does not exist an adequate the U.S. Gulf Coast and in West Germany material properties (mechanical, thermal, respectively. In addition, the West and chemical) data base for predicting the Germany design incorporated data from long-term response of rock salt subjected model testing of rock salt. to cyclic loadings of pressure, tempera- ture, and humidity typical of a CAES Cavern shape, size, spacing, and facility as depicted in Fig. 3. Some number obviously can all vary depending creep data exists for laboratory specimens upon the requirements of any particular of rock salt subjected to elevated temper- CAES facility operation. A systems opti- atures and varying load paths. Also, a mization analysis should be used as a number of associated material "laws" have basis for establishing these parameters. been proposed for numerical modelling Intercavern air flow effects, e.g., fric- applications. However, for CAES opera- tion and thermal losses, along with total tions, the important long-term effects of air volume flow requirements would neces- creep-rupture and low-frequency fatigue sarily be coupled with cavern system sta- of rock salt in a cyclic environment are bility requirements in such an analysis. largely unexplored phenomena. The general CAES criteria for dome Deterioration of salt around caverns configuration requires that an adequate also can be anticipated to depend upon volume of salt must exist at a workable site-specific material responses to a air reservoir depth so as to provide a rapid drop of cavern pressure. For ex- surrounding envelope of salt sufficient ample, a pressure drop from 2,219 psi to prevent the formation of connections (15,300 kPa) to 114 psi (786 kPa) over from the reservoir to the dome surface. 158 hours apparently caused a significant Such connections could be of a long-term roof fall in an experimental gas storage progressive character, and would be due cavern in a German salt dome.1* to a combination of creep-rupture, mech- anical and thermal fatigue, air penetra- The topic of material properties of tion, and hydrofracturing effects. rock salt and the relationship to cavern stability is relatively complicated and The loading cycles for CAES operations somewhat lengthy. Details on this topic must range within bounds obtained from an are beyond the scope of this review re- engineering analysis and a verified data port. A later section outlines a pro- base. A verified data base necessarily will include field experiments within a 378 potential CAES salt structure with ob- survey relative to long-term stability of served test results correlating with pre- caverns in salt domes. Attention was di- dicted results from an appropriate analy- rected to information sources on aspects sis. The analysis should be based on a of stability which were particularly rele- combination of laboratory and/or in-situ vant to potential CAES operations in salt tests and numerical modelling. Since domes. Thus a literature survey was under- cavern pressure is the loading parameter taken, and also discussions were held with most easily controlled under actual oper- persons experienced in the geostorage in- ating conditions, it should receive maxi- dustry involving salt domes. The princi- mum attention in tests and numerical mod- pal findings are reported in this section elling simulating prototype CAES loading in summary format. environments. However, at the same time the dependent and coupled loading param- Field data associated with storage of eters of temperature and humidity also natural gas in salt dome caverns currently must be properly represented. It should is most directly applicable to long-term be noted that most of the factors affect- stability studies of CAES caverns. Review ing long-term cavern stability under CAES of th<= geostorage related literature and operations could be strongly coupled. discussions with operators of a variety of storage operations (hydrocarbons and/or The planned cavern pressure cycle at brine; in salt domes have revealed signifi- the Huntorf CAES facility includes charg- cant amounts of data. However, the data ing for eight hours to 70 atm, then dis- typically is fragmented and not directly charging for two hours to 50 atm. An- applicable to stability criteria for CAES other CAES study includes an example cycle operations. The most similar field opera- with 24 hr pressure variations of approx- tions to CAES exists currently in the nat- imately 13 atm. The daily cycle is em- ural gas storage industry. Experiences in bedded in a weekly cycle with approximate storage of natural gas in salt dome cav- maximum and minimum values of 78 and 46 erns should be carefully evaluated in atm respectively. future planning of CAES operations. In addition to effects of CAES load- CAES storage raises more concerns with ing cycles on the rock salt around air- stability problems than liquid hydrocarbon reservoir caverns, effects on well cas- storage, solution mining, or natural gas ings and "casing seats" into the salt storage currently operational in salt dome also must be considered. Such effects caverns. Long term cyclic loadings of are probably better understood (than pressure and temperature could have more effects on salt) by the storage industry deleterious effects on cavern stability operating in Gulf Coast salt domes. How- than current psuedo steady-state storage ever, the progressive deterioration of operations in salt dome caverns. However, casing-seat grouts should be explored the extent of the possible increase in under cyclic pressure and thermal load- deleterious effects is currently unknown. ings. The Huntorf CAES plant utilized Roof falls reportedly have occurred in an inner "free" tube to transmit air, thus conjunction with rapid pressure drops in uncoupling the outer well casing from ther- salt dome natural gas storage caverns, mal or pressure "shock" loading effects. which implies rates of pressure drops In summary of this section, general associated with daily CAES cycles could criteria for long-term stability of CAES affect cavern stability. Surface contact caverns in salt domes have been reported; by brine, as contrasted to gas, is be- however, a data base specific to the lieved by some investigators to increase cyclic loading character of CAES opera- the plasticity characteristics of rock tions must be developed prior to specify- salt. If this is true, spalling of the ing quantitative criteria with any degree surfaces of CAES caverns will be more of confidence. As noted previously, a severe than for brine filled caverns methodology is proposed later within this under similar pressure and temperature report for establishing site-specific and loading cycles. quantitative criteria. A number of workers with field exper- STATE OF THE ART SURVEY; ience in the domai storage and mining PRINCIPAL FINDINGS industry are dubious about the signifi- cance of conventional rock mechanics lab- A principal part of this report inclu- oratory testing programs relative to pre- ded the findings from a state-of-the art dictions of salt dome cavern behavior. 379 They share a strong conviction that only atively sophisticated monitoring equipment field tests and related experiences under now exists for such purposes, e.g., a actual site-specific conditions can be laser ranging device that can monitor relied upon at the present time to yield distances within a gas-filled salt cavern reliable results for predicting cavern under pressure with an independent sensi- behavior. tivity of ± 10 cm.9 Other sensitive mon- itoring equipment also may be useful for Test caverns have been employed in recording cavern performance, e.g., micro- West Germany to obtain field data for seismic monitoring systems and tiltmeters. performance of natural gas storage cav- erns. Apparently no such scaled down cav- The study of fundamental rock salt erns with tests have been utilized thus mechanics for many kinds of storage in far in the United States (U.S.). Some domes is still in a relatively early stage borehole test procedures have been pro- The more complicated loading conditions posed both in the U.S. and abroad as an associated with CAES caverns will entail a intermediate method for obtaining data on considerable amount of careful investiga- in-situ rock salt performance that could tion before full confidence can be placed be related to cavern stability. in associated specific and detailed quanti- tative long-term stability criteria. Laboratory testing programs for determining rock salt behavior have been Concerns with compensated (constant conducted mainly with compression tri- pressure), as contrasted to noncompensated axial or uniaxial tests. However, more (constant volume), CAES caverns would in-"" infrequently used triaxial extension tests elude possible (1) erosion of salt in the were found to be of primary value in at air-reservoir "walls' and connecting open- least one major study of creep closure of ings due to liquid flow, and (2) abrupt salt dome caverns.7 Triaxial extension loss of cavern pressure due to a champagne- tests permit a more comprehensive evalua- effect blowout of the compensating water tion of creep-rupture, which is a signifi- (brine) "leg". However, compensated cav- cant effect in long-term stability of erns would possess an apparent advantage openings in rock salt. of low amplitude cyclic pressure loading of surrounding rock salt. Crystal sizes (around 0.25 in (0.64 mm) "diameter") in polycrystaline rock PRINCIPAL RECOMMENDATIONS salt are relatively large, thus scale effects are probable in laboratory tests A relatively brief set of summary with relatively small specimens. Some principal recommendations are presented test data imply dependence on size of cyl- in this section. They are based on the indrical rock salt specimens falls off findings from the state-of-the-art survey sharply for 4 in. diameter (and above) and additional study relative to require- cylinders with standard length to diameter ments for long-term stability of CAES ratios of 2 or slightly larger. caverns in salt domes. Numerical modelling of rock salt be- Long-term stability criteria for havior with the finite element method has CAES caverns should be concerned with been in use for some time. Numerical mod- effecting acceptable limits on time de- elling can be employed as a powerful aid pendent creep closure, creep rupture in developing and verifying stability cri- (spalling or slabbing), and superimposed teria during the laboratory and in situ deleterious cyclic loading effect. How- testing phases as well as during the pilot ever, short-term effects that could cavern (if used) and field operational threaten the overall integrity of the air- phases of prototype CAES air-reservoirs. reservoir also must be limited, e.g., hydrofracturing, well casing failure, and Empirical methods combined with field cavern roof and wall "falls" due to experience have been used successfully in abrupt depressurization. cavern stability studies." The CAES cav- erns at Huntorf, West Germany, were de- Specific long-term stability cri- signed primarily on this basis. teria for CAES caverns should be devel- oped in a closely coordinated and bal- Stability criteria ultimately must be anced program utilizing complementary: verified by field data from pilot and/or (1) laboratory and/or bench-scale test- operational prototype CAES caverns. Rel- ing of site-specific salt. (2) numerical 380 modelling, and (3) field testing. Later stages which incorporate the three listed sections of this report include schemat- requirements for development of long-term ics illustrating this concept in detail. stability criteria for CAES caverns in salt domes. Correlation studies should be per- formed on results from laboratory and/or bench-scale tests, borehole and/or model Start Stage (1), Laboratory Testing cavern tests, and proto-type cavern tests. Consistent findings relative to long-term cavern stability then could be (a)J, used to develop an optimum testing pro- Formulate Stability Criteria With. gram for determining necessary site- Latest Data Base specific input parameters for stability 1 criteria. Such a program would be ex- )„ tremely valuable in planning additional CAES cavern systems and operations in Update Analysis salt domes. Figure 5 is a summary list of topics icl I which should be studied to obtain addi- Model Numerically Current Stage tional information relative to long-term With Latest Data Base stability of CAES caverns in salt domes'." (d) w Agreement Between Numerical Model noJ Long-Term Creep, with Creep Rupture, And Physical Behavior? of Rock Salt. Vis • Effects of Pressure and Temperature Repeat Steps (a)-(d) For Stage: Loading Rates. (2) Pilot CAES Cavern Low Frequency Fatigue, with Coupled (3) Prototype CAES Cavern Cyclic Pressure, Temperature, and Wetting Conditions. Progressive Air Penetration of Salt Criteria Verified Fabric. Fig. 6. Development of Stability • Cavern Monitoring Methods. Criteria • Correlation of Laboratory, Analytical, and Field Results. IMPLEMENTATION OF STABILITY Fig. 5. Topics for Additional Study On CRITERIA FOR CAES CAVERNS Long-Term Stability of CAES Caverns in Salt Domes. In this section the implementation criteria enhancing long-term stability of CAES caverns is presented in a three- DEVELOPMENT OF SITE-SPECIFIC phase format. The three phases generally STABILITY CRITERIA follow a work plan sequence by which a CAES facility could be established and The development of a complete set of made operational. quantitative site-specific stability cri- teria requires: (1) the gathering of site- The three phases of stability cri- specific technical data; (2) an appropriate teria implementation are associated re- analysis which will yield predictive quan- spectively with: (1) site selection, (2) titative results; and ultimately, (3) the cavern system design, and (3) facility verification of criteria by agreement be- operation. Figure 7 illustrates schemati- tween predicted and monitored performance. cally the methodology whereby site- specific long-term stability criteria In this section a three-stage format could be implemented for CAES caverns in is proposed for developing site-specific salt domes. criteria. Figure 6 illustrates the three 381 Start Phase (1): Implementation of Stability Criteria in Site Selection (a) , Apply Cavern Stability Analysis and Criteria: Propose Salt Dome Site for CAES Caverns Stable stress state in salt around caverns (based on time independent analyses)? Acceptable rates of creep closure and Investigate: creep rupture over life of facility Dome Utilization history (based on time dependent analyses)? Megascopic features of geomechanical Acceptable rate of progressive failure system by geological, geophysical, due to cyclic pressure and temperature I and corehole studies. loads over the life of facility (based on [fatigue tests and associated modelling)? , (c) 7T • Apply Long Term Cavern yes no Stability Criteria: %^Return to (2a)' Relative minimum of extensive early "wild" brining *nd sulphur mining Start Phase (3): Implementation of activities? Stability Criteria for Operating Program Absence of strong evidence of ongoing dissolution of dome at the caprock - (a) salt interface? Propose Operating Program For Relatively small amounts of gas, brine, CAES Caverns and inclusions in salt stock? (*>), Competent, homogeneous rock salt from approximately -1000 to -6000 feet, and Investigate and Utilize Effects I also adequate in volumetric extent? , of Operating Program Via: yes no Numerical modelling of operating 'Return to (la)- program for stability effects. Start Phase (2): Implementation of Monitor operating effects. Stability Criteria in Site Selection (cL (a) Apply Cavern Stability Analysis and Criteria Application Propose Design Configuration For CAES Cavern System Preliminary field tests indicate good agreement between results of numerical (b) , modelling and test data? If not, reconcile results. Investigate Via: Operating program results in acceptable Laboratory and bench scale testing of cavern performance, as indicated by site specific salt. field monitoring program? Acceptable cavern performance includes measured Numerical modelling of proposed design acceptable cavern closure, creep closure configuration. . irates, and surface spa!ling rates. To 3(d) 382 From 3(c) 3. Chang, G. C, Loscutoff, W. V., and yes no Schneider, J. R., (Symposium Cochair- ^Return to (3a) men) 1978 Compressed Air Energy Stor- age Symposium, Proc. of Meeting on May 15-18, 1978, Asilomar Conference (d) Grounds, Pacific Grove, California, Continue to Monitor. Update Numerical In Press. , Stability Analysis Periodically. . 4. Rohr, H. U., Mechanical Behavior of a Gas Storage Cavern in Evaporitic Rock, Fig. 7. Implementation of Stability Proc. 4th International Symposium on Criteria 1974! JigklSr*116"1*106801- S°C" CONCLUSIONS 5. State of Louisiana, statewide Order No. 29-M, "Rules and regulations per- CAES technology, relative to long- taining to the use of salt dome cavi- term stability or air reservoir caverns ties for storage of liquid and/or gas- in salt domes and salt anticlines., cur- eous hydrocarbons, etc.", Department rently is in an early stage of investi- of Conservation, Baton Rouge, July, gation. Much work remains to be done, particularly in relating site-specific field results to results predicted from 6. Mattick, W., Weber, 0., Stys, Z. S , engineering analyses based upon test data and Haddenhorst, H. G., Huntorf-The and numerical methods. World's First 290 MW Gas Turbine Air Storage Peaking Plant, Proc. Amer Adequate numerical programs and Power Conf., Illinois Institute of testing methods exist in principle to Technology, Chicago, 111., April 1975. perform stability analyses of CAES cav- erns. At this time a substantial data Boresi, A. P., and Deere, D. U., Creep base should be developed specific to en- Closure of a Spherical Cavity in an vironments anticipated for rock salt Infinite Medium (with Special Applica- surrounding CAES caverns in domes and/or tion to Project Dribble, Tatum Salt anticlines. Verification of one or more Dome, Mississippi), for: Holmes and competing methods of analysis also should Narver, Inc., May, 1963. be completed as expeditiously as possible, with verification based on site-specific Dreyer, W. E., Results of Recent Stud- field results. ies on the Stability of Crude Oil and Gas Storage in Salt Caverns, Proc. 4th International Symp. on Salt, V II REFERENCES Northern Ohio Geological Soc, 1974, P« 65-92. 1. Johnson, K. S., and Gonzales, S., Salt Deposits in the United States Nolte, Wierczeyko, Problems Occurring and Regional Geologic Characteristics During the Sonar Logging of Storage Important for Storage of Radioactive Caverns, Vth International Symposium Waste. Report Y/OWI/Sub-7414/1, on Salt, May, 1978, Hamburg, Northern Earth Resources Associates, Inc., Ohio Geol. Soc, Proc. In Press Athens, Georgia, March, 1978. 2. Thorns, R. L., and Martinez, J. D., Preliminary Long-Term Stability Cri- teria For Compressed Air Energy Stor- age Caverns In Salt Domes, Prepared ^°^ Battelle, PNL, Special Agreement B-548O4-A-L, Prime Contract EY-75-C- 06-1830, Institute For Environmental Studies, Louisiana State University, DRAFT, August, 1978. 383 PROJECT SUMMARY Project Title: Stability and Desiqn Criteria Studies for Compressed Air Energy Storage Reservoirs-Rock Mechanics and Geology Components Principal Investigator: Howard J. Pincus Organization: University of Wisconsin-Milwaukee P. 0. Box 413 Milwaukee, MI 53201 414/963-4017, 4561, 4962 Project Goals: To evaluate the effects of pressure-temperature cycling with compressed air on underground reservoir rocks and their associated caprocks. We are concerned with changes in physical properties and in nicrostructure that can be attributed to ventilation. We seek to contribute to development of criteria for identifying and evaluating candidate sites, to identify geological parameters to be monitored in field tests, and to assist in developing monitoring systems for operating installations. Project Status: Now starting on third phase. Through September 1, 1978, 137 specimens were processed in some way; 61 of these were ventilated. Pressure-tenperature conditions were at lower end of anticipated operating range. In the phase just started, higher temperatures and pressures will be used. Contract Number: Jan. 1, 1977 - Sept. 30, 1978 - Battelle PNL Award B-37774A-K Oct. 1, 1978 - Sept. 30, 1979 - Battelle Subcontract B-21286A-K Contract Period: (1) Jan. 1, 1977 - Sept. 30, 1977 (2) Oct. 1, 1977 - Sept. 30, 1978 (3) Oct. 1, 1977 - Sept. 30, 1979 Funding Level : (1) $25,000 (2) $73,800 (3) $62,000 Funding Source: Battelle Pacific Northwest Laboratories 385 FABRIC ANALYSIS OF ROCK SUBJECTED TO CYCLING WITH HEATED, COMPRESSED AIR Howard J. Pincus Departments of Geological Sciences and Civil Engineering University of Wisconsin-Milwaukee Milwaukee, Wisconsin 53201 ABSTRACT An essential component of current investigations of compressed air energy storage in aquifer-type rocks is the characterization of changes with ventilation in pore-and- grain structure of the rocks, particularly as related to porosity and permeability. In addition to conventional microscopic examination, we are using. Fourier optics to express changes in spatial frequency and orientation of grains, pores and microcracks. Optical Fourier amplitude transforms depict the distribution of spatial frequencies and orienta- tions in their input-images. We can also express the similarity between pairs of input images by cross-correlating their transforms. Each of these Fourier optical procedures has detected some changes with ventilation in rock fabric. Differences between rocks are often greater than differences in the same rock associated with ventilation. Differences in gross physical characteristics among some rocks are consistent with contrasts in fabric. INTRODUCTION The goals of this project are: with ventilation. 1) To evaluate the effects of pres- 2) Changes in permeability asso- sure-temperature eyeling with ciated with ventilation have not been compressed air on underground detected. reservoir rocks and their asso- ciated caprocks. 3) Changes in porosity and heat capacity of St. Peter Sandstone, asso- 2) To contribute to the development ciated with ventilation, have not been of criteria for identifying and detected. evaluating candidate sites. 4) Microscopic examination in re- 3) To identify geological parameters flected light has revealed no diagnostics to be monitored in field tests. for distinguishing with confidence be- tween ventilated and unventilated rocks. 4} To assist in developing monitor- ing systems for operating in- 5) Heating of Berea Sandstone and stallations . Bedford (Salem) Limestone to 260°C. for five days has resulted in decreases in We have been concerned chiefly with the value of Young's modulus. Compressive detecting and measuring changes in physi- strength of the Berea also decreased. cal properties and in microstructure that can be attributed to ventilation. In the ventilation experiments carried out, typical air temperatures at This paper is concerned with some of the specimen were 110°C., with confining the methods and results of studies of the pressures up to 820 kPa. microstructure of some candidate rocks. It will perhaps be useful here to outline The most striking result of the work results obtained so far in other phases described and partially documented in of our ^ this paper is that initial variations in fabric within the same unit, e.g., the 1) Values of Young's modulus and St. Peter Sandstone, appear to exceed the compressive strength of sandstones studied changes resulting from ventilation under so far do not show systematic changes the experimental conditions used so far. 386 Observation of the St. Peter in iarge ex- treated surfaces of rock specimens and posures indicates considerable variabili- their optical transforms are shown in ty in lithology and structure. Figs. 6 and 7. Note the horizontal elongation of the Fig. 7 transform, cor- ROCKS STUDIED responding to the input's vertical "grain". Host of the specimens studied so far are St. Peter Sandstone, collected in an The scanning of transforms yields active quarry in south-central Wisconsin profiles that greatly facilitate compari- (Fig. 1). Experimental work has also been son. Earlier work ' has shown changes done on Platteville Limestone and Jordan associated with ventilation that seem to Sandstone, also collected in south-central indicate some transport of fines in some Wisconsin, and on two members of the U.S. rocks when airflow is unidirectional; Bureau of Mines-A.R.P.A. Standard rock this is manifested as, for example, larger suite, Berea Sandstone and Bedford (Salem) high spatial-frequency content in the Limestone. transform of the low-pressure end of the specimen. Specimen blocks weighing 13-40 kg. have been collected and cored (Figs. 2,3). Optical correlation provides another The time-based eyeling system shown in means for quantifying comparisons of Fig. 4 has been used to ventilate all of spatial information. the specimens for which results are pre- sented in this paper. This equipment has As we have developed and applied been eyeled at up to almost 1000 cycles in this method in our work, three genera- four days with mean air velocity through tions of optical transforms are produced. some specimens exceeding IS cm/s. First, transforms of the two images to be compared are generated. These two trans- FABRIC ANALYSIS forms are then used as an input, side-by- side, and their joint-transform is genera- We are concerned with the size, shape, ted (Fig. 8, top). The transform of the arrangment and orientation of grains, joint-transform is then generated (Fig. 8, pores, and microfractures, with the dis- bottom), and the intensity of its first- tribution of cement and other fine-grained order diffraction spots is a measure of material in the pores, and with changes in the degree of similarity of the spatial time (or ventilation) of any of the fore- information recorded in the first-genera- going. Microscopic analysis, which is tion transforms. These cross-correlations being continued, has so far not yielded are normalized by dividing the raw cross- useful diagnostics. correlation reading by the geometric mean of the autocorrelations of each of the To augment fabric analysis, we have two first-generation transforms. Some of been using a Fourier-optical approach to the results obtained so far are presented spatial analysis, supported by a recently- in Table 1, adapted from our last annual developed method of optical correla- report1. tion1'-'3'4 The top half of the table represents Utilizing coherent light and suitable data for two unidirectionally ventilated optics, the two-dimensional Fourier ampli- specimens and the bottom half presents tude transform of a transparent image of data for two specimens, each with one a rock surface is produced as a diffrac- ventilated and one unventilated input. tion pattern. The directions of radii to points in the transform are perpendicular In the upper half, the correlation to the directions of the corresponding in- is smaller between the high- and low- put elements; the radial coordinates of pressure ends of the red St. Peter than dots in the pattern vary linearly with the between the high and low ends of the gray spatial frequency of the corresponding in- St. Peter. In both the field and the put linear elements. Some basic Fourier laboratory, the red St. Peter crumbles transform relations are presented in Fig. more easily than the gray. 5, taken from a pictorial digital atlas-* intended for seismologists. In the bottom half of the table, the correlation between unventilated and Photographic images of specially ventilated red St. Peter is smaller than 387 between corresponding yellow St. Peter ting result is that the introduction of specimens. In the field, the yellow is lots of fines into pore spaces seems to somewhat more durable than the red. have a marked effect on the correlation coefficients (cxe vs. cxd), a result also Table 1. Optical correlation of venti- achieved with other inputs not presented lated specimens. here. Table 2. Normalized optical correlations- Normalized Cross- Lath-like artificial inputs (Fig. 9). Rock Correlation a b c d e St.Peter(gray) High vs. low 0.97 BE-I-18 pressure(gray) a .13 .13 St.Peter(red) High vs. low 0.88 b .34 .37 .18 BE-I-25 pressure(red) c .47 .28 St.Peter(gray and High-gray 0.93 red) BE-I-1S,25 vs. high-red d .21 St.Peter(gray and Low-gray 0.83 red) BE-I-18,25 vs. low-red CONCLUDING REMARKS St.Peter(red) Vent. vs. 0.82 We are continuing our work with BE-I-25 unvent. (red) artificial inputs, investigating the effects of several key fabric variables, St.Peter(yellow) Vent. vs. 0.91 singly and in combination. Concurrently, BE-I-22 unvent. (yellow) we are continuing analysis of rock speci- mens. Recently we have completed con- St.Peter(red and Red vs. yellow, 0.85 struction of ventilation equipment which yellow) BE-1-25,22 unvent. will permit use of air at temperatures and pressures near the probable upper St.Peter(red and Red vs. yellow, 0.84 limits of an operating system. yellow) BE-I-25,22 vent. The assistance of Dr. Karl Scheiben- Further, correlation between red and graber and Mr. Jack Hopper in preparing yellow unventilated specimens is about this paper is acknowledged with thanks. the same as between red and yellow venti- lated specimens; both of these correla- Support for this work has been pro- tions indicate less spatial similarity vided by D.O.E. through B.P.N.L., and than is indicated by the correlation the University of Wisconsin-Milwaukee. between yellow unventilated and ventilated. REFERENCES Variations within the same unit, say the St. Peter, may exceed changes resulting 1) Pincus, H. J. (1978),FY 1978 Progress from ventilation. Report to BPNL and DOE, 57 pp. OPTICAL CORRELATION OF 2) Smith, G. C. et al. (1978), FY 1977 ARTIFICIAL INPUTS Progress Report, PNL-2443, UC-94b, Ch. 4, Laboratory Testing. Interpretation of the optical corre- lation coefficients obtained so far, re- 3) Pincus, H. J. (1978), U.S. Natl. Comm quires further investigation. We are Rock Mech., Proc, 19th U.S. Symp., currently determining relative sensitivi- v. 1, 215-220. ties to contrasts in grain size, grain shape, grain arrangement, preferred 4) Pincus, H. J. (1978), Int. Assn. Eng. orientation, and sorting. Geol., Proc, III Cong., Sec. Ill, v. 2, 107-116. A sample of the work underway and partly reported* is shown in Fig. 9. 5) Peterson, R. A. and Dobrin, M. B. Some of the normalized correlation coeffi- (1966), "A Pictorial Digital Atlas", cients are shown in Table 2. One interes- United Geophysical Corp., Pasadena. 388 \ •-;/• Fig. 3 Cored specimens in several stages of preparation. Host spec- Fig. 1 Faulted St.Peter Sandstone i in quarry near Verona,Wisconsin. imens are St.Peter 3andston«,5 *flu». Scale: Haraner in foreground. in diameter. Rock breaks out in large blocks which disintegrate quickly. Darker beds are red. Fig.2 Specinen blocks weighing 15- Fig. 4 Time-based cycling system. kO kg. Scale is 15 en. long. Specimen chamber is in front of Two of the blocks shown have been three circular pressure gauges, drilled for NX(54*a) cores. right-center. Air is heated in tall column at extreme right. At left, three pyroneters(rectangular). 389 TIME DOMAIN FREQUENCY DOMAIN OR FREQUENCY DOMAIN OR TIME DOMAIN - 1111 I 111 J__L WINDOW FUNCTION If -•- J_L b) nnnnfjinnnn o nnifinn c) Fig. 5 Fourier transforms of form equivalents are in the right- truncAted functions. The window hand column ; the * denotes conv- functions correspond to apertures olution, by which the transform of of illumination. In a),b),and c) the truncated function is obtained in the left-hand column, the inf- from the functions above. The ir.ite function times the window square wave in b),left side, corr- function yields the truncated esponds to a profile across white function iwaediately below. Trans- and black lines of equal width. 390 Fig. 6 Top. Light gray Berea Sand- Fig. 7 Top. Yellow and pink St. stone. Specimen pores are filled Peter Sandstone.Same treatment and with whits paint.Short bar: O.kmn. scale as in Fig. 6. Bottom.Optical Bottom. Optical transform of above. transform of abovo(O.D.A.23?3). (O.D.A.2372).Scale bar:2.5 lines/mm Same scale as Fig.7. 391 Fig* 8 Joint-transform optical correlation. Top. Joint-transform ot two optic- al transforms. Regularity and dis- tinctness of vertical bands are greater the more similar are the two transforms being compared. Bottom. Optical transform of the Fig.9 Top. Artificial inputs sim- joint-transform above. The bright- ulating lath-like grains. From ness of the two first-order diff- a to c preferred orientation de- raction dots, marked by inverted creases and porosity increases. V's, varies with the regularity The large grains in c.d.and a and distinctness of the vertical are identical, with fine materials bands in the joint-transform. in the pores increasing from c to e , Bottom. Optical transforms of the inputs above.Mote that the number of radiating bands increases from the transforms of a to c, and that the high spatial-frequency content increases from the transform of c to that of e . 392 PROJECT SUMMARY Project Title: Numerical Modeling of Behavior of Caverns in Salt for Compressed Air Energy Storage (CAES) Principal Investigator: Shosei Serata, Ph.D. Organization: Serata Geomechanics, Inc. 1229 Eighth Street Berkeley, CA 94710 (415) 527-6652 Project Goals: (1) To identify the practical ranges of the variables that are important to the geomechanical design of CAES cavities in rock salt. (2) To develop procedures for the computer simulation of the effects of these variables on the behavior of CAES cavities. (3) To perform parametric studies of the behavior of typical solution cavities in domed salt formations. (4) To develop general stability and design criteria for CAES cavities having projected lifetimes of about 50 years. (5) To identify areas of further research required to arrive at satisfactory stability and design criteria for CAES caverns in rock salt. Project Status: (1) A literature and data base review has been completed, although current literature is constantly checked for new developments relevant to CAES projects. (2) A limited laboratory investigation of the behavior of rock salt under cyclic loading conditions is in progress. A number of cyclic and static tests have already been completed. (3) Preliminary parametric studies of cavern depth, cavern air pressure, excess in situ lateral stress, and material prop- erties of rock salt are in progress. Preliminary parametric ranges have been established but will be modified as the program progresses. Contract Number: Special Agreement No. B-54809-A-P Prime Contract No. EY-76-C-06-1830 Contract Period: Mar. 7, 1978 - Feb. 28, 1979 Funding Level: $85,000 Funding Source: Battene Pacific Northwest Laboratories 393 NUMERICAL MODELING OF BEHAVIOR OF CAVERNS IN SALT FOR COMPRESSED AIR ENERGY STORAGE Thomas E. Cundey, Senior Math Analyst and Shosei Serata, President Serata Geomechanics, Inc. 1229 Eighth Street Berkeley, California 94710 ABSTRACT Finite element techniques are being used to establish general design criteria for caverns in salt formations that could be used for compressed air energy storage (CAES). The three types of parameters being studied are geological (in situ stresses, material properties), geometric (cavern dimensions, cavern spacings), and operational (air pressures, air temperatures). The numerical studies are being supple- mented by a laboratory testing program developed to determine the effect of a cyclic loading environment on the material properties of rock salt. The five phases of the overall program are described, and the work com- pleted to date is discussed. Preliminary ranges of the most significant parameters are presented. Areas in which additional research is necessary are identified, with particular regard to laboratory testing. INTRODUCTION might influence the long-term behavior of CAES caverns in this Among the unknowns that affect material. the technical and economic feasi- bility of compressed air energy In order to develop suitable storage (CAES) in rock salt are the stability and design criteria for responses of underground openings CAES caverns in rock salt, a to fluctuations in pressure and multi-phased program of study has temperature that would occur in the been initiated. The first phase, daily charge and discharge opera- a state-of-the-art survey and tions of CAES plants. The thermo- formulation of preliminary mechanical stresses generated by stability and design criteria, is such fluctuations and by the ele- currently underway at Louisiana vated storage temperatures might State University (LSU). The cause mechanical damage to the second phase, which is the subject walls of CAES caverns, thereby of this paper, will use numerical affecting their overall stability. modeling of the behavior of CAES The material properties of rock caverns in rock salt. The third salt, including creep behavior, and fourth phases will be a have been studied previously. laboratory investigation of the However, many uncertainties remain material properties of rock salt concerning the responses of rock relevant to CAES caverns and field salt to very long-term loading, studies of CAES cavern behavior. cyclic loading, thermal variations, and volumetric changes, and the ways in which these responses 394 — PROJECT DESCRIPTION underground openings in bedded salt formations was a major con- The five principal research sideration in giving priority to activities in this second phase of solution cavities in salt domes. the program and the studies to be undertaken within each activity The data available from are described in the following existing sources is insufficient sub-sections. to establish satisfactory general criteria for the effects of cyclic IDENTIFICATION OF RANGES OF loading on the behavior of rock VARIABLES salt. A limited program of lab- oratory testing is therefore There is considerable uncer- being undertaken to investigate tainty about some aspects of the these effects. This work is in- mechanical and thermal behavior of tended to give qualitative and rock salt surrounding CAES caverns. some quantitative information to Therefore, for purposes of numeri- supplement existing knowledge of cal simulation studies, a practical fatigue behavior of earth materi- range of values must be established als . A small quantity of rock for each variable that has signi- salt has been obtained for these ficant influence on the behavior tests. The basic properties of of the caverns. The following the salt are derived from tri- sources of data.are being used to axial tests and triaxial creep develop ranges for these variables: tests on some of the samples. (1) the published literature; Other samples are subjected to (2) the data base established by different numbers of cyclic re- the LSU state-of-the-art survey; versals of deviatoric stress under (3) operational criteria developed confining pressures typical of by previous CAES studies; and those expected in rock salt sur- (4) data and experience accumulated rounding CAES caverns. After by Serata Geomechanics, inc. (SGI) cyclic loading, the elastic and in previous studies of the behavior creep properties of the salt are of salt caverns used for the stor- measured and compared with the age of compressed air, natural gas, properties of the salt samples LPG, and crude oil. that were not cyclically loaded to determine the effects of cyclic The variables for which para- loading on salt properties. metric ranges are being investi- gated include the following: COMPUTER SIMULATION TECHNIQUES cavern geometry, cavern depth, cavern separation (in multiple The primary computer simula- cavern systems), in situ earth tion studies are being performed stresses, the rheological proper- using SGI's finite element com- ties of rock salt, the permeabili- puter program (REM). This program ty of rock salt, cavern operating has been developed and used to mode (compensated or uncompensated), study the time-dependent behavior and cycles of operating tempera- of underground openings in con- tures and air pressures. Major ventional and solution mines and emphasis is being placed on uncom- the stability of storage cavities pensated caverns created in salt for oil, natural gas, LPG, and domes by solution mining. The compressed air. highly site-specific nature of the geologic and geometric parameters To study the effects of involved in the numerical analysis thermo-mechanical stresses and of of abandoned conventionally-mined temperature changes on material 395 properties, the finite element being developed. The criteria program HEAT, developed at the being developed include cavern University of California, Berkeley, geometry and depth, in situ earth is being used. This program stresses, cavern spacing, material analyzes steady-state and transient properties of rock salt, the per- linear heat conduction in plane and meability of rock salt, the magni- axisymmetric regions. Output from tudes of operating pressures and the HEAT program provides insights temperatures, and the maximum flow into the magnitude and distribution and heat transfer rates. of rock temperatures which are used as input to the geomechanical RECOMMENDATIONS FOR ADDITIONAL analyses to be performed using the RESEARCH REM program. Using the results of the cur- PARAMETRIC STUDIES rent research and other experi- ences in the design of underground The parametric studies are cavities, additional geotechnical designed to investigate the effects research necessary to develop of variations in the principal geo- satisfactory criteria for the logical, geometric, and operational stability and design of CAES variables on cavern behavior and caverns in rock salt will be stability. The principal variables identified and recommended. These to be studied using the computer recommendations will include guide- simulation models include cavern lines for laboratory testing of geometry and depth, material prop- rock salt and field studies of erties of the rock salt, in situ CAES caverns. stresses, and operating conditions (e.g., magnitudes and rates of RESULTS change of pressure and temperature). LITERATURE REVIEW The range of each variable used in the parametric studies is A review of the published based on the results of the litera- literature was conducted at the ture and data-base review. However, outset of the research in hope if the numerical analyses indicate that some of the environmental that other variables or variable conditions surrounding CAES caverns ranges should be investigated, the in salt had been investigated analysis program will be modified previously. Because much work had appropriately. A limited number of been done by the authors and other simulations will also be conducted investigators in studying the to investigate modes of cavern static mechanical properties of failure induced by extreme condi- rock salt with particular emphasis tions such as very high or very on creep, this area was de-empha- low cavern pressure, elevated air sized in the literature review. temperature, extreme cavern dimen- Major emphasis was instead placed sions, and small cavern separation in three other areas: the effect distances in multi-cavern networks. of cyclic loading on the behavior of rock salt, the permeability of CRITERIA FOR STABILITY AND DESIGN rock salt under various stress states, and the effect of elevated Drawing upon the results of temperature on rock salt behavior. the studies at LSU and the computer simulation studies described above, Cyclic Loading. No literature general criteria for the stability could be found on the effect of and design of CAES caverns are cyclic loading on rock salt 396 behavior, and in fact little work secondary stage with has been done on the behavior of constant creep rate, and any rocks under cyclic loading con- finally a stage in which ditions. Soil materials under the strain rate increases cyclic loading conditions, on the until failure occurs; other hand, have been studied ex- tensively in recent years, partic- (4) The cyclic creep curve ularly to advance the art of was invariably bounded designing earthquakt;-resistant by the complete stress- structures and for offshore oil- strain curve of the drilling platform foundations. The respective rock. most comprehensive laboratory in- vestigation of rock behavior under Only granite was tested in cyclic loading conditions was done cyclic triaxial compression, and "by Professor B. C. Haimson and his its behavior was similar in every colleagues at the University of respect to that under cyclic uni- Wisconsin. Although the differ- axial compression, in addition, ences in the behavior of rock salt an increase of fatigue strength and the four hard rock types with increasing confining pressure (marble, limestone, sandstone, and was observed. Other cyclic tests granite) used in the Wisconsin were conducted under uniaxial ten- study are appreciated, it is never- sion and uniaxial compression- theless useful to summarize some of tension, but those loading condi- Professor Haimson's findings. tions are not relevant to CAES caverns and will not be discussed. In cyclic uniaxial compression tests on all four types of hard Attewell and Farmer proposed rock, the following observations a hypothesis to explain the resul- were made: tant deformation of a fine-grained dolomitic limestone in cyclic uni- (•1) All rock types showed a axial compression which appears to fatigue strength of 60 be compatible with Haimson's find- percent to 80 percent of ings and is based on strain-energy the uniaxial compressive dependent crack propagation. Above strength; a threshold stress level at which cracks are initiated, the deforma- (2) Stress-strain curves tion from successive sub-failure showed large hysteresis load cycles is cumulative, with in the first cycle, failure occurring when the strain decreasing hysteresis in energy stored in the rock exceeds the next cycles, nearly a critical level equivalent to constant hysteresis loops failure under non-cyclic loading. in the following cycles, and a reopening of the Permeability. The permeability of loops in the final cycles rock salt under various stress prior to failure; states was studied by Chia-Shing Lai at Michigan State University (3) Three characteristic in 1971. A special high pressure stages of cyclic creep triaxial cell was designed and con- were observed in all structed for the permeability four rock types: a pri- tests. Lateral and axial loading mary stage, in which was controlled by two independent strain increases from hydraulic systems. Uniform flow cycle to cycle at a de- of the permeating fluid (kerosene) celerating rate, a was applied to the specimens under 397 a constant fluid pressure. All relating creep rate with tempera- tests were conducted at room tem- ture, stress, and time: perature . 3.2 x 10-36 fc-0.65 Numerous tests were conducted using a wide range of values of where d is the pillar strain rate mean stress and octahedral shearing (10~6 per hour), T is the tempera- stress, and empirical relationships ture (degrees Kelvin), 398 Other samples are being loaded Cavern air pressure = 0 kN/n»2 to cyclically to determine the effect 15,215 kN/m2 (0 atm to 150 atm) of cyclic loading on the material properties as determined in the Excess in situ lateral stress = static tests. -10,978 kN/m2 to +10,978 kN/m2 (-1,591 psi to +1,591 psi) Although a number of the cyclic tests have been conducted, Octahedral shear strength of salt = the analysis of the test results 3,795 kN/m2 to 5,175 kN/m2 has not yet been completed. One (550 psi to 750 psi) series of tests was conducted at a confining pressure of 17,250 Other salt properties = Weak to kN/m2 (2,500 psi) and axial load strong varying between 17,940 kN/m (2,600 psi) and 53,130 kN/m2 In the first studies the (7,700 psi) at a frequency of 1 behavior of a horizontal section cycle per hour. This extremely of a single vertical cavern in an high deviatoric stress, equivalent infinite salt mass was investigated to about 90 percent of the tri- using an axisymmetric model. A axial strength at this confining set of parameter base values were pressure, is unrealistic to expect used: in a CAES environment but was chosen to provide evidence of Cavern depth = 610 m (2,000 ft) fatigue characteristics of rock salt. A second series of tests, Cavern air pressure = 5,070 kN/m2 still in progress, employs a con- (50 atm) fining pressure of 17,250 kN/m2 (2,500 psi) and axial load varying 2 Excess in situ lateral stress = between 21,840 kN/m (3,165 psi) 4,391 kN/m2 (636 psi) and 29,190 kN/m2 (4,230 psi) at a frequency of 2 cycles per hour. Octahedral shear strength of salt = After a predetermined number of 4,140 kN/m2 (600 psi) cycles (which varies from test to test) are completed, the samples Other salt properties = Weak are tested in triaxial compression creep tests, and the material One parameter was varied through properties therein determined will the ranges described above while be compared with those of the non- che others held their base values. cyclical ly loaded samples. Cavern air pressure in these first analyses was static rather than PRELIMINARY NUMERICAL MODELING cyclic. Final graphs of the out- STUDIES put data from these runs are not complete. Parametric ranges for some of the principal variables to be in- FUTURE WORK vestigated have been established based on the results of the liter- Once the preliminary para- ature review, previous CAES studies, metric studies of cavern depth, and the authors' experiences in cavern air pressure, excess in salt cavern design, as follows: situ lateral stress, octahedral shear strength of salt, and other Cavern depth = 153 m to 1,529 m salt properties are completed, the (500 ft to 5,000 ft) single cavern models will be modi- fied so that other effects can be investigated. Cyclic cavern air 399 pressure will be studied, and pro- Although our numerical gram HEAT will be used to compute studies are far from completion, it temperature distributions around is evident that much work has to be single vertical caverns. The done to evaluate possible coupling degree of air permeability will be effects among the key operational assessed based on Lai's studies, parameters—cyclic air pressure and material properties of salt and air temperature. As discussed near the cavern boundary will be earlier, no studies have been adjusted to simulate the deterior- reported on the effects of cyclic ation of the boundary salt. All loading on salt, but there is these analyses will employ the literature on rock salt permeabili- geometry of a horizontal section ty and temperature effects on salt of a single vertical cavern in an properties. It is hoped that infinite salt mass, as this is the laboratory investigations of the most economical geometry. combined effects of elevated tem- perature and cyclic loading can be The next step after the para- completed and the numerical models metric studies of the operational re-evaluated after these labora- and geological variables is a tory investigations have been parametric study of geometric completed. Furthermore, final variables. The first of these verification of the models will be cavern diameter and cavern requires the collection and length; and again a single vertical analysis of field data from a CABS cavern in an infinite salt mass demonstration site. will be analyzed axisymmetrically. but the entire cavern will be SELECTED REFERENCES modeled rather than just a horizon- tal section. Single-cavern Attewell, P.B. and I.W. Farmer, analyses will be followed by 1973. Fatigue Behavior of Rock, double-cavern and other multi- international Journal of Rock cavern simulations, so that the Mechanics and Mining Sciences, effect of separation distance Vol. 10, pp. 1-9. between adjacent caverns can be investigated. Although the Bradshaw, R.L., et al., 1968. realistic ranges of most of the Properties of Salt important in variables will have been estab- Radioactive Waste Disposal, lished by this time, a number of Special Paper No. 88 of the extreme cases will also be run to Geological Society of America, gain insight into potential cavern pp. 643-658. failure modes. Haimson, B.C., 1973. Mechanical The results of all the numeri- Behavior of Rock under Cyclic cal studies will be carefully Loading, Final Technical Report studied to determine general to the Bureau of Mines. Contract stability and design criteria for H0220041. CAES caverns. The numerical re- sults will be supplemented by the Lai, Chia-Shing, 1971. Fluid results of the literature study. Flow through Rock under Various Professor Thorns' state-of-the-art Stress States, Ph.D. Thesis, report, and our experiences in Michigan state University. salt cavern design. Recommenda- tions for future work in both the Serata, S., T.E. Cundey and C. Lai, laboratory and the field will be 1978. Rock Mechanics Problems made. in the Design of Solution Cavi- ties and use of Abandoned Mines 400 for the Storage of Compressed Air Energy, Proceedings of the CAES Technology Symposium, Asilomar, California. Taylor, R.L., 1975. *HEAT*, A Finite Element Computer Program for Heat-Conduction Analysis, Report 75-1 for the Civil En- gineering Laboratory, Naval Construction Battalion Center, Port Huename, California, Purchase Order N62583/75-M-Y695. 401 PROJECT SUMMARY Project Title: Compressed Air Energy Storage Studies Principle Investigator: Dr. G. T. Kartsounes Organization: Argonne National Laboratory 9700 South Cass Avenue T-12-3 Argonne, IL 60439 312/972-7695 Project Goals: The objective of this program is to extend the technology base of the the compressed air energy storage (CAES) concept for application to electric utility peak-power generation. During FY 1978, three studies were conducted. The objectives of the studies were: to develop a gen- eral design optimization procedure for CAES plants having aouifer stor- age reservoirs; to evaluate the performance and capital and'operating costs of possible turbomachinery options for CAES plants; and to evaluate the performance, cost, and manufacturing feasibility of a unique reci- procity expander/compressor which could have a major impact on CAES viability. Project Status: Due to decreased funding for FY 1979, no further work will be done on the development of the design optimization procedure for CAES plants having aquifer reservoirs. The evaluation of turbomachinery options has been completed, and no further work is anticipated. A preliminary evaluation of the reciprocating engine has been completed, and further evaluation will continue through FY 1979. Contract Number: W-31-1O9-ENG-38 Contract Period: FY 1978 Funding Level: $310,000 BO * Funding Source: Department of Energy, Office of Conservation, Dtviston of Energy Storage Systems * Included in this project are the following: George T. Kartsounes and Choong S. Kim, "Evaluation of Turbomachinery for Compressed Air Energy Storage Plants," page 416; George T. Kartsounes and James G. Daley, "Evaluation of the Use of Reciproeating Engines in Compressed Air Energy Storage Plants," page 425. 403 THE DESIGN OPTIMIZATION OF AQUIFER RESERVOIR- BASED COMPRESSED AIR ENERGY STORAGE SYSTEMS Frederick W. Ahrens Energy and Environmental Systems Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439 ABSTRACT The application of a general Compressed Air Energy Storage (CAES) power system design optimization methodology to the class of CAES plants having aquifer air storage reservoirs is discussed. The resulting procedure incorporates performance and economic models for the aquifer reservoir, wells, piping, and air compression system. Its use allows identi- fication of designs which minimize the subsystem power generation cost (mills/kWh), while satisfying constraints related to the geology, equipment, and utility load curve. The design specification resulting from the optimization procedure includes: land area to be purchased, well depth, number of wells, well spacing, wellbore diameter, main pipeline diameter, required compressor system power and discharge pressure, and required compres- sion time durations for each day of the week. A capital and operating cost summary for the optimum design is a final output of the procedure. This paper reviews the models and constraints incorporated in the optimization procedure. Although the basic framework is well-developed, some refinements or additions to the modeling may be necessary to improve the results; these possibilities are discussed. Results of case studies are included in the paper in order to illustrate the power and potential economic impact of the techniques described, to demonstrate some of the economic tradeoffs which occur' in the optimal design of aquifer reservoir-based CAES systems, and to show the influence of certain cost para- meters . INTRODUCTION A major portion of the Department of conditions. Energy research program on Compressed Air Energy Storage (CAES) is devoted to addres- When aquifers are considered for CAES sing air reservoir concerns. In the case reservoir application, the number of design of constant-pressure (hard rock) caverns, parameters to be selected is much larger it is quite easy to design a reservoir and, in addition, there are many constraints which satisfies the planned operating imposed by the operating cycle, the inter- cycle of the CAES plant, once the turbine action with aboveground machinery, and the system air supply requirements have been geological characteristics. These factors specified. It is then straightforward to give a great incentive to the careful ex- include the effect of cavern costs in eco- ploration of aquifer system design options, nomic studies of turbomachinery options so that the economic benefit of the plant (e.g., see Ref. 1). The situation is some- can be maximized while simultaneously in- what more complicated in the case of con- suring its long-term capability to meet the stant-volume (usually salt cavity) reser- plant operational criteria. The primary voirs. For these reservoirs, the peak goal of the work reported here has been to storage pressure (related to the amount of develop an appropriate, comprehensive, cushion air) must be selected, which in- means for performing these design studies. volves finding the economic balance between The performance, economic and design opti- air compression and cavern volume costs. mization considerations which form the The relative design simplicity of hard rock bases of the design procedure are described and salt cavity reservoirs has resulted in and illustrated with example case study a DOE-sponsored research emphasis on the results in subsequent sections. long-term stability or reliability of the reservoirs undergoing CAES plant operating 404 DESIGN STRATEGY PHILOSOPHY Because of the complexity of an aqui- fer reservoir-based CAES system, both in terms of subsystem interactions and design constraints, it was decided to orient the formulation of the design procedure toward the utilisation of generalized, computer- oriented techniques for solving nonlinear, constrained optimization problems.2 With this approach, a basic framework for system design can be developed so that new improve- ments in the technical or economic models can be easily incorporated. Implicitly, the choice of basing the design procedure on modern optimization techniques reflects Fig. 1. Typical CAES Power System a recognition of the confusion and inade- quacy which could result from applying the Subsystem 2 is the power generation train. more traditional "parameter study" approach The motor/generator and the utility grid to a problem of this magnitude. are incorporated in the third group (sub- system 3). It is important to note that SYSTEM CONSIDERATIONS this particular decomposition is general, in the sense that it is not dependent upon In principle, and probably in practice, the internal design of any particular sub- the design optimization of an entire CAES system. Furthermore, it minimizes the num- system could be handled as one large prob- ber of coupling variables. That is, the lem, especially if the models employed for interactions of subsystems 1 and 2 with individual system components were not too subsystem 3 are dependent on only one coup- detailed. However, consideration of a ling "variable" — the utility load cycle. general formulation led to the development The interactions between subsystems 1 and of an advantageous decoupling strategy 2 (the ones of principle concern to the which enables separate optimization of par- plant designer) are dependent on only three ticular subsystems without compromising coupling variables — the inlet pressure to the optimal system design. Each of these the power generation train (pt.), the spe- suboptimizations is, of course, simpler to cific air mass flow rate (&') and the util- perform than that of the full problem. ity load cycle. In broad terms, a CAES power system The criterion for optimum design is comprises the following: the air compres- chosen to be the total normalized cost (C) sion train (compressors, intercoolers, af- of the system (i.e., cost per unit of elec- tercooler); compressed air piping; air tricity generated by the CAES power plant). storage reservoir (any type); power genera- This total cost is the sum of the individual tion train (e.g., turbines, combustors, re- subsystem normalized operating costs.* The cuperator) ; reversible motor/generator and costs have to be minimized subject to vari- the utility grid. Although the utility ous performance and technical constraints. grid is not physically part of the CAES The implication for CAES plant design is plant, this interaction should be consid- that, for a given utility load cycle, a sub- ered in designing the plant, since the de- optimization of subsystem 1 would provide sign (cost) of the plant can influence the minimum subsystem operating cost (c9) utility usage (operating cycle). Conver- and values for the corresponding subsystem sely, the utility load cycle affects the design variables, as a function of the plunt design (i.e., a coupling exists). coupling variables — pt and &'. Similar For the purpose of design optimization the optimization for subsystem 2 would yield overall system can be decomposed into three subsystems (see Fig. 1). The first subsys- tem (subsystem 1) comprises the air com- pression train, the main piping and air Typically the normalized operating costs distribution system and the air storage include fuel costs, maintenance, charge reservoir. rate on capital investment, etc. 405 C? (the minimum operating cost of subsystem For the purpose of analysis, it helps 2) and its optimum design, as a function of to make a distinciton between edge-water the coupling variables only. Finally, the and bottom-water reservoirs. Edge-water sum of c!jl and C? can be minimized by inspec- reservoirs, shown in Fig. 2, are character- tion to determine the optimum values of the ized by relatively thin formations, a cap- coupling variables, the minimum plant cost rock of appreciable dip, an underlying im- (C*) and the optimal plant design. The permeable layer, and water driven to the process can obviously be expanded (in prin- edge of the field during bubble development. ciple) to include variations in the utility load cycle and consideration of the resul- ting economic benefits or penalties to the utility. A noteworthy corollary of the approach described is that changes in the design options considered (e.g., turbines vs. piston expanders) in one subsystem do not invalidate the optimization results for the other subsystem. The remainder of this paper is confined to consideration of the design of a particular variety of subsystem 1 — one with an aquifer reservoir. AQUIFER RESERVOIR TECHNICAL CONSIDERATIONS Fig. 2. Edge-Water CAES Aquifer The design of aquifer reservoirs for Reservoir CAES requires integration wif.h the charac- teristics of the aboveground machinery, In bottom-water reservoirs, depicted in piping, and the utility load cycle. The Fig. 3, a water-air interface lies in a design also depends greatly upon site-spe- nearly horizontal plane beneath the air cific geol gical properties like porosity, bubble. This commonly occurs in thick discovery pressure, permeability and formations. A characteristic unique to threshold pressure.3"* Some of these in situ bottom-water reservoirs is the phenomenon properties enter into the flow performance; of water coning. Because the bottom-water others impose design constraints. Compli- reservoirs involve more design variables cations are introduced by way of distribu- and constraints, they have received the ted flow resistance, formation heterogene- greatest attention in the present optimi- ities and possible two-phase flow of water zation study. and air. Due to the complexities, however, aquifer reservoirs appear to offer a signi- AQUIFER-RELATED DESIGN CONSTRAINTS ficant potential for economic optimization. As related to CAES systems, potential For an underground porous formation constraints imposed by the aquifer charac- to be suitable for storing compressed air, teristics have been discussed in Refs. 3 it should have certain structural features. and 4. These constraints were largely Suitable aquifers are usually in the ap- identified from experience in natural gas proximate! shape of an inverted saucer. The storage. The DOE-funded work in progress top consists of a tight porous caprock, on aquifer reservoir stability may result saturated with water. The interfacial in identification of additional ones. The property of the air-water system in these constraints presently appearing to be im- tight pores does not permit the flow of portant for inclusion in the CAES design air. Thus, the dome shape will prevent procedures are as follows. any lateral or vertical migration of com- pressed air. The compressed air is con- Air Bubble Sl^e. After growing the air tained in the pores of the rock between bubble to the desired equilibrium size, the caprock qnd the bottom layer of water further growth or shrinkage due to the and/or rock. In aquifers, the adjacent daily variations in pressure is to be nul- water moves under an applied pressure lified. This concern is reflected in two gradient and therefore requires careful related constraints. First, the total mass monitoring to ensure zero net movement of air stored during a weekly cycle should over a period of time. equal the total mass removed. Second, for 406 TOP PROJECTION optimization to be determined. This infor- Aid)- PROJECTED AREA mation includes the total bubble volume as OF AIR BUBBLE a function of bubble thickness (measured at tbe apex of the dome) and the spill point (maximum bubble thickness for which the air will remain trapped by the caprock). The °o°o corresponding constraints are that the area o o o ooo occupied by the well-field must not exceed o o o the projected bubble area and the bubble size must not exceed the spill point volume. BOUNDARY OF THE ACTIVE REGION (Aoct) Water Coning. The problem of water coning in bottom-water reservoirs means that, for given reservoir conditions and well pene- tration depth, a critical flow rate of air exists above which air cannot be withdrawn from the reservoir without simultaneous %m INJECTION/ production of water. The critical flow WITHDRAWAL rate is extremely sensitive to in situ WELLS reservoir heterogeneities. It is known that the presence of an impermeable bar- TE THAT PARTIAL rier like a shale streak below the well H IS THE MEAN UTILIZATION OF would drastically inhibit bottom-water HEIGHT OF THE RESERVOIR VOLUME DOTTED IS ALLOWED from coning into the well. The phenomenon FIGURE of water coning has been studied exten- IMPERVIOUS ;APROCK sively in the past under the assumption of steady state flow, but an order of magni- tude estimation for CAES applications (short discharge time) indicates that a non-steady analysis is required to ade- ^WATER LEVEL quately determine the maximum well pene- tration that permits withdrawal of com- Fig. 3. Bottom-Water CAES Aquifer pressed air without co—production of water. Reservoir Little attention has been given to this situation in the literature. Therefore, long-term constancy of bubble size, a pres- the coning height is presently treated as sure schedule having the average weekly a parameter in the design procedure. The pressure (corresponding to the mean mass intention is to calculate the cost and of air in the bubble during the week) equal performance sensitivity of the aquifer sys- to the aquifer discovery pressure should be tem to this parameter. This will help es- adopted. tablish the priority to be assigned to the study of transient coning. Charging Pressure. In operating CAES plants, no apparent advantage results from Well Spacing. It can be observed5 by con- using high injection pressures; actually, sidering the dynamics of flow in porous there are economic benefits of injecting media that, for a given charging or dis- air at the minimum possible pressure com- charging time, a critical distance exists patible with the reservoir dynamics and around each well beyond which only a negli- power availability. However, during air gible amount of compressed air storage can bubble development, a high pressure will occur. This gives rise to an economic con- reduce the development time. An upper straint on maximum well spacing (i.e., limit on charging pressure is imposed to greater spacings would be wasteful of land avoid exceeding either the caprock and bubble volume). The critical spacing threshold pressure or the overburden pres- can be calculated from a diffusion time sure. formula.3'5 Aquifer Geometry. It is obvious that the AQUIFER FLOW MODELING reservoir design and storage capacity must be compatible with the site-specific geom- Due to the design requirements imposed etry of the formation. Contour maps for by the system coupling variables (turbine the site enable the information needed for inlet pressure and mass flow rate) and the 407 need to select a proper pressure ratio for saturation,4 be used in all calculations. the air compressor train, a prediction of Second, since, in a radial geometry, the the formation pressure drop is needed in pressure losses are concentrated around the the design optimization procedure. Based wellbore, which should be relatively dry, on an extensive study of the modeling re- it seems justifiable to use the dry perme- quirements for flow in porous radial disc ability values in estimating the pressure geometry (as applied to CAES applications drop in the reservoir. in edge-water aquifers),5 It can be reason- ably expected that a simple quasi-steady OPTIMAL DESIGN OF A CAES model will suffice. To allow consideration SUBSYSTEM WITH AQUIFER RESERVOIR of bottom-water reservoirs (having a coning constraint), as well as to insure that po- The decomposition concept described tential economic benefits of partially- earlier suggests that the aquifer reservoir, penetrating (i.e., shallower) wells can be compressed air piping, and air compression examined by the optimization procedure, a train should be designed concurrently as a • more general, two-dimensional, version of subsystem. This grouping has minimal in- 2 the quasi-steady model is employed in the teraction with the rest of the CAES system. present study. In order to easily use the It should be realized that any attempt to formation pressure drop equation (i.e., to design and optimize only part of this sub- employ a single "typical well" model), the system (namely, the aquifer well-field active part of the actual dome-shaped res- alone) would be less satisfactory and, ervoir has to be represented by an equiva- possibly, misleading. The resulting "solu- lent cylinder having the same projected area tion" would be dependent on assumed values and a height equal to the ratio of actual of parameters such as piping pressure drops storage volume to projected area. This in- and would not directly allow the compres- formation is determined from-contour maps. sion costs to impact the reservoir design. The amount of mass stored or removed Site-specific reservoir design studies during a given charge or generation pro- for CAES have been discussed in previous cess is used, together with the void volume literature. For example, a conceptual de- of the active well-field within the bubble, sign of a complete CAES plant using the to determine the change in mean formation Brookville aquifer as the reservoir was pressure occurring during that process. conducted by General Electric.6 The design Combining this information with the quasi- was based on more or less state-of-the-art steady formation pressure drop prediction equipment and was used to test some general enables the maximum and minimum wellbore conclusions concerning technical and econo- pressures occurring during the week to be mic feasibility of compressed air storage. found. These values, in turn, are needed Also, Katz and Lady1* have analyzed (and in assessing the compatibility between the partially optimized) an aquifer and a reef reservoir design and the aboveground equip- system to illustrate a design philosophy ment. for CAES plants using underground porous media. The general techniques resulting The chief uncertainty in the flow from the present project should aid in modeling just described is that it assumes conducting optimal design studies for CAES values for the effective permeability and systems in the future. porosity of the porous medium are known. These values are influenced not only by SUBSYSTEM PERFORMANCE MODELING AND heterogeneities in the rock, but also by DESIGN CONSTRAINTS the distribution of water throughout the formation following bubble development and The performance modeling and design subsequent dryout (to the extent it occurs). constraints associated with the aquifer Although moisture effects have been con- were discussed in an earlier section. sidered, 3 more work is required to resolve These aspects have received the greatest the issues. As a reasonable measure, for attention because they are complex and design study purposes, the following sim- reservoir costs are dominant in subsystem plifications are used. First, to account 1. Rather simplified compressor train and for the reduction of storage space because piping system performance models are used of moisture remaining after bubble develop- in the present subsystem 1 design procedure. ment, a modified porosity has to be defined. However, the incorporation of more detailed It is recommended that, until more accurate models would not alter its basic structure. information becomes available, the dry por- osity value, reduced by the connate water 408 The design considerations are best power generation process occurring at the illustrated by reviewing a typical step in time of the week for which the mass stored the iterative search for the optimum CAES (bubble pressure) is minimum. This proce- subsystem design. The typical design step dure enables the determination of the mini- includes compression train design, based mum pressure available to the power genera- on flow rate and pressure drop calculations tion train for the design being considered. for a charging process, and checking of the For a design to be acceptable, this pres- available turbine inlet pressure, based on sure must be at least as high as the speci- pressure drop calculations for a power fied inlet pressure. generation process. In the compressor design stage de- First, the compressor train mass flow scribed above, the total charging time was rate is calculated from the known turbine used; it influenced the predicted charging flow rate and ratio of weekly power genera- flow rate and power. It should be noted tion to storage time. This calculation in- that this charging time duration and power corporates the non-growth constraint for level must be checked for compatibility with the air bubble and also assumes (for sim- the specified utility load cycle. An ideal- plicity) that the mass flow rate during ized utility load cycle is shown in Fig. 4, every charging period is the same. Next, together with the corresponding reservoir the required compressor train discharge air storage cycle. pressure is calculated by adding the pres- sure drops in the wells and compressed air piping to the maximum wellbore bottom pres- , nKnaat sure, predicted with the aquifer model. •r MB PLANT The maximum wellbore pressure depends on the weekly mass charging/discharging cycle, mmr the wellbore diameter, depth of well pene- tration, well spacing, and number of wells. From knowledge of the compressor train dis- charge pressure and flow rate, and specifi- -.BMCWll MASS STOKED cation of the pressure ratio across the •MS JISMD low-pressure compressor* (either 11:1 or \A_A 16:1), the total compression power is cal- culated from available data.3'5'7 MOMW TUBBKI KMBMT MUMMY RIMr MrUKW MS». NOMMT The pressure difference from well-head to well-bottom reflects friction and grav- Fig. 4. Idealized Weekly Utility Load ity effects using standard relationships • 3>i l and Air Storage Cycles The piping system friction pressure drop is patterned after the simplifications employed The power generation level and time sche- by Katz.1* It is assumed that the majority dule is considered invariant, reflecting of the pressure drop in the surface piping the power demand for which the CAES plant system occurs in the main pipeline and that is to be designed. The excess power level an equivalent pipe length (L) can be de- for 'storage and its daily available time fined to account for pressure drops in the durations, however, have maximum values feed, cross-feed and branch pipelines. The but these may not be entirely needed by most significant design variable of the the CAES plant which is being designed. piping system is then the diameter of this Since the compressor power and the time main pipeline. Standard relationships are variation in air storage over the week can used in these calculations.3 A 2% addi- both influence the subsystem costs, the tional pressure drop in the aftercooler is tradeoffs between the two* allowed by the added. present optimization procedure can lead to potential operating cost reduction. After the pressure drop analysis of the compression process, some similar pres- sure drop calculations are done for the The beginning and ending time for each charging process and the compression power The compressor train is modeled as com- (assumed uniform for simplicity) are all prising a low-pressure compressor, a considered as design variables, subject to booster compressor, and appropriate inter- the maximum value constraints imposed by coolers and aftercooler. the utility. 409 SUBSYSTEM ECONOMICS: THE OBJECTIVE OPTIMIZATION PROCESS FUNCTION A detailed discussion of the nonlinear At every stage in the optimization programming (optimization) algorithms, or process, the trial design being considered their computer code implementations (OPT8 has an associated set of costs. To put the and BIAS9), which were employed in this costs on a common basis, it was decided to study, will not be given here. In essence, minimize the total operating cost (per unit these generalized procedures interact with of power generated) attributable to subsys- computer subroutine representations of the tem 1. In the terminology of optimization subsystem performance and cost models, and theory, this operating cost is the objec- the constraint definitions, in order to tive function to be minimized. It com- find that combination of design variables prises the annual carrying charge on capi- which minimizes the objective function and tal, subsystem operating and maintenance satisfies all the design constraints. (O&M) costs, and the cost of compression During the course of the search for the energy (electricity) derived from the base optimum, many (e.g., hundreds) of trial plant off-peak power. The specific capital designs are considered. The computer codes costs included are: used work only on the continuous design variables. Discrete variables (those re- (1) Main piping and distribution system - stricted to only a few allowed values, such dependent on piping design and number as pipe diameters) must be examined "manu- of wells. ally" by repeated application of the com- puter code. In the present formulation, (2) Wells - dependent on number, depth, the number of wells is approximated as a and diameter. continuous variable, because it is typically (3) Land - for simplicity, assumed propor- a large number (e.g., a few hundred). The tional to projected area of air bubble. present CAES design optimization procedure results in the specification of the follow- (4) Compressor train - based on data from ing independent variables: air bubble size, Ref. 7. number of wells, well depth, wellbore di- (5) Bubble development - dependent on ameter, well spacing, compression (charging) equilibrium bubble volume (air com- time duration for each day, compression pression cost) . ratio of the low-pressure (L.P.) compressor and main piping diameter. Much additional The O&M cost is assumed proportional to the information can subsequently be derived capital charge cost. Further details on from these results (e.g., booster compres- the cost calculations and data are given sor pressure ratio, land area to be pur- in Refs. 2 and 3. The design variables in- chased, etc.). fluencing the various cost components are depicted in Table 1. ECONOMIC TPADEOFFS IN DESIGN Table 1. Subsystem 1 Cost Factors The number of design variables (re- lated to the flexibility of the model) re- Cost Items quires the investigation of many tradeoffs Bubble during optimization. Although some of Design Com- Develop- Air Com- these tradeoffs are perhaps obscured by the Variables Wells Land PipinR pressor ment pression "automatic" nature of the optimization, the Utility Load formulation of the procedure and operational X Cycle X experience have led to the identification Wellbore of several tradeoffs. Diameter X X X Well Pene- (1) Active Bubble vs. Total Bubble Size. tration X X X Little incentive exists to sinking wells Bubble near the outer periphery of the reservoir. Thickness X X X X For a given bubble thickness at the apex Number of of the dome, as the active land area in- X X Hells X X creases, the average well thickness de- Well Spacing X X creases (see Fig. 3) so that the perfor- Main Pipeline mance per well suffers and the number of Diameter X X X wells increases. However, fewer, deeper, h.V. Compres- wells concentrated near the center of the sor Pressure X X bubble results in development of a largely 410 Inactive bubble and In greater cost per The major testing of the capabilities well. of the design optimization procedure has been for the example of a hypothetical 600 (2) Well Penetration. Greater penetration MW CAES plant using the Media, Illinois, of wells into the bubble reduces the number Galesville aquifer as the reservoir. Con- of wells, but increases the cost per well. tour maps and "material properties" for (3) Well Spaaing. Closely spaced wells this aquifer were taken from Ref. 4. The have less pressure drop (compression cost) geometrical information on storage volume but also less storage volume associated and projected bubble area as a function of with each well. bubble thickness, based on the contour maps, is tabulated in Ref. 3. Other pertinent (4) Bubble Thickness. Greater thickness parameter values used in the study are permits deeper wells (fewer needed) but re- given in Table 2. quires more surrounding land (projected area of bubble). Table 2. Galesville Study Parameters (5) Compression Time. Use of all the charging time available minimizes the capi- Aquifer Discovery tal cost of the compressor train (lower Pressure 840 psla flow rate). Use of reduced time (higher Effective Porosity 14.3% flow rate) can alter the shape of the res- Closure (top struc- ervoir mass storage cycle, reducing the ture to spill point) 110 ft maximum pressure swing. This could reduce Average Horizontal the number of wells needed to meet the tur- Sand Permeability 448 md bine inlet pressure requirements. Average Vertical Sand Permeability 354 md EXAMPLE APPLICATIONS Specific Flow Rate 10.4 lbm/kWh Turbine System The CAES subsystem 1 design optimiza- Inlet Pressure 750 psla tion procedure described in the preceding Utility Cycle: sections has been successfully implemented. Power Generation 600 MM Results of applying it are presented in Power Generation Tine 5 days, 10 hrs/day this section for illustration purposes, to Max. Compression examine the potential economic impacts that Power 590 MW can be achieved with aquifer reservoir sub- Max. Compression Time 6 days, 10 hrs/day system optimization, and to examine the (excludes Friday) effects of certain cost parameters. Storage Temperature 150^ Base Plant Electri- Originally, it was planned to apply city Cost 15 mllls/kWh the new procedure to the design of a 600 Land Cost $1200/acre MW plant at Brookville, Illinois, so that the optimized design could be compared with Other Cost and Sub- system Parameters see Ref. 3 the G.E. design.6 In preparing to do this, however, it was noted that the G.E. design appears to violate the spill point con- As a starting point, a feasible (but straint for the Brookville aquifer site. nonoptimal) design for the Galesville plant That is, the storage volume encompassed was developed intuitively, although this is by the G.E. Bvookville reservoir design not essential for the implementation of the exceeds that available above the spill optimization procedure using the OPT8 or point, as determined from contour maps of BIAS9 algorithms. Table 3 compares the 1 the aq . ifer layer. In the Brookville initial intuitive design with two optimized study, the actual site-specific properties designs. The first one of these is the re- (porosity, permeability, average aquifer sult of the formulation described in pre- thickness) were used, but the reservoir vious sections. The second design was ob- was approximated as a constant thickness tained by employing a further simplification circular disc without water-related con- in which the charging time variables were straints. Application of the procedure held constant. In all these cases, the developed in the present study, which at- discrete design variables were held fixed temps to account for geometrical limita- at the values: main pipeline diameter tions more correctly, led to a design with (48 in.), wellbore diameter (7 in.), L.P. about 700 wells; 308 wells were recommended compressor pressure ratio (11:1). in the G.E. report6 using the less restric- tive aquifer geometry assumption. 411 Table 3. Sample Galesville Study Results The oprimization also underlines the compromise necessary between compression power requirements and capital costs of the Optimized Fixed reservoir system. Higher compressor power Initial Optlalzed Storage and cost are tradeoffs for lower land, bub- Design Design Tlae Design ble development, and well construction costs. Aquifer Reservoir In the first optimum design, the weekly res- Specifications ervoir pressure variation is reduced by an Surface area to be even distribution of air storage over the bought (acres) 5895 2777 2944 entire cycle. This is accomplished by re- Active well-field ducing the weekend storage process dura- area (acres) 2755 2038 2944 tions. On the other hand, the simplified Air bubble thickness (ft) 105.0 79.9 81.7 optimization, with fixed storage times, Hell depth (ft) 1385 1369 1380 uses a larger active reservoir volume and Well spacing (ft) 467 S30 533 reduced reservoir formation pressure drop Kuaber of wells 700 402 575 (larger number of wells) to decrease the cyclic pressure fluctuation. A noteworthy Systea Pressures. Flow Rates and Powers feature of the optimum design is partial Minimum available turbine utilization of the air bubble. This is systea inlet pressure (pal) 775.5 750.0 750.0 caused by the high cost of constructing Total storage process additional wells in the outer region of tlae (hrs) 59.4 51.4 60.0 the bubble, where they yield only minimal Air flow rate during benefit due to the tapering of the aquifer storage processes (lbB/sec) 1459 1686 1444 formation. Compressor power required (MW) 385 449 384 The most important results are the re- Coapreaaor discharge ductions achieved in the subsystem 1 costs. pressure (pal) 850.3 879.3 874.6 The optimization procedure described herein System Coats yielded a 39% lower capital cost and 20% Land cost ($, millions) S. 843 4.165 4.417 smaller operating cost, compared to the Bubble development coat initial design! Restricting the storage ($, million*) 6.818 2.580 2.779 processes to fixed values caused these im- Hell construction cost provements to be only half as much. Al- ($. millions) 73.379 41.915 60.190 though substantial design improvements have Lou pressure compressor been made, further cost reductions are ex- cost ($, millions) 4.642 4.996 4.619 pected as the optimization algorithms are Booster compressor cost fine tuned and the models improved. ($, millions) 4.455 4.897 4.511 TOTAL CAPITAL COST ($, millions) 101.59 62.00 79.97 Further optimization runs for the Galesville problem have been made, using REDUCTION IN CAPITAL COST (%) - 39.0 21.3 different starting point designs, to deter- 3ASE LOAD ELECTRICITY COST (mllls/kUh) 11.45 11.54 11.53 mine whether the "global" optimum has been TOTAL SUBSYSTEM OPERATING found. The best of these solutions has an COST (aills/kWh) 24.25 19.36 21.60 operating cost of only 17.8 mills/kWh, a REDUCTION IN OPERATING reduction of 8% from the optimum value given COST (*) 20.2 10.9 in Table 3. This design has only 252 wells, an active area of 1294 acres, a bubble thickness of 91.6 ft., and a 53 hr. charging There are many interesting observa- time. Interestingly, the fixed charging tions to be made from the results in Table time (60 hr.) version of this solution is 3. Both of the optimum designs reduce the very similar in design and cost. number of wells, average well depth and bubble size, indicating that the starting When a CAES plant using the Media point was a case of overdesign. This con- Galesville aquifer was investigated by Katz 1 clusion can also be drawn from a compari- and Lady, * they concluded "... use of 100 son of available turbine system inlet pres- input/output wells seem reasonable for full sures in the three designs. Thus, a care- development (600 MW)." This number, not fully formulated constrained optimization based on detailed optimization, is consider- problem has allowed a reduction in "safety ably less than found in the present study factors" required in an intuitive design (252). The discrepency may be partially process. due to the imposition of the diffusion time- 412 related constraint on well spacing in the costs were varied, thus explaining the optimization procedure. Whether that con- linear relationships in the figures. Al- straint is conservative or the assumptions though this observation may not be of gen- of the previous investigators overestimates eral validity, it would be comforting to the reservoir flow capability under CAES know that a CAES design would remain opti- cycling conditions remains to be determined. mum if the cost of base-plant electricity were to increase in the future! Examination of the various Galesville optimization runs, and those done for the 19.00 Brookville site, shows that the optimum wells penetrate nearly to the bottom of the bubble, often being limited by the coning constraint. Increasing the coning distance parameter from 1 ft to 5 ft increases the operating cost by about 1 mill/kWh, indica- 18.75- ting that the coning problem should be studied further. For the Galesville problem, the effect of wellbore and main pipeline diameters on optimum design operating cost are shown in 18.50- Figs. 5 and 6, respectively. The main pipe size has little effect. 21.0 18.25' Mill PIPE DIMETER (IK.) Fig. 6. Effect of Main Pipe Diameter in Galesville Problem 21.0' 2.0 4.0 6.0 8.0 10.0 12.0 HEIUORE DIAMETER (IN.) Fig. 5. Effect of Wellbore Diameter in Galesville Problem The effect of certain cost parameters 27.0' on the optimum Galesville subsystem 1 opera- 10 20 JO « 60 ting cost has also been investigated (see KU.-DMU.IIK COST it/n.) Figs. 7-9). It was found that the optimum Fig. 7. Effect of Well-Drilling Cost- design variables did not change as these Galesville. 413 CONCLUDING REMARKS The design procedure described in this paper appears to be the most complete method available for designing aquifer reservoir-based CAES plants. Limited com- parisons with published results using more simplified methods of analysis suggests a possible inadequacy in those methods. Fur- ther work is recommended to resolve these issues. The design optimization procedure is general in its structure, but its current computer implementation is somewhat res- tricted (e.g., bottom-water reservoirs, equal compression power for each charge process, etc.). It is also based on a somewhat idealized aquifer model and on particular judgements on important con- straints. However, extensions and refine- ments can be readily incorporated as re- quired. Utilization of the design optimization 6 5000 10000 15000 20000 25000 procedure can be valuable, when carefully LAND COST (S/ACRE) applied. It can: Fig. 8. Effect of Land Cost - Galesville - result in actual capital and operating cost savings in plant design, - give insight into the economic trade- offs among design variables, and - assess the influence of uncertainties in cost data. Furthermore, if combined with a similar optimization model for the turbine system, a complete CAES plant design optimization could be performed. Some additional information on the work presented is available in Refs. 2 and 3. A final report is in preparation which will provide full documentation, including listings of the computer subroutines embody- ing the optimization-oriented CAES model. ACKNOWLEDGMENTS The important contributions of Ajay Sharma (University of Illinois at Chicago Circle), Rajesh Ahluwalia (Argonne National Laboratory), and Ken Ragsdell (Purdue University) in the development and imple- mentation of the methods and models des- 20 40 60 80 100 cribed herein, and in the preparation of previous manuscripts on the project, are ELECTRICITY COST (MILLS/KKH) gratefully acknowledged. This work was supported by the Division of Energy Storage Systems, Office of Conservation, U.S. Fig. 9. Effect of Base Plant Electricity Department of Energy. - Galesville 414 REFERENCES 'Kim, C.S., and G.T. Kartsounes, A Para- metric Analysis of Turbomaahinery Options for Compressed Air Energy Storage Plants, Froc. of Compressed Air Energy Storage Technology Symposium, Pacific Grove, Cal. (May 1978). 2Sharma, A., F.W. Ahrens, K.M. Ragsdell, R.K. Ahluwalia, and H.H. Chiu, Design of Optimum Compressed Air Energy Storage Sys- tems, to be presented at the 1978 Midwes- tern Energy Conf., Chicago, Nov. 19-21, 1978; to be published in ENERGY (journal). 3Ahluwalia, R.K., A. Shamra, and F.W. Ahrens, Design of Optimum Aquifer Reser- voirs for CAES Power Plants, Proc. of Compressed Air Energy Storage Technology Symposium, Pacific Grove, Cal. (May 1978). "Katz, D.L., and E.R. Lady, Compressed Air Storage for Electric Power Generation, Ulrich's Books, Inc. Ann Arbor (1967). 5Ahluwalia, R.K., F.W. Ahrens, A. Sharma, H.H. Chiu, and G.T. Kartsounes, Dyanmic Analysis of Porous Media Reservoirs for Compressed Air Energy Storage, Argonne National Laboratory, unpublished infor- mation (1977). 6Bush, J.B., Jr., Principal Investigator, Economic and Technical Feasibility Study of Compressed Air Storage, ERDA Report No. 76-76 (March 1976). 7Davison, W.R., and R.D. Lessard, Study of Selected Turbomaahinery Components for Compressed Air Energy Storage Systems, prepared by United Technologies Research Center for Argonne National Laboratory, Report ANL/EES-TM-14 (Nov. 1977). "Gabriele, G.A., and K.M. Ragsdell, OPT; A Nonlinear Programming Code in Fortran- IV-Users Manual, Purdue Research Founda- tion (June 1976). 9Root, R.R., and K.M. Ragsdell, BIAS: A nonlinear Programming Code in Fortran-TV- Users Manual, Purdue Research Foundation (Sept. 1978). 415 EVALUATION OF TL1RBOMACHINERY FOR COMPRESSED AIR ENERGY STORAGE PLANTS George T. Kartsounes and Choong S. Kim Energy and Environmental Systems Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439 ABSTRACT This paper presents a study of possible turbomachinery options for compressed air energy storage plants. The plant is divided into four subsystems: a turbine system, compressor system, motor/generator, and an underground air storage reservoir. The tur- bine system comprises a high-pressure turbine and combustor, a low-pressure turbine and combustor, and a recuperator. The compressor system comprises a low-pressure compressor, booster compressor, intercoolers, and an aftercooler. A water-compensated mined cavern constitutes the underground air-storage reservoir. Plant performance is presented in terms of five parameters: specific air flow rate, specific heat rate, specific storage volume, specific compression rate, and overall plant efficiency. The capital and oper- ating costs of the plant as a function of the turbomachinery options are presented. De- sign variables of the turbomachinery are the reservoir pressure and inlet gas tempera- tures to the turbines. INTRODUCTION Compressed air energy storage is a The components of the subsystems of a near-term technology for the load leveling CAES plant are delineated here for preci- and peak shaving strategies being consi- sion of reference in this paper. The tur- dered by electric utilities. Assessments bine system consists of a low-pressure gas of the technical and economic feasibility turbine (LGT) and combustor, a high-pres- of this storage system indicate that it is sure gas turbine (HGT) and combustor, and economically competitive with conventional a recuperator (see Fig. 1). The LGT is a gas-turbine peaker units. The CAES concept turbine modified from a conventional gas- is based on a split Brayton cycle with an turbine peaker unit. For proposed CAES accompanying underground air storage res- plants, the HGT is a modified steam turbine ervoir. During periods of off-peak power operating at gas temperatures of about demand, air is compressed with base-plant 1000°F. Optimized designs for compressed- power and stored in the underground reser- air turbines that operate at high tempera- voir. For power generation, the air is tures have been investigated.1 The com- discharged through a combustion turbine bustors can be designs modified from con- during the peak demand period. ventional gas-turbine peaker units. Pre- liminary studies indicate that recuperators Because the storage reservoir is usu- can be designed that are economically fea- ally the most costly single component in a sible for CAES application. These differ CAES plant, its volume is a sensitive de- from conventional gas-turbine peaker units sign parameter. The volume required is because of the high-pressure air leaving affected by storage pressure and tempera- the reservoir. ture, power level, generation time, reser- voir type, air quantity required by the The compressor system contains a low- turbine system, and pressure ranges per- pressure (LC), high-pressure (HC), and mitted by the turbomachinery (turbines and booster compressor (BC), intercoolers, and compressors). Compressed air can be stored an aftercooler (see Fig. 1). Intercooling underground in caverns or in the pore space is required to operate the compressors of porous rock formations. within limits tolerable for standard 417 COOLING AIR T T MOTOR/ GENERATOR T T GENERATOR RECUPERATOR CAVERN (Po.To) Fig. 1. Schematic Diagram of CAES Plant materials. An aftercooler is used to cool The cost of a CAES plant can be char- the air to avoid possible thermal-stress acterized in terms of capital cost and damage to the storage reservoir. operating cost. Capital cost includes the direct cost of the air storage facility, The performance of a CAES plant can the turbomachinery, the balance of plant, be characterized in terms of four specific and the indirect cost due to a contingency parameters and an overall plant efficiency: allowance, engineering and administration, and escalation and interest during con- • Specific air flow is the mass flow struction. The operating cost of the plant rate of air supplied to the turbine includes the capital- charge, cost of fuel system per kilowatt power generated. to the combustors, off-peak electricity to It is the major factor in determining the compressors, and operation and mainte- the size of the turbines, compressors, nance costs. and air-storage reservoir. This paper presents a study of pos- • Specific heat rate is directly propor- sible turbomachinery options for CAES tional to fuel consumption and is plants with particular emphasis on the tur- equal to the product of specific fuel bine system. The performance and cost of consumption and the lower heating the complete plant resulting from differ- value of fuel. It therefore affects ent turbomachinery options are presented. the operating cost of the turbines. The turbine system design parameters con- • Specific storage volume, the volume sidered are the reservoir storage pressure of reservoir required per kilowatt of and the inlet gas temperatures to the LGT power generated, is dependent on the and HGT. The LGT was based on a nominal specific air flow rate and the tem- pressure ratio of 16:1.* A water-compen- perature of stored air. sated mined cavern was chosen as the com- pressed air storage reservoir. • Specific compression rate is the ener- gy equivalent of the power supplied to the compressors per kilowatt of power generated. This parameter is the amount of off-peak energy required to Studies have indicated that pressure ra- operate the compressors. tio has a minor effect on performance and • Overall plant efficiency is equal to conventional low-pressure turbines (from the total energy output from the tur- peaker units) having a nominal pressure bines divided by the sum of the ener- ratio of 10-16:1 can be used.2'3 gy input from the fuel and off-peak energy to the compressor system. PERFORMANCE EVALUATION following parameters were assumed to be known or specified. THERMODYNAMIC ANALYSIS Adiabatic efficiency of compressors: A thermodyanmic analysis was carried out on each subsystem of a CAES plant, and ^HC " \C m nBC = °-90; the results were combined to evaluate over- all plant performance. Design parameters Temperatures: T . = 77°F, T... = T,_ = T._ considered in the analysis include: air storage pressure and inlet gas temperatures = 100°F, Tlg » 120°F; and to the high-pressure gas turbine and low- pressure gas turbine. The details of the Pressures: p... = 1 atm, p., = 16 atm. analysis are presented in Ref. 4. PERFORMANCE RESULTS Underground Air Storage System. The under- ground air storage reservoir considered is Results of the parametric study are a water-compensated cavern. Therefore, the presented in terms of the five performance pressure variation in the cavern during the parameters: specific air flow rate, spe- operating cycle is negligible. The air cific storage volume, specific heat rate, temperature of the storage cavern (To) was specific compression rate, and overall assumed as 120°F (322°K) and four different plant efficiency. These values are given air storage pressures (p0) were considered as a function of air storage pressure and in the analysis: 30, 50, 70, and 100 atm inlet gas temperatures to the HGT and LGT. (3 x 106, 5 x 106, 7 x 106, and 1 x 107 Pa) . Specific air flow rate is the flow Turbine System. The selection of the tur- rate of air coming out of the cavern per bine system (see Fig. 1) evolved from the unit output of the turbine system. It is results of a previous study.2 The follow- directly proportional to the turbine and ing values of system parameters were con- compressor sizes, and, thus, is an impor- sidered: tant factor in determining the cost of the above-ground facility. A plot of specific Turbine efficiencies: nLGT = nHGT = 0.90, air flow rate against air storage pressure at different turbine inlet gas temperatures Recuperator effectiveness: e = 0.8 (Fig. 2) shows that the air flow rate varies from 6.6-12.0 lb/kWh (3.0-5.4 kg/ Temperatures: T, = 1000°,1600°,2000°,2400°F kWh) for the conditions specified in this (811°,1144°,1366°,1589°K) study, and it decreases as air storage pressure'increases• In addition, higher T = 1600°,2000°,2400°F turbine inlet gas temperatures result in 3 (1144°,1366°,1589°K), smaller air flow rate, even though cooling air is required. Pressures: 16 atm (1.6 x 106 Pa). Specific storage volume, the required Subscripts given in the above parameters storage cavern volume per unit work output, correspond to the components or stations is directly related to the cost of the un- in Fig. 1. The efficiencies of turbines derground facility for a CAES plant. This and combustors are based on state-of-the- storage volume depends on the required art values of available equipment.1 Recup- specific air flow rate as well as on cav- erator effectiveness is a function of the ern conditions, such as pressure and tem- heat exchanger specifications. Because the perature of stored air. Consequently, re- temperature of the inlet gas to the tur- sults for the specific storage volume show bines must be kept low enough to avoid a trend similar to that for the specific thermal damage of the turbine blades and air flow. Figure 3 shows the effects of vanes, cooling air is required for higher air storage pressure and turbine inlet gas inlet gas temperatures. The amount of temperatures on the storage volume. It is cooling air required was determined from seen that smaller storage volume results data presented in Ref. 1. from higher air storage pressure or higher turbine inlet gas temperatures. Specific Compressor System. The study was extended storage volume in this study varies from to the compressor system in order to com- 0.96 ft3/kWh (0.027 m3/kWh) to 5.84 ft3/ plete the analysis of the CAES plant. The kWh (0.162 mVkUh). 3,700 30 40 50 60 70 80 90 100 30 40 50 60 70 80 90 100 STORAGE PRESSURE, Po (aim or 105 Pa) STORAGE PRESSURE, Po (aim or l()5 Pol Fig. 4. Effect of Storage Pressure on Specific Heat Rate Fig. 2. Effect of Storage Pressure on Specific Air Flow Rate T5 CK) 1200 1300 1400 1500 0.175 4,200 6.c I I I I 4.4 r T3=IO0O*F(8ll«K)iT5c|600*(ll44«K) * HGT INLET TEMPEMTURE, Tj= # v/ /V V«00 F(ll44»K) 0.150 4,100 £ 5.0 7\/ TjsTjtaooo'FdJss'K) 0.125 "3 H Yvk/ TjtT5«2400'F(IJ89*K) 4.0 3 oioo _• § 3.0 _V\NS\v\ /. _ u 0.075 £ 2.0 4.0 - 0.050 STO R 3.9 1.0 — ^-^^s 0.025 1600 1800 2000 2200 2400 0 1 1 1 1 1 1 0 30 40 50 SO 70 80 90 100 LGT INLET TEMPERATURE, T5 CF) STORAGE PRESSURE, Po (aim or I05 Po) Fig. 5. Effect of Turbine Inlet Tempera- tures on Specific Heat Rate Fig. 3. Effect of Storage Pressure on Specific Storage Volume Specific compression rate is .the fuel equivalent of the off-peak electrical ener- Specific heat rate is a measure of gy input to the compressor system per unit premium-fuel usage for the combustors per power output of turbine system. In this unit power output of the system. It varies study, specific compression rate is based in this study from 3700 Btu/kWh (3.98 x 106 on an off-peak heat rate, including elec- J/kWh) to 4280 Btu/kWh (4.52 x 106 J/kWh). trical and mechanical losses, of 10,400 The effect of storage pressure on the heat Btu/kWh (1.097 x 107 J/kWh). For the con- rate is given at different turbine inlet ditions of this study, the rate varies from gas temperatures in Fig. 4: higher stor- 5280 Btu/kWh (5.57 x 106 J/kWh) to 7790 age pressure results in lower heat rate. Btu/kWh (8.22 x 106 J/kWh). Figure 6 shows Figure 5 shows that heat rate increases as that, in general, compression rate increases the LGT inlet gas temperature increases slowly with increasing storage pressure and and that the HGT inlet gas temperature has smaller compression rate is required by a minor effect on the heat rate. higher turbine inlet gas temperatures. 420 _ 9.000 a reasonable basis for the economic analy- sis, the following operating cycle was I chosen: 20-hr nominal cavern storage capa- I «.000 a > a > T3>KH>0 F((M K)iTs'!(00 F(ll44 K) city and 2190-hr/yr generation time. 7,000 0.59 7 i"* .24OO'F(B.»'K) 6,000 = fa. 3 0.58 £ ^. 6 •••i* S _ 13*15.2400^ 0589'K)' 0.57 * T3=TS.tl600*F(H44»K) C 4,000 I I I I I I 0.56 T5=2000»F{I366 K) 3O405O607OS090 100 STORAGE PRESSURE, Po (otm or 10$Pa) 0.55 _ TjS|000*F The effects of turbine inlet gas tem- I6OO*F(II44*K) peratures on plant efficiency are given in Fig. 8. It shows that higher plant effi- 1- 0.55 ciency is obtainable with higher HGT inlet gas temperature. It also shows that effi- IOOO*F(eil*K). ciency increases with the LGT inlet gas tem- perature for T3 - 2000°F (1366°K) or 2400°F (1589°K), and it has a minimum at about T5 - 2000°F (1366°K) for T3 - 1000°F (811°K) g 053 _ Po*70lt*(7ilO*Po) or 1600°F (1144°K). I I I 1600 1800 2000 2200 2400 ECONOMIC ANALYSIS LGT INLET TEMPERATURE, T5fF) An economic analysis of the CAES plant was made to show the effects of the para- Fig. 8. Effect of Turbine Inlet Tempera- meters on capital and operating costs. The tures on Overall Plant Efficiency analysis was based on the performance re- sults described above. In order to pro"*''" 421 CAPITAL COSTS OPERATING COSTS Direct capital cost of the CAES plant Operating cost of the CAES plant con- was divided into the following: cost of sists mainly of capital charge, cost of underground air storage cavern and water- fuel to the combustors, off-peak electri- compensating reservoir, cost of turbomach- city to the compressors, and operation and inery equipment, and balance of the plant. maintenance. Annual capital charge was estimated from the total capital cost based The storage cavern cost included the on the fixed capital charge rate of 18% per cost of the air and water shafts, cavity, year. Estimation of the cost of premium development and mobilization, and comple- fuel was made by multiplying the specific tion. The cost of the air and water shafts heat rate by the cost of No. 6 oil. Cost was estimated based on the cavern depth of tne off-peak electricity to the compres- which was determined by the air storage sors was estimated from the specific com- pressure. The cost of the cavity was esti- pression rate and the electricity cost mated based on specific storage volume with from the base plant. A value of 2 mills/ a 10% capacity margin. Since the storage kWh was used as the cost of operating and cavern considered in the analysis is water- maintenance for all cases. compensated, the cost of the water reser- voir was also included. The storage-rela- ECONOMIC RESULTS ted, costs were based on Ref. 5. Results of the economic study are Estimation of the turbomachinery cost given in terms of the two specific costs: was based on Ref. 1. In this reference, capital cost ($/kW) and operating cost the selling price is estimated for 10, 20, (mills/kWh). The values are presented as and 50 units. Based on the evaluation of a function of the storage pressure (po) regional markets and development potential and the turbine inlet temperatures (T3 for CAES conducted by Harza Engineering and T5). Company,6 a 50-unit selling price was used in this study. The cost of the low-pres- Capital Costs. Capital cost of a CAES sure gas turbine with a cycle-pressure ra- plant varied from $285/kW to $406/kW for tio of 16:1 was determined by the inlet gas the range of design parameters specified temperature and the cost of the high-pres- in the study. The cost of the underground sure gas turbine was determined by both the storage cavern was found to be the highest inlet gas temperature and air storage pres- component cost for most cases varying from sure. Costs of the LC and HC with the 26-46% of the total capital cost and the overall compression ratio of 1:16 were es- cost of the turbomachinery equipment varied timated from the air flow rate, and the from 16-31% of the total direct capital cost of BC was determined by the air flow cost. In general, it was found that higher rate and air storage pressure. A 25% al- turbine inlet temperatures result in higher lowance was given for the ducting and in- turbomachinery cost. stallation of the turbomachinery equipment. Total capital cost is given in Fig. 9 The remainder of the plant equipment, as a function of storage pressure for four which includes the clutches, motor/genera- different combinations of inlet gas tem- tor, recuperator, combustors, fuel storage, peratures to the HGT and LGT. Capital coolers, electrical power system, land, and cost sharply decreases with increasing plant structure was denoted as the balance storage pressure for all the cases up to of plant. This equipment is relatively in- 70 atm (7 x 106 Pa) and either slowly de- sensitive to CAES design parameters and a creases or increases thereafter. Higher fixed cost of $80/kW was used for the bal- turbine inlet temperatures result in lower ance of plant for all cases of this study. capital cost at low storage pressures, for example 30 atm (3 x 106 Pa). However, at Total capital cost of the plant was storage pressures greater than 70 atm estimated from the direct capital cost con- (7 x 106 Pa), higher turbine inlet tem- sidering the following allowances: 15% for peratures result in higher capital cost. contingency, 10% for engineering and admin- Among the cases considered in the study, istration, and 30% for escalation and in- the design parameters that result in the terest during the construction period. lowest capital cost are those when T3 = T5 = 1600°F (1144°K) and pn = 100 atm (1 x 107 Pa). 422 operating cost. Consequently, the opera- ting cost in Fig. 10 shows a similar trend i to that of the capital cost. The figure 400 shows that the operating cost decreases with increasing air storage pressure for all cases but T3 - T5 - 2400°F (1589°K), which 380 has a minimum at about 70 atm (7 x 106). '1000* P(8II'K); Ts'l*(»«F<1.44'K) It also shows that, among the cases studied, the lowest operating cost results when To » I 360 f T5 • 1600°F (1144°K) for po > 58 atm (5.8 x 6 10 Pa) and T3 • T5 - 2400°F (1580°K) for 6 340 p0 < 58 atm (5.8 x 10 Fa). However, in the pressure range of 50-90 atm, which is the 1 Tj«T . MOO'F 1569 5 •K) COS T ( J most likely range for CAES with a water-com- 320 hi pensated reservoir, the difference in opera- ting cost between different turbine systems Is less than about 3 mills/kWh. IAPITA L 300 --— _T5.T5.SWCF IM6*K 260 BUEAAfI'Hiiil 'IWV 1 III** «^ 260 1 1 1 i | 1 30 40 50 60 70 80 90 100 * 92 LL \\T3*l000'F(8ll*K)iT9c|(00*F(ll44*K) STORAGE PRESSURE, Po (atm or »5po) Fig. 9. Effect of Turbine Options on Capital Cost The dotted curve in Fig. 9 represents the cost of a plant using a modified steam turbine (%JT " 78Z) for the HGT, which operates at I000°F (811°K) inlet gas tem- perature. The solid curve for T3 * 1000°F (811°F) and T5 - 1600°F (1144°K) is based on the assumption that the cost of this new, high-efficiency HGT (nngj " 90Z) would be the same as that of the modified 30 40 SO 60 70 80 90 100 steam turbine. The actual cost of this 5 new turbine would depend upon the develop- STORAGE PRESSURE, Po (otm or I0 Pa) ment cost and the number of units sold. Thus, the actual high-efficiency cost Fig. 10. Effect of Turbine Options curve should be somewhere between the on Operating Cost solid and dotted curves. The net result is that the cost differences would be neg- Figure 10 also illustrates that for ligible, therefore favoring the use of the T3 = 1000°F (811°K) and T5 = 1600°F (1144°K) modified steam turbine because of proven the difference in operating cost between a reliability and equipment availability. modified steam turbine for the HGT and a new, high-efficiency design is negligible; Operating Costs. The operating cost of the i.e., less than 1 roill/kWh. Thus, the use CAES plant is given in Fig. 10 as a fun- of a modified steam turbine would be fa- tion of the design parameters. The cost vored because of proven reliability and of premium fuel was selected as $2.50/106 equipment availability. Btu and the electricity cost was 15 mills/ kWh. In this figure the operating cost Effect of Electricity and Fuel Costs on varies from 44.8-55.5 mills/kWh. Operating Costs. Table 1 illustrates the effect of different electricity and pre- The capital charge was found to be mium fuel costs on the overall operating much higher than the cost of fuel or elec- cost of a CAES plant. Two plant designs tricity; it amounts to 52-60% of the total are compared: Plant A where T3 = 1000°F 423 Table 1. Effect of Electricity and Fuel Costs on Operating Costs Operating Cost (mills/kWh) Electricity 17tta1 fV^afr Cost (mills/kWh) ($/106 Btu) Plant Aa Plant Bb X Decrease0 15 2.50 47.4 45.3 4.4 15 3.75 52.2 50.0 4.2 15 5.00 56.9 54.7 3.9 20 2.50 51.1 48.5 5.1 20 3.75 55.8 53.3 4.5 20 5.00 60.0 58.0 4.3 25 2.50 54.7 51.8 5.3 25 3.75 59.5 56.6 4.9 25 5.00 64.2 61.3 4.5 turbine inlet temperatures: T3 - 1000°F (811°K> (rw,_ - 78X) W1 T5 - 1600°F (1144°K). Turbine inlet temperatures: T3 - T5 - 1600°F (1144°K), C100(Plant A - Plant B)/Plant A. (811°K) (TTHQJ) - 78%) and T5 - 1600°F other two reservoir types (i.e., aquifer (1144°K), and Plant B where T3 - T5 - reservoirs and salt caverns), but further 1600°F. These two plants bracket the study is recommended to fully evaluate the highest and lowest estimated operating affect of reservoir type on CAES plant per- costs. formance and cost. From this table it is seen that for From the performance analysis, the fixed electricity cost, the difference be- following trends were observed: tween plant designs decreases as the fuel cost increases. This means that if fuel 1. Specific air flow rate and storage costs increase faster than electricity volume decrease as pQ, T,, or Tj costs, then the type of plant design (i.e., increases. the selection of turbine system) becomes less significant. This scenario would Specific heat rate decreases as po favor using Plant A because of proven increases and increases as Tj in- reliability and equipment availability. creases; but is relatively insensi- tive to T,. For fixed fuel cost, the difference between the operating costs of the two de- Specific compression rate, in general, signs increases as the electricity cost in- slightly increases as po increases; creases. This means that if base plant it decreases with increasing T3 or T5. power increases in cost at a faster rate than fuel costs, then Plant B would be 4. In general, overall plant efficiency favored. In this case, the use of a new, increases as T3 increases; is only high-efficiency HGT would be justified. weakly affected by p or T-. CONCLUSIONS From the above, it can be concluded that optimum performance results from the use This paper has considered the perfor- of a high storage pressure and high inlet mance and cost of possible turbomachinery gas temperature to both turbines. options for CAES power plants. Particular emphasis was directed toward the turbine The economic analysis, however, illus- system of the plant. The main design vari- trates that minimum cost (capital and ables were the reservoir storage pressure operating) does not necessarily correspond and the turbine inlet gas temperatures. A to optimum plant performance. Considering water-compensated mined cavern was selected the specific operating cost (i.e., mills/ as the storage reservoir. The results of kWh), which can be considered the true this study should be applicable to the indicator of plant cost, at storage pres- 424 aures below about 60 atm, the highest tem- V' Specific storage volume s perature turbine system considered in this Recuperator effectiveness study (i.e., T3 = T5 = 2400°F (1589°K)) re- e sults in the lowest cost; whereas, above n Efficiency 60 atm, the turbine system with 13=15= Overall plant efficiency 1600°F (1144°K) results in the lowest cost. overall A significant result is that for the Subscripts pressure range of 50-90 atm, which is the Booster compressor range of present interest for water-com- BC pensated caverns, the operating costs for Cl Combustor 1 all of the turbine systems considered in Combustpr 2 this study are within about 3 mills/kWh of C2 each other; the average cost is about 47 HGT High-pressure gas turbine mills/kWh. Furthermore, it was observed Low-pressure gas turbine that if the cost of premium fuel increases LGT as a faster rate than the cost of base HC High-pressure compressor power electricity, which seems to be a LC Low-pressure compressor logical scenario for the future, the cost difference between turbine systems de- 0-19 Correspond to Fig. 1 creases . REFERENCES The economic study indicated that for a turbine system with T, = 1000°F (811°K) Davidson, W.R., and R.D. Lessard, Study and T5 = 1600°F (1144°K), the use of a new, of Selected Turbomachinery Components for high-efficiency, high-pressure turbine Compressed Air Energy Storage Systems, could not be justified and a modified prepared by United Technologies Research steam turbine could be used with little Center for Argonne National Laboratory, cost penalty. Report ANL/EES-TM-14 (Nov. 1977). Based on the above factors, the 2Kartsounes, G.T., Evaluation of Turbine overall conclusion of this study is that Systems for Compressed Air Energy Storage the turbine system can be constructed Plants, Argonne National Laboratory using available turbines with proven reli- Report ANL/ES-59 (1976). ability without significantly sacrificing cost. The HGT can be a modified steam tur- 3Kim, C.S., and G.T. Kartsounes, A Para- bine and the LGT can be obtained from a metric Study of Turbine Systems for Com- peaker unit which operates at an inlet gas pressed Air Energy Storage Plants, temperature of about 1600°F (or lower), Argonne National Laboratory Report ANL/ requiring little, if any, cooling air. ES-64 (April 1978). Interestingly, this is the approach being used at the Huntorf Plant,7 which is the "•Kim, C.S., and G.T. Kartsounes, A Para- world's first CAES plant. metrio Analysis of Turbomaahinery Options for Compressed Air Energy Storage Plants, ACKNOWLEDGMENTS Proc. of the 1978 Compressed Air Energy Storage Technology Symposium (May 1978) . The research activities in compressed air energy storage, which formed the basis 5Giramonti, A.J., Preliminary Feasibility of this paper, were funded by the Division Evaluation of Compressed Air Storage of Energy Storage Systems, Office of Power Plants, United Technologies Research Conservation, U.S. Department of Energy. Center, R76-952161-5 (Dec. 1976). NOMENCLATURE ^Underground Pumped Hydro Storage and Com- pressed Air Energy Storage: An Analysis E^ Specific compressic • .• ..e of Regional Mcwhets and Devsloix-mt Poten- 1 tial, prepared by Harza Engineering Co. A Specific air flow rate for Argonne National Laboratory, Argonne p Pressure Report ANL-K-77-3485-1 (March 1977). Q' Specific heat rate 7Stys, Z.S., Air Storage System Energy T Temperature Transfer (ASSET) - Huntorf Experienae, ERDA/EPRI CAES Workshop (Dec. 1975). 425 EVALUATION OF THE USE OF RECIPROCATING ENGINES IN COMPRESSED AIR ENERGY STORAGE PLANTS George T. Kartsounes and James G. Daley Energy and Environmental Systems Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439 ABSTRACT The application of reciprocating engines to compressed air energy storage (CAES) plants is presentee! in this paper. The expected advantages compared to plants using tur- bines and compressors are reduced reservoir size and cost, reduced compression energy, and increased overall plant efficiency. The performance of possible engine and plant con- figurations are presented. One configuration uses a reversible, reciprocating expander/ compressor engine. Power generation results from engine operation as an internal-combus- tion expander; compression is accomplished using the same engine operating as a recipro- cating compressor. Another possible configuration results when an internal-combustion engine is used as a high-pressure expander and a gas turbine is used as a low-pressure expander. Compression is accomplished using either separate turbocompressors or opera- ting the high-pressure expander as a reversible-reciprocating compressor in series with " low-pressure turbocompressor. Capital and operating costs of plants using reciproca- ting engines are estimated and compared with that of turbine-based CAES plant designs. It is shown that using reciprocating engines can reduce capital and operating costs by about 11% and 8£, respectively, compared to a plant using available turbomachinery. INTRODUCTION Electric utilities commonly use die- combustion temperature approaching 3000°K. sel/generator sets or gas turbine/compres- Thus, although the power output of a die- sor units for peak power generation. The sel engine and gas turbine/compressor may use of a gas turbine offers advantages in be equivalent, the higher combustion tem- higher power ranges due to its compara- perature of the diesel causes a larger tively smaller size and lower maintenance enthalpy change, hence a lower required requirements; however, many applications air flow. This reduction in air flow rate favor the better part-load efficiency of would benefit a CAES plant through reduced a diesel engine. reservoir size and reduced electrical energy needed for air compression. Another feature of diesel engine oper- ation that has significance for CAES sys- A conventional diesel engine cannot tems is a greatly reduced air flow require- be used directly in a CAES plant because ment compared to gas turbines. This re- it would not be possible to utilize the duced need for air is related to the maxi- high-pressure reservoir air. However, a mum combustion-gas temperature. In either reciprocating engine similar to a diesel a gas turbine system or diesel engine, the which decouples the compression process power output equals the product of the flow from expansion can be used. Power genera- rate of combustion products and the change tion results from engine operation as an in specific enthalpy across the engine. internal-combrstion expander and compres- This enthalpy change is approximately pro- sion is accomplished using the same engine portional to the maximum absolute tempera- operating as a reciprocating compressor. ture. In a gas turbine, the. maximum gas A basic description of the operating temperature is about 1350"K due to mater- characteristics and design features of the ial limitations, although considerable reciprocating expander/compressor engine research is in progress to eventually per- is described in Ref. 1. mit temperatures as high as 1900°K. A diesel engine can have an instantaneous 427 PLANT DESIGN Cooling water is circulated through the water jacket of each engine removing Two possible plant configurations part of the heat of compression. This heat using reciprocating engines are addressed removal reduces compression energy since in this paper. One design uses a rever- the compression process approaches isother- sible, two-stage expander/compressor con- mal compression. Heated water leaving the cept and the other uses a compound engine engines is then coded in a radiator (e.g., comprising reciprocating expander/compres- cooling tower). sor engines for high-pressure duty and a gas turbine for low-pressure expansion. In the expansion (i.e., power genera- tion) mode of operation, valves B, F and H PLANT USING REVERSIBLE, EXPANDER/COMPRES- are closed. The M/G operates as a genera- SOR ENGINES tor, and the LP and HP engines operate as expanders. Air from the reservoir is pre- A schematic diagram of a CAES plant heated in the recuperator using exhaust gas using expander/compressor engines is pre- from the LP expander. Preheated air and sented in Fig. 1. In this arrangement, a injected fuel are burned and expanded in single high-pressure (HP) engine is con- the HP expander. The products of combus- nected through a shaft to a single low- tion are further expanded in the LP expan- pressure (LP) engine. Both engines are der and then flow through the recuperator. then connected to a reversible motor/gen- The cooling water removes part of the heat erator (M/G). In this practice, the HP of combustion. This energy loss is neces- engine and LP engine may each consist of sary to reduce engine metal temperatures several separate engines that are mani- and to improve engine reliability. folded together. M/G units may be con- nected to each separate engine or a combi- Comparing the plant configuration of nation of engines. Fig. 1 with that of a turbine-based CAES plant, the HP expander/compressor engine replaces the expansion turbine and combus- tor, and the centrifugal booster compres- sors; the I.P expander/compressor engine re- places the low-pressure turbine and combus- tor, and the axial compressor. Furthermore, the intercooler and aftercooler are smaller than that of a conventional plant because part of the heat of compression is removed in the cooling water. However, engine cooling water requires an additional cir- culation pump and radiator. PLANT USING COMPOUND ENGINES The plant configuration previously considered is based on the use of expander/ compressor engines for both high- and low- pressure functions. However, a compound engine design which combines expander/com- Fig. 1. Schematic Diagram of a CAES Plant pressor engines with conventional turbo- Using Expander/Compressor Engines machinery offers improvement over either all-turbine or all-reciprocating engine designs. This configuration is depicted During the compression mode of opera- in Fig. 2. tion, valves C, D, E and G are closed. The M/G operates as a motor receiving off-peak In this configuration, the high-pres- power from a base-load power plant, and the sure turbine and combustor, and booster LP and HP engines operate as reciprocating compressors of a conventional CAES plant compressors. Ambient air is compressed in are replaced with a reversible HP expander/ the LP compressor and the heat of compres- compressor engine. A low-pressure turbine sion is removed in the intercooler. The and axial compressor are used for the low- air is further compressed in the HP com- pressure duty cycle of the plant. However, pressor and then cooled in the aftercooler since the exhaust of the HP expander will before delivery to the storage reservoir. 428 be sufficiently high (e.g., 1360°K), a low- using an additional heat exchanger between pressure combustor will be unnecessary. the air inlet and the combustion gas outlet of the HP reciprocating engine. This heat exchanger would lower the temperature of the gas entering the LP turbine and in- crease the temperature of the air entering the HP engine. Because of the large tem- perature difference between the fluid streams, the required effectiveness of this heat exchanger would be low (estimated to be about 0.4) and hence it would be inex- pensive. The main advantage of this plant design is a reduction in reservoir size which results because turbine cooling air can be reduced (or eliminated). In addi- tion, simpler, less-expensive piping would be required to transport the combustion gas between the engine and turbine. The disadvantage would be potential thermal problems in the engine inlet valve due to the higher air temperature. Fig. 2. Schematic Diagram of a CAES Plant PERFORMANCE EVALUATION Using Compound Equipment The proposed reciprocating engines would function only as a single component During the compression mode of opera- of a CAES plant. Therefore, engine perfor- tion, valves C, D and G are closed, both mance will be given in the context of over- M/G units operate as motors receiving power all plant operation as well as in terms of from a base plant, coupling B is disconnec- engine efficiency, temperature, and pres- ted, and the HP engine operates a compres- sure when operating as a power generator or sor. During expansion, valves B, F and E compressor. Comparisons will be made of are closed, both M/G units operate as gen- CAES plants having the proposed engines, erators, coupling A is disconnected, and the Huntorf facility,2 which is the world's the HP engine operates as an expander. In first CAES plant, and two other turbine- both modes of operation, cooling water cir- based designs — one near-term and the culates through the HP engine. other an advanced design.3 This plant configuration will have MEASURES OF PERFORMANCE essentially the same main air flow rate as that of a plant using reversible HP and LP The following parameters were selected engines with some additional turbine to measure CAES plant performance: cooling air possibly being required. Thus, by using compound equipment, the size of • Specific Air Flow Rate - the amount of the reservoir can be significantly reduced reservoir air required per unit of leading to significant savings. generated power. However, the possible cost advantage • Specific Heat Rate - the product of of reversible equipment cannot be entirely specific fuel consumption and the realized with this configuration. The cost lower heating value of the fuel. advantages of completely reversible equip- • Expansion Power/Compression Power ment could be realized using a HP expander/ compressor engine and a reversible LP tur- • Overall Plant Efficiency - plant out- bine/compressor. The latter equipment put per total energy input. could be a radial-bladed, centrifugal, single or multi-stage compressor which can THERMODYNAMIC ANALYSIS be reversed to function as a radial turbine. A thermodynamic analysis was conducted ALTERNATIVE COMPOUND ENGINE PLANT of the plant schematically illustrated in Fig. I. : The analysis considered two A possible alternative to the engine/ types of power generation — constant pres- turbine system depicted in Fig. 2 involves sure and pressure-limited combustion. 429 The analysis was extended to consider extensions to the state-of-the-art of gas plant performance using compound engines. turbine technology. The performance of the turbines and com- pressors was based on Ref. 3. The performance of both reciprocating engine combustion processes is seen to be Simplifying assumptions used in the very similar. The chief difference between analysis include: the two processes is the higher maximum operating temperature and pressures during (1) thermodynamic equilibrium during com- pressure-limited combustion. bustion, It is seen that the reciprocating (2) reservoir pressure remains constant engine concepts have significantly lower at 4.48 MPa (this condition would be specific air flow rates than the turbine- approximated by a water-compensated based designs. This situation is reflec- reservoir), ted in the favorable values shown for the (3) reservoir temperature equal to 322°K, ratio of expansion power to compression power and overall plant efficiency. Sev- (4) negligible pressure losses in piping, eral factors enter into the calculated low (5) negligible heat transfer in system specific air flow value for the engine, Piping, with the most important of these being the air/fuel ratio. Actual air/fuel values will (6) recuperator effectiveness of 0.8, be dictated by engine operation — that is, (7) fuel being octane (C8H16) with a by the need to limit emissions or to lower lower heating value of 44.4 MJ temperatures. Other influences on specific (19,000 Btu/lbm), and kg air flow are shown in Figs. 3 and 4. The important influence of heat loss to cylin- (8) turbine efficiency of 90%. der-jacket cooling water is shown in Fig. 4. Cylinder heat transfer in the proposed DISCUSSION OF PERFORMANCE engine will be quite different from conven- tional engines due to the use of pressur- The three turbine-based CAES plant ized reservoir air. Actual values of heat designs selected for performance compari- transfer will need to be measured experi- son with the proposed reciprocating engine mentally, although improved methods of concepts are listed in Table 1. The analysis can greatly reduce the uncertainty Huntorf plant is included for the obvious in the value of this parameter. The in- reason that it is the world's only opera- fluence of reservoir pressure on air flow tional CAES plant. The principal reason is shown in Fig. 3. As can be seen, higher that the performance of the Huntorf plant reservoir pressure leads to reduced flow is lower than the two other systems is that rate, although the change will be only presently the plant does not include a re- cuperator. Addition of a recuperator would 5-7% over the range of expected reservoir bring all system performance parameters pressures. closer to those of the near-term turbine system. The plant heat rate is lower for the reciprocating engine concepts than the un- recuperated Huntorf design, but higher than The near-term system can be imple- either of the two turbine-based designs. mented using conventional turbomachinery. This higher heat rate is partially caused In this system, a modified steam turbine by the higher operating temperatures of the can be used for the 1000°F HP turbine. The reciprocating engines. The advanced, high- 1600°F LP turbine can be a modified gas temperature turbine system has a higher turbine from a peaker unit. This turbine heat rate than the low-temperature system will not require vane or blade cooling but for the same reason; that is, fuel is re- will require some auxiliary cooling for the quired to raise combustion temperatures. disk, blade and vane attachments, air seals, The incentive for using this turbine de- and bearing buffers. sign is to improve all other performance parameters. The heat rate is shown to be The adva'noed-turbine system will re- strongly influenced by cylinder heat trans- quire significant air cooling of the HP fer in Fig. 4. High reservoir pressure is turbine to operate at a gas inlet tempera- seen to provide lower heat rates in Fig. 3. ture of 2400°F; the LP turbine will be un- This result is due to the additional energy cooled. The two turbines for this system provided by the high-pressure air. do not exist but represent attainable 430 Table 1. CAES Plant Performance Comparisons Plant Using Plant Using Plant Using Plant Using Plant Using Constant- Pressure- Compound Near-Term Advanced Pressure Limited Engines Huntorf Turbine System Turbine System Combustion Combustion HP Engine Plant (1000,1600°F)a I[1500,2400°F)a Engines Engines LP Turblneb Reservoir Pressure, MPa 4.48C 4.48 4.48 4.48 4.48 4.48 (psla) (650) (650) (650) (650) (650) (650) Specific Air Flow Rate, kg/MJ 1.46 1.42 1.32 .656 .659 .914 (lbn/kWh) (11.59) (11.2) (10.4) (5.21) (5.23) (7.26) Specific Heat Rate, J/J 1.61 1.12 1.35 1.46 1.46 1.41 (Btu/kWh) (5500) (3808) (4600) (4972) (4994) (4805) Expansion Power/Compression 1.24 1.38 1.65 3.30 3.29 2.37 Overall Efficiency, Z P°"er 41.5 54.2 51 56.8 56.6 54.4 Inlet temperature to HP and LP Engine Characteristics: turbine, respectively. Max., temp., K 2136 2552 1995 (3885) (4134) (3130) 2000°F inlet gas temperature. (°F) Max., pressure, MPa 4.48 13.8 4.48 Inlet pressure to HP turbine. (psia) (650) (2000) (650) Air/fuel ratio 20 20 22.6 CONSTANT-PRESSURE COMBUSTION /_ 0.72 -1.5 CONSTANT-PRESSURE COMBUSTION ~A/F»20 0.74 1.60 - A/F«20 RESERVOIR PRESSURE»4.48 MPo/ / HEAT L0SS-25M.HV 0.68 - S /-~ 1.4 0.72 i / / ~0.64 1.3 £ o / e _j i. £0.60 i.2 i u |0.56 - 1.1 £ v> EO EXPECTED 0.62 - 0.52 — VARIABLE — -1— 0 RANGE 1.30 run 1 1 1 1 1 1 0.9 4 6 8 10 4 8 12 16 20 24 28 RESERVOIR PRESSURE, MPa HEAT LOSS TO JACKET WATER, %LHV Fig. 3. Influence of Reservoir Pressure Fig. 4. Influence of Jacket Water Heat on Specific Air Flow Rate and Loss on Specific Air Flow Rate Specific Heat Rate and Specific Heat Rate Figure 5 is presented to indicate Throttling of the air flow would be elimi- operational flexibility possible using the nated and the reservoir size would be re- pressure-limited combustion process. It duced. is seen that a constant maximum pressure can be maintained over a wide range of The use of a compound engine is seen reservoir pressure. The penalty is an in- to offer performance advantages over the crease in maximum engine temperature and use of an all-turbine system. As would be heat rate. This suggests that the engine expected, plant performance using the com- could be used with aquifer or salt-cavern pound engine falls between the values using storage to provide constant plant output the all-turbine or all-reciprocating system. with varying reservoir pressure by con- The compound engine operates at a signifi- trolling the fuel injection rate. cantly lower specific air flow rate than 431 any of the turbine systems, which results The compound engine would consist of in a higher power ratio and overall plant eight Delaval Model RV-20-4 diesel engines efficiency. operating in parallel and feeding the LP turbine. This model is a large, slow speed (450 rpm) engine designed for sta- tionary power or marine application. The PRESSURE-LIMITED COMBUSTION cost of $46/kW listed in Table 2 includes MAXIMUM PRESSURE. 13.6 MPo all auxiliaries and accessories for a stan- A/Ft 20 dard installation and will, therefore, .3000 HEAT LOSS'20% IHV yield a conservative cost estimate since many standard items (such as intercoolers I and turbochargers) will not be needed in 12800 this particular application. In addition, some reduction in price should result from 5 large-quantity production. In fact, the estimate may be even more conservative g2600 since the engine eventually chosen for this application may be a lower-priced, compact, high-production engine of less rugged de- sign such as used in diesel locomotives. £2400 The cost of the compressor system to supply about 375 kg/sec (827 lb/sec) of air 2200 at 5.92 MPa (865 psia) to the storage res- 4 6 6 10 ervoir is estimated from Ref. 4 to be $18/ RESERVOIR PRESSURE, MPo kW of power input. Assuming an expansion power to compression power ratio of 2.37 Fig. 5. Effect of Reservoir Pressure on (i.e., the value listed in Table 1), the Maximum Engine Temperature total cost of the expansion/compression equipment for the compound-engine concept can be calculated as $41/kW. However, PRELIMINARY COST ESTIMATES this estimate does not consider the pos- sible cost reduction using the engines as COST OF COMPOUND ENGINE reversible, HP reciprocating compressors. From Ref. 3, the expansion power to com- Performance and cost data of a HP tur- pression power ratio for the all-turbine bine, LP turbine, and reciprocating engine system can be calculated as 1.67. Using that can be combined to function as a tur- this value, the total cost of the expan- bine system or compound engine for a CAES sion/compression equipment for the all- plant are presented in Table 2. The tur- turbine system can be calculated as $38/ bines were selected to match the flow re- kW. quirements of the reciprocating engine. Table 2. Performance and Cost of Expanders HP HP LP Reciprocating Turbine Turbine Enginea Inlet Pressure, MPa (psia) 5.92(865) 1.15(167) 5.92(865) Inlet or Maximum Temperature, K (°F) 1367(2000) 1367(2000) 2100(3324) Total Mass Flow (including cooling air), kg/sec (lb/sec) 296(653) 375(827) 296(653) Net Power Output, MW 112 200 238 Cost, $/kW 41 19 46 8 - Delaval Model RV-20-4 Diesel Engines 432 COST OF REVERSIBLE-RECIPROCATING ENGINES The operating cost is equal to the sum of the capital charge, fuel cost, elec- As previously discussed, a CAES plant tricity cost, and operating and maintenance using reversible-reciprocating engines for costs. By definition, the reduction in the power generation and compression could be capital charge is equal to that of the to- designed (Fig. 1). It is demonstrated in tal capital cost. The fuel cost of the Ref. 1 that four LP expander engines would compound engine is about 26% greater than be needed to match the volume flow of each that of the turbine system because of its HP expander. Therefore, using the Delaval higher heat rate. Offsetting this increase engines, power generation would require is a reduction of about 42% in the electri- thirty-two LP engines for every eight HP city cost. The operating and maintenance engines. The LP engines would not require charge of 3.0 mills/kWh for the compound fuel injection and thus should be slightly engine represents a conservative estimate lower in cost than the HP engines. Assu- based on Ref. 6. The total operating cost ming the same power output as the compound of a CAES plant using a compound engine is engine (i.e., 238 MW for HP expansion and about 8% less than that of a plant using 200 MW for LP expansion) and equal cost a near-term turbine system. for the HP and LP expander engines, the total capital cost of the engines would be OTHER APPLICATIONS $109/kW. Some additional cost would be OF RECIPROCATING ENGINES necessary to make the engines reversible so that they could operate as reciprocating The performance and cost estimates compressors. Thus, the cost of the expan- presented in this paper are based on large- sion/compression equipment would be greater scale peak-power generation (e.g., greater than $109/kW. than 200 MW) for electric utilities. How- ever, reciprocating engines are also adap- This estimated cost is much greater table to store energy in the form of com- than that of the compound-engine or the pressed air using dispersed solar or wind all-turbine system. The conclusion is energy conversion systems. These systems that reversible-reciprocating engines for could generate 1 kW - 10 MW of peak power. both high- and low-pressure duty should not be used. CONCLUSIONS COST COMPARISON BETWEEN A NEAR-TERM The use of reciprocating engines in TURBINE SYSTEM AND COMPOUND ENGINE CAES plants has been evaluated in this paper. Two plant configurations were con- A comparison of the capital and oper- sidered — one using reversible-expander/ ating costs of a CAES plant using a near- compressor engines and a plant using a term turbine system and one using a com- compound engine concept composed of recip- pound engine is presented in Table 3. In rocating engines and turbomachinery. The both plant designs, compression is accom- expected advantages compared to turbine- plished using turbocompressors. based designs are reduced reservoir size and cost, reduced compression energy re- A water-compensated mined cavern was quirements, increased overall plant effi- selected as the air storage reservoir. ciency, and reduced capital and operating The operating characteristics of the res- costs. ervoir were 20-hr nominal storage (as de- fined in Ref. 5), 2190-hr generation time, It is shown that using reversible- and a storage pressure of 4.48 MPa (650 expander/compressor engines in place of a psia). The storage-cavern and surface- near-term turbine system reduces the re- reservoir costs were estimated using data quired air flow by about 53%, decreases from Ref. 5. The cost of the balance of compression energy by about 58%, and in- plant and the capital cost of the expan- creases overall plant efficiency by about sion/compression equipment were assumed to 5%. However, all of the cost advantages be the same for both plant designs. achievable from this improved performance are more than offset by the increase in the The use of the compound engine results cost of the engines. It was estimated that in a cost reduction of about 25% for the the cost of the engines would be about storage cavern and surface reservoir. This three times that of a turbine-based system. reduction is due to the lower air-flow re- This result is due to the high cost of quirement of the compound engine. The to- using reciprocating engines for low-pres- tal capital cost is reduced to about 11%. sure expansion. Thus reversible, 433 Table 3. Cost Comparison Between Near-Term Turbine System and Compound Engine Water-compensated mined cavern Storage pressure = 4.48 MPa (650 psia) 20-hr nominal storage; 2190-hr/yr generation time Near-Term Compound Turbine System Engine Capital Cost, $/kW Expansion/Compression Equipment3 41 41 Storage Cavern 89 66 Surface Reservoir 8 6 Balance of Plant 80 80 Indirect Costs'3 136 121 Total $354/kW $314/kW Difference -11.3% Operating Cost, mills/kWhC Capital Charge 29.5 26.2 Fuel 9.5 12.0 Electricity 10.8 6.3 Operating and Maintenance 2.0 • 3.0 Total 51.8 mills/kWh 47.5 mills/kWh Difference -8.3% Engines and/or turbines and compressors 15% contingency, 10% engineering and administration, 30% escalation and interest during construction. CCapital charge - 18%/yr; fuel - $2.50/106 Btu; electricity - 15 mills/kWh. reciprocating, expander/compressor engines work is planned to evaluate the economics should not be used for both the high- and of the substitution of the high-pressure, low-pressure duties of a CAES plant. reciprocating engine for a gas turbine in a CAES plant. In addition, the use of re- The compound engine evaluated is com- ciprocating engines in dispersed solar or posed of a reciprocating engine for high- wind energy conversion systems will be pressure expansion, a gas turbine for low- evaluated. pressure expansion, and turbocompressors for compression. Comparing a plant using ACKNOWLEDGMENTS a compound engine concept to one using a near-turbine system, the air flow is re- The research activities in compressed duced about 35%, compression energy re- air energy storage, which formed the basis quirements are decreased by about 42%, and of this paper, were funded by the Division the overall plant efficiency remains about of Energy Storage Systems, U.S. Department the same. However, the capital cost of the of Energy. expansion/compression equipment is about the same (or possibly lower) than the tur- REFERENCES bomachinery of the near-term system. It is estimated that the capital and operating 'Kartsounes, G.T., and J.G. Daley, The Use costs are reduced about 11 and 8%, respec- of Reciprocating Engines in Compressed Air tively. Energy Storage Power Plants, Proc. of the 1978 Compressed Air Energy Storage Tech- The evaluation presented in this paper nology Symposium (May 1978). indicates that a considerable economic ad- vantage can be realized using a compound engine concept in a CAES plant. Further 434 2Stys, Z.S., Air Storage System Energy Transfer (ASSET) - Huntorf Experience, ERDA/EPRI CAES Workshop (Dec. 1975). 3KIm, C.S., and G.T. Kartsounes, A •Para- metric Analysis of Turbomaahinery Options for Compressed Air Energy Storage Plants, Proc. of the 1978 Compressed Air Energy Storage Technology Symposium (May 1978). ''Davison, W.R.., and R.D. Lessary, Study of Selected Turbomaahinery Components for Compressed Air Energy Storage Systems, prepared by United Technologies Research Center for Argonne National Laboratory, Report ANL/EES-TM-14 (Nov. 1977). 5Giramonti, A.J., Preliminary Feasibility Evaluation of Compressed Air Energy Stor- age Power Plants, United Technologies Research Center, R76-952161-5 (Dec. 1975). 6Raymond, R.J., et al., Cost and Applica- tion of Coal Burning Diesel Power Plants: A Preliminary Assessment, prepared by Thermoelectron Corporation for National Science Foundation, NSF 75-SP-0917 (Aug. 1975). 435 SESSION VIII: COMPRESSED.AIR ENERGY STORAGE 437 PROJECT SUMMARY Project Title: Compressed Air Energy Storage—Advanced CAES System. Studies Principal Investigators: W. V. Loscutoff, M. A. McKinnon Organization: Pacific Northwest Laboratory PO Box 999 Rich!and, WA 99352 Telephone: (5095 946-2768 Project Objective: Develop advanced CAES systems that require little or no firing of the gas turbine with natural gas or oil. Project Status: Alternate technologies and fuels are being investigated to determine their potential for reducing the use of or replacing premium fuels at a CAES plant. Among the Alternative fuel systems beinq investigated are coal fired magnetohydrodynamics, fluidized bed combustion, nuclear waste decay heat, and coal gasification. Alternate technologies investigated include the use of hybrid systems using varying degrees of thermal energy storage and adiabatic systems. The following concepts have been found to hold the greatest promise: coal gasification, fluidized bed combustion, and thermal energy storage. PNL is directing research at United Technologies Research Center to study fluidized beds. PNL will concentrate its effort on the study of utilization of thermal energy storage with CAES. This includes Dlanned activity with a utility. The primary responsibility for integration of coal gasification with CAES lies with EPRI. PNL will maintain information exchange with EPRI on this subject. Contract Number: EY-76-C-06-1830 Contract Period: FY 1978, continuing Funding Level: $150,000 60 Funding Source: Department of Energy, Division of Energy Storage Systems 439 ADVANCED CAES SYSTEM STUDIES M. A. McKinnon Pacific Northwest Laboratory PO Box 999 Richland, Washington 99352 ABSTRACT The primary objective of this program during FY-1978 was to evaluate and screen a number of advanced concepts available for integration with compressed air energy storage (CAES) systems. This paper summarizes the principal results and conclusions reached from a preliminary assessment of four concepts for heating air without using premium fuels. The concepts considered were Nuclear Waste Heat Augmented CAES, Magnetohydrodynamics (MHD) combined with CAES, Coal Gasification used with CAES, and Fluidized Bed Combustion (FBC) used with CAES. In addition to the above studies a comparative economic analysis of CAES in a hard rock cavern was made for three systems, a conventional fired system, a hybrid system using a single stage of thermal energy storage (TES), and a no fuel double stage TES system. Conclusions reached from the studies are: 1) nuclear waste decay heat utilization in conjunction with a CAES system is technically feasible but the potential fuel savings will not compensate for environmental concerns or justify the additional system complexities, 2) .MHD has potential for a good match with CAES but should not be studied further until the MHD problems have been solved, 3) Coal Gasification and FBC require more study, 4) Hybrid CAES systems and no fuel CAES systems using TES can be competitive with conventional CAES if there is a large enough differential between compression power costs and fuel costs. INTRODUCTION and without augmentation by thermal energy storage (TES). In addition, PNL examined Large scale energy storage systems how nuclear waste heat, magnetohydrodynamics, are useful in an electric utility grid as fluidized bed combustion and coal gasifi- a means.for load leveling, i.e., storing cation may be integrated with CAES. This energy during periods of excess capacity paper reports the results of those studies. for use during periods of excess demand. There are several alternatives for INCREMENTAL COST ANALYSIS storing energy under investigation. Compressed Air Energy Storage (CAES) is The FY 1977 progress report for CAES one of these alternatives that appears advanced concepts concluded that turbine to be economically and technically viable. fuel could be saved if Thermal Energy There is however a potential long term Storage (TES) was added, but at the weakness of conventional CAES concepts. cost of increased system size and It is the reliance of conventional CAtS complexity, reduced overall energy utili- on clean petroleum fuels whose future zation efficiency, and increased capital availability for power generation is cost. The preliminary economic analysis under question due to increasing demand, concluded that there was currently no decreasing domestic reserves and the apparent economic incentive to include TES uncertainty of potential import embargos. in a CAES system. However, because of the preliminary nature of the analysis and The objective of the advanced CAES because of the minor cost savings of concept studies at Pacific Northwest non TES systems over that of systems using Laboratories is to find and develop TES, additional economic analysis was CAES systems whose dependence on petroleum recommended. This section of the report derived fuels is minimized or entirely gives the results of that analysis. eliminated. During this past year five such systems have been investigated. CAPITAL COSTS A comparative economic analysis was made of CAES systems at a hard rock site with This analysis considered the three 440 systems shown schematically in Figures 1, 2 and 3. The first figure is that for the reference or base CAES system. The second and third are for the hybrid and the no- fuel systems, respectively. The operating and performance variables used in the analysis are shown in Table 1 and were taken from reference 1. The no-fuel system was selected such that all of the compression energy would be saved. We had felt that the no-fuel system would not look economically competitive; however, as will be seen from this study, chat is not neces- sarily the case. The no-fuel system can look good under the proper set of condi- tions. Fig. 2. High Pressure CAES Cycle With Recuperation and Single Stage TES/Regeneration Table 1. Operating and performance Variables* Reference Hybrid No Fuel System/State Parameter Units CAES System . System TES System OPERATING VARIABLES LPC (Axial Outlet Temperature °F 437 700 896 HPC (Centrifugal) Outlet Temperature °F 437 4Sil 525 HPT Inlet Temperature °F 1,000 1.000 502 LPT Inlet Temperature °F 1,500 1,500 855 HPT Expansion Ratio 3.1 5.8 6.1 Fig. 1. Base System High Pressure CAES PT Expansion Ratio 15.0 8.0 7.6 Cycle PERFORMANCE VARIABLES Compressor Specific Mass Flow Rate lb/kwh 13.5 12.2 11.0 Costs generated for the reference CAES Turbine Specific Mass Flow Rate lb/kUh 10.6 10.6 16.0 system were based on estimates developed Coefficient of (kWOout in a recent Acres American, Inc. feasibility Performance (klihiln 1.27 1.15 0.69 study of CAES for peak shaving in Califor- nia.2 These estimates are also used as a Turbine Heat Rate Btu/kHh 3.771 3.279 basis for developing the costs of the TES System Heat Rate Btu/kHb 11.420 11,810 14,100 systems. Where equipment requirements •F-Y-iy77 Progress Report Compressed Air Enerqy Storaqe Advanced Systems are similar to those of the base case but of a different size, costs were assumed to be directly proportional to size. Where new equipment was used, other cost sources A capital cost comparison of the were sought. Becaus? this is an incremental three systems is given in Table 2. cost analysis, the primary focus of the These costs assume a 750 MW CAES facility cost comparison is on capital equipment constructed over 6.5 years with the plant differences and on the differences in going on line in 1984. All costs were turbine fuel and compression energy use. price level adjusted to represent the total More interest was shown in relative costs installed cost in January, 1985 dollars/kW. than in absolute costs. 441 Table 2. Capital Cost Comparison ($106) Reference Mo FwT CMS 5y»tt» Hybrid !nl» TES S»it» Land, Structures tnd Plint i S.5 $ 5.5 $ 5.2 Equipment Turbo Machinery Intercoolers 1.7 1.2 0.4 Aftereoolers 0.8 . Compressor Recuperator - 22.0 - Compressors 25.0 25.0 37.5 Compressor Motors 17.3 19.2 31.9 Turbines 18.4 18.4 27.5 Turbine Exhaust Recuperator 8.5 8.5 - Other Turbo Machinery 27.7 27.7 26.2 Storage Facilities Air Storage s Associated Costs 34.5 34.5 51.8 Thenui Storage - 2.0 6.1 Fuel Storage 0.4 0.4 - Electrical and Miscellaneous _ .1.0.6. .10.6 $150.4 $175.1 $197.3 Engineering and Construction ManigeMnt (151 of Direct Costs) 26.3 $173.0 $201.4 $226.9 Contingency (15X of Above) 25.9 30.2 34.0 Escalation During Construction 50.4 58.7 66.1 Interest Durinq Construction 31.4 36.6 41.2 July 1984 $ «280.7 $326.9 $368.2 January 1985 S 287.8 335.1 377.4 Plant Capacity 750 MW 750 HI 750 HI Cost/kN 1985 S S384/W S447AH $S03/kH Table 3. Reference Case Assumptions for Comparative Analysis Plant Storage Capacity 10 hours Plant Capacity 750 MU Operation and Maintenance Costs Fixed $4.85/kW(a) Variable .32 m1]ls/kWh(a) Operation and Maintenance Cost Escalation Rate 5.5I/yr Fuel Cost $4.00/10.6 Btula) Fuel Cost Escalation 7.0%/yr Compression Energy Cost 30mills/kHhla> Compression Energy Cost Escalation Rate 7.0J/yr Fig. 3. No-Fuel TES High Pressure CAES Cost of Money 12* System Fixed Charge Rate 15.5% Life of Plant 35yrs Calculations for the base CAES Capital Cost Escalation Rate 51 system cost of electricity and the Capacity Factor 251 other incremental costs are based on the la)19B5 Price Levels same assumptions as were used in the Acres American study. These assumptions are listed in Table 3. Using these assump- Table 4. Levelized Cost of Electricity tions, the levelized cost of electricity for CAES Systems for the three CAES systems were calculated. These costs are shown in Table 4. Costs (mills/Ml) Cost Reference Hybrid No Fuel The sensitivity of the levelized cost Component CAES System System TES System of electricity to capacity factor, turbine Capital 27.2 32.7 35.6 fuel cost, and compression energy cost is 0 S « 4.5 4.5 4.5 shown in Figures 4, 5, and 6. Figures 4, Turbine Fuel 31.5 27.4 0 Compression Energy 5 and 6 can be confusing unless one 19.3 54.7 91.2 recognizes that the power cost from the Total 11?.5 118.3 131.3 base CAES system (represented by a dashed line at zero in the figures) increases with 442 4Q • REFERENCE CASE • REFERENCE CASE 30MILUKW-H 30MILL/KW-H $4.0410*) BTU 30 / NO-FUEL/TES 20 / HYBRID . 10 0 1 1 i -10 -10 20 30 40 50 60 CAPACITY FACTOR, * COMPRESSION ENERGY COST, MILLS /KW-HR 1657) (986) (1,314) 11,643) (1.971) (2,3001 (UTILIZATION RATE, 10° KW-HR/VEAR) Fig. 6. Compression Energy Cost Fig. 4. Capacity Factor (Utilization Rate) Sensitivity Sensitivity energy cost of 20 mills/kW-hr were assumed, instead of the 30 mills/kW-hr that was 40 used to obtain the results for Figures 4, • REFERENCE CASE 30MILLKW-H 5, and 6 the no-fuel system would be more M.O/UO6! BTU economical than the reference system for fuel costs greater than $4.75/106 Btu. NO-FUEL/IES This study has shown that a particular no-fuel system can look attractive if the differential between compression energy and turbine fuel costs is large enough. However, that no-fuel system had assumed a compressor outlet temperature that is beyond the limits of current technology. Even so, we are encouraged by the results and feel that additional analysis should -10 be conducted on no-fuel systems that use 3.00 3.50 4.00 4.50 5.00 5.50 equipment that represents current techno- FUll COST, * HO4 81U logy. Fig. 5. Fuel Cost Sensitivity UTILIZATION OF NUCLEAR WASTE DECAY HEAT fuel cost and compression energy cost and A preliminary assessment has been decreases with increased capacity factor. performed of the utilization of decay These three figures only look at the heat from canisters containing solidified incremental cost above that of the base waste fission products to augment a CAES system. As seen from these figures, compressed air energy storage (CAES) the incremental cost for this particular system. The primary focus of the study hybrid system does not reach zero for any was to perform a technical assessment of value of the three parameters examined. the concept and then, if the results However, the variance does not become were sufficiently positive, to follow very large over the range of values studied. that study with a detailed economic The incremental cost for the no-fuel TES assessment. The conclusion reached system exhibits a much higher degree of is that even though a CAES system augmented sensitivity. For compression energy costs by nuclear waste decay heat may be techni- less than 15 mills/kW-hr in 1985, this cally feasible, the potential fuel savings particular no-fuel system could operate will not be sufficient to compensate for less expensively than the reference CAES environmental concerns or to justify the system. In addition, if a 1985 compression additional system complexities. 443 The expansion side of the system used populous area would present a number of for the analysis is shown schematically in drawbacks. In addition to considering Figure 7. It uses exhaust heat recuperation safety when selecting a site, the potential and two stages of expansion. The compres- impact on the environment must be taken sion side of this system is the same as into account. Problems with licensing shown in Figure 1. The nuclear waste is will also undoubtedly arise. contained in a secondary loop and heat is recovered by means of a heat exchanger. The obstacles and costs associated Calculations performed based on heat trans- with the utilization of nuclear waste fer considerations indicated that the decay heat with CAES appear to be much nuclear waste would have to be contained greater than an alternate system, CAES in finned canisters in order to obtain the with TES. The nuclear system will suffer required heat transfer and to maintain high penalties and costs due to finned canisters, air temperatures. Current canister techno- remote handling, siting and licensing. logy, even though ill defined, restricts In addition, the nuclear system does not cylinder designs to 12 to 24 inch dia- use the compression heat. meter, 10 to 15 feet lengths, maximum skin temperatures between 500°F and 800°F, heat The TES system can store the compression loadings of 3.5 kW/can to 13 kW/can, and heat that the nuclear waste system does maximum allowable cyclic temperature fluc- not use. The temperature level of the tuation of 200°F. compression heat is comparable to the allowable skin temperature of a canister in a nuclear system. A TES system can be very simple, have a low capital cost, require minimal maintenance, and have insignificant impact on the environment. A nuclear system will not be able to com- pete with a simple TES system. MAGNETOHYDRODYNAMICS Magnetohydrodynamics (MHD) was found to be a prime candidate for CAES applica- tions since approximately twenty percent of its output is used to compress air. Optimum use of MHD and CAES requires that they are located at the same facility. This would also allow the heat of compres- sion of the CAES system to be used to heat feedwater for the MHD bottoming cycle. Fig. 7. Expansion Side of a CAES System Utilizing Nuclear Waste Decay Heat CAES incorporated with MHD would and Turbine Exhaust Recuperation allow a facility to have base and inter- mediate load capability. The facility The cycle calculations performed may be able to supply peak loads also. indicate that the use of waste decay The system would derive all of the air heat can decrease fuel consumption storage compression power internally and by up to 18%. This fuel savings would not have to rely on power from the must be balanced against the cost of utility grid during off-peak periods. The adding fins to the canisters and ultimate goal of the continuously operated providing waste storage caverns with CAES/MHD plant might be to supply the remote handling equipment. This entire spectrum of the demand curve and to system will also need air circulating do so at efficiencies close to that of base- equipment, heat exchangers and air load power generation. monitoring equipment Presently the MHD system with best Location is another factor that compatibility with CAES is an open-cycle works against the concept. Ideally, MHD with a steam bottoming plant as a CAES plant should be situated near a shown schematically in Figure 8. To load center, thereby minimizing trans- combine the system of Figure 8 with CAES, mission losses. However, locating a the compressor must be replaced by a CAES nuclear waste storage facility near a plant which includes an expander turbine 444 ELECTRICAL STACK During the previous fiscal year a SEED POWER-OUOWER T GAS preliminary study of Fluidized Bed Combus- _L tion (FBC) integrated with CAES was FUEL COMBUSTION MHD initiated. The material reviewed indicated CHAMBER GENERATOR that FBCs could provide a clean efficient means of burning coal without the need for AIR stack gas scrubbing. FBC was also reported PREHEATER" to have low NO emissions and reduced carry STEAM over of volatiTes. The bed's temperature GENERATORS also eliminates slagging problems and ELECTRICAL produces a soft unsintered ash. POWER-OUT These positive findings prompted additional study of FBCs. A request for THREE-PHASE COMPRESSOI STEAM proposals was advertised and an evaluation GENERATOR TURBINE •• WATER of the responses has been completed. A contract with United Technology Research CONDENSER Center to further that study is being finalized. The overall contracted effort A R consists of four tasks which are: INLET •=FEE0D PUMP . Review and assess the state-of-the- art of pressurized and atmospheric Fig. 8. Open Cycle MHD with a Steam FBC as they relate to CAES applications. Bottoming Plant . Develop schematic diagrams of possible whose outlet pressure is ten atmospheres. FBC/CAES configurations. Perform a During retrieval of air from storage, the preliminary screening analysis to ten atmosphere expander turbine exhaust' identify the system or systems which will become the air supply to the combus- have the greatest potential for tion chamber of the MHD system. Such a successful development as near term CAES/MHD combination would allow an MHD CAES peaking plants. plant with a base output of 970 MW to become a plant that can deliver a range of . Develop a pr.conceptual design of the power from 628 MW to 1346 MW. Preliminary best system as determined from the calculations indicate the combined CAES/MHD screening analysis. Determine avail- plant can supply mid range power at effi- ability and estimate cost of system ciencies close to base load efficiencies. components. Before MHD/CAES can be demonstrated, . Project the potential market available MHD will have to be successfully demon- to a CAES/FBC system. strated on a commerical scale. Many MHD problems remain to be solved prior to Meanwhile, additional studies of FBCs that time. The most obvious and hardest have been conducted at PNL to identify MHD problems are associated with high potential problems or limitations imposed temperature materials for electrodes. by combining FBC with CAES. The major Several other problems which are currently problems are associated with either the preventing MHD from being a near term design and cost of a compatible heat option for power generation are: end exchanger for an atmospheric FBC (AFBC) or current losses, electrode arcing, large the particulate clean-up problems needed scale DC-AC power conversion, avoidance of to insure the survivability of the turbines shocks, and seed separation. in a pressurized FBC (PFB) system. The technology to clean the gas at high temper- Our preliminary evaluation indicates atures does not exist at the present time. that MHD would be a good companion for CAES, however, we recommend that addi- Most of the available literature on tional development of MHO/CAES systems AFBCs addresses their use as boilers in a be delayed until most of the MHD problems steam system. As boilers the systems can are solved. be small due to optimum use of the enhanced heat transfer coefficient in the fluidized FLUIDIZED BED COMBUSTION bed. The tube wall temperatures are kept 445 low by the cooling effect of the boiling multitude of different processes. Gasifiers water on the inside of the boiler's tubes. are generally classified according to The superheat region of the boiler can be their coal flow relative to the gas flow. located out of the bed. Under this system, three types can be identified: fixed bed, fluidized bed, and When an AFBC is used as an air heater entrained flow gasifiers. Other classifi- the tube wall temperatures pose more severe cation systems are also used, such as design problems. The cooling effect of pressure level, source of oxygen (air or the air on the inside of the air heater oxygen blown), number of stages, or ash tubes does not provide as great a cooling removal method. effect as boiling water; therefore, the tube wall temperatures and tube surface It is not the intent of this part of area can be expected to be higher. The the paper to provide a complete description higher wall temperature will demand more of gasifier processes but only to indicate exotic materials to withstand the pressures that there are many types to choose from. and corrosion potential. It appears that For more detailed information one is Incoloy 800 will be required for the heat referred.to the references on gasifica- exchanger so that wall temperatures of tion J3"6) 1200°F to 1300°F.can be attained. The increase in tube surface area requires Power generation systems that use more tubes, but the density of tubes is coal as their primary fuel are high in limited by its impact on the bed fluidizing capital cost with respect to oil and and mixing processes. Consequently, to natural gas systems. Coal gas fired get more tubes in the bed requires a bigger systems would be even higher in capital bed. cost. High capital costs and the inability for daily turn off of gasifiers dictates Material problems and bed size will be that the gasifier facility be operated included in next year's effort to assess the continuously and preferably at full viability of a FBC/CAES. capacity. The continuous operation would also help to defray their high capital COAL GASIFICATION cost. A preliminary study was made to assess The application of coal derived fuel the possibilities of augmenting CAES with gas to CAES has now been narrowed to three coal gasification. The study identified general alternatives. The options are as one gasification system that is recommended follows: for further study. . The gaseous fuel produced during Gasification is the process of con- compression periods would be sold verting solid coal to a clean gaseous fuel. and distributed to industrial users Most processes used air or oyxgen with steam that are located nearby. to react with the coal. The product xuel gas may have a heating value range of 100 . Excess fuel produced during compres- to 1000 Btu/scf depending on the process. sion periods would be stored to be The heating value of the product gas has used during generation periods. been divided into three major categories. They are: low-Btu gas (LBG) with heating . The turbines of the CAES operate values from 75 to 175 Btu/scf. Inter- continuously using all the fuel gas mediate-Btu gas (IBG) covers the 250 to as it is being produced. 400 Btu/scf range, and high-Btu or pipe- line quality gas (HBG) with a 925 to The first option has control difficul- 1000 btu/scf range. Air blown reactors ties and is limited to sites near large produce LBG whereas MBG is produced by industrial users. The user must be will- oxygen blown reactors. HBG requires expen- ing to use off-peak power. If he is willing sive shift reaction and methanation process- to use off-peak power then he could be used ing of MBG. In general, the thermal effi- to load level directly without the expense ciencies of the conversion process decrease of a coal gasification/CAES plant. with increasing heating values while the ease of storing and transporting the gas The second option requires storage increases. of gaseous fuel. The technical prob- lems associated with isolating gas and Gasifier development has produced a air storage could make site selection 446 very difficult. Current gasifiers are Assessment of Low- and Intermediate-Btu operated below 200 psi which would require Gasification of Coal, FE/1216-4, National large storage volumes for the gas or a Academy of Sciences, Washington, D.C., means of compressing the gas into smaller 1977. volumes. The additional gas storage cost added to the already high cost of the Handbook of Gasifiers and Gas Treatment gasifier makes the option less attractive Systems, Dravo Corp, Pittsburgh, PA, than the other options. 1976. The last alternate allows continuous operation of the gasifier without the need to sell or store the product gas. The main operational characteristic is continuous combustion of the product gas as it is being produced. The acronym CGC/CAES (Continuous Gasification and Consumption CAES) will be used to denote this alternate. This option is conceptually different in that the plant's output would change by switching compressor trains on or off. The compressors would be on during storage and off during operation. The turbines would be operated continuously. Conventional CAES switches on and off both the compressor and turbine trains. The CGC/CAES bypasses the major disad- vantages of other gasification options namely storing or selling fuel gas. It is this advantage along with the ability to accommodate a coal gasifier that operates continuously and at near full capacity that makes CGC/CAES the most attractive Coal Gas/CAES. The CGC/CAES concept will be studied in greater de- tail next year. The extent of the study will depend on the results and direction of the coal gas/CAES work being done by EPRI. REFERENCES 1. Kreid, D.K. and M. A. McKinnon, "FY-1977 Progress Report Compressed Air Energy Storage Advanced Concepts," Report No. PNL-2464/UC-946, March 1978. 2. M.J. Hobson, et al., "Feasibility of Compressed Air Energy Storage, as a Peak Shaving Technique in California," Acres American, Inc. for the California Energy Commission. May 1978. 3. Donald L. Katz, Dales E. Briggs, Edward R. Lady, John E. Powers, M. Rasin Tek, Brymer Williams, and Walter E. Lobo, Evaluation of Coal Conversion Processes to Provide Clean FueTsT EPRI 206-0-0 Report to Electric Power Research Institute by University of Michigan College of Engineering, 1974. 447 PROJECT SUMMARY Project Title: Application and Design Studies of Compressed Afr Energy Storage for Solar Energy Applications Principal Investigator: Gerard T. Flynn Organization: M.I.T. Lincoln Laboratory P- 0. Box 73 Lexington, MA 02173 617/862-5500 Ext. 7456 Project Goals: The objective is to investigate the applicability of compressed air energy storage in combination with thermal energy storage for use in electric utility power generation. The combination of solar thermal and off-peak energy will be used in determining cost and performance models. Par- ticular attention will be focused on the cost and performance of the packed beds for thermal eneroy storage. Project Status: Work nearly completed on the initial 15 month program. Final report is in preparation. Contract Number: EX-76-A-01-2295 Contract Period: Oct. 1978 - Jan. 1979 Funding Level: $150,000 Funding Source: U. S. Department of Energy 449 SOLAR THERMAL AUGMENTATION OF CAES* G. T. Flynn MIT/Lincoln Laboratory P.O. Box 73, Lexington, Massachusetts 02173 ABSTRACT This paper describes a storage system which combines an adiabatic compressed air storage plant with a solar thermal central receiver. A packed pebble bed is used to store the heat of compression generated during the charging of the compressed air reservoir. This heat is then returned to the air on discharge of the reservoir. This stored thermal energy minimizes the solar thermal energy required to raise the air temperature to 1500°F at the turbine inlet The system described is based on a central receiver design proposed by Boeing. The turbomachinery is within the state-of-the-art, but would be built specifically for this application. Economic screening curves are given which compare conventional and adiabatic CAES for both coal fired and nuclear base load pumping. This hybrid system is marginally competitive with conventional CAES on the basis of levelized fuel costs of $3.55 per MBtu and nuclear off-peak pumping costs of 7.0 mills per kWh. 1.0 INTRODUCTION before expansion in the turbine. However, an adiabatic system will always have an The storage of off-peak energy in the energy return ratio (efficiency in this form of compressed air could provide a case) of less than unity. The energy considerable saving in the fuel used for return ratio is defined as the energy peak power generation. An open-cycle gas delivered to load from storage divided by turbine peaking plant normally would the energy used for charging the storage. provide its own compressed air in a Brayton In the case of a fired CAES, the energy cycle. Approximately two-thirds of the return ratio could be as high as 1.4. power from the turbine is used to run the compressor, and the remainder is available If a hybrid scheme were used where for electric power generation. Therefore, thermal energy was added to an adiabatic the efficiency is typically 30% or less system before turbine expansion, consider- and the corresponding turbine heat rate is ably less fuel-derived thermal energy 12,500 Btu/kWh. would be used to achieve the same turbine inlet temperature utilized in fired CAES A CAES (Compressed Air Energy Storage) design. If the thermal energy is to be system which provides compressed air from provided by a solar central receiver, the an air storage reservoir can generate importance of minimizing the heat rate is power with a heat rate as low as 4000 even greater than for the fuel fired case Btu/kWh (1). If the compressed air reser- in order to minimize the capital cost of voir is pumped (charged) by nuclear or the solar central receiver and heliostat coal base load capacity during off-peak field. hours, CAES could reduce high distillate fuel consumption by a factor of three. 2.0 SYSTEM DESCRIPTION If the air were compressed isentro- Figure 1 shows the hybrid solar/CAES pically, i.e., without intercooling and system. Off-peak energy is stored as the heat of compression were stored in a compressed air in an underground cavern TES (Thermal. Energy Store) , the system at a pressure of 35 BAR (BAR=14.5 p.s.i.a.) could be operated adiabatically (2,3,4) In order to maintain a constant pressure and no additional heat would be required during charging and discharging, the cavern is hydraulically compensated with *This work was sponsored by the Division a water shaft to a surface sited pond. of Energy Storage Systems of the U.S. The depth of the cavern from the surface Department of Energy. 450 - If $ is determined by the system pressure re- 2.1 CAES CHARGING CONSIDERATIONS quirement. The depth would be 33.5 feet x storage pressure in BARs. This calcu- The first compressor increases the lation limits the system pressure to be pressure from atmospheric to 16 BAR, not greater than the local pore water which is a reasonable value for available pressure, which eliminates any possibility compressors operating without intercooling. of air leakage through saturated over- With an inlet temperature of 70°? (530°R), burden. At a pressure of 35 BARs, the the outlet temperature would be 780°F depth would be 1200 feet. (1240°R). This air must be cooled before entering the booster compressor. Rather than discarding the thermal energy, it is stored in a TES (Thermal Energy Store). The TES (5) consists of a large packed bed of pebbles. The air passing through transfers heat to the pebbles. The air velocity is quite slow and at any point in the bed, there is only a slight difference in temperature between the air and the pebbles. Therefore, as the air passes through the bed and heat is trans- ferred from air to rock, a steep thermo- cline (temperature gradient) is developed. This thermocline moves gradually to the bottom of the bed on charging. On dis- chaiging, air flows in the reverse direction through the bed and the thermo- Figure 1 Hybrid Solar/CAES System cline would move back to the top. Because of irreversibilities in the hfcat transfer, the thermocline would gradually disperse A packed pebble bed T£S is used to on successive charge/discharge cycles. As store the heat of compression £iom the this dispersion increases, the outlet first compressor which provides a com- temperature of the bed would rise towards pression of 16:1. The second booster the end of the charge cycle. To prevent compressor has a compression ratio of the inlet air temperature from exceeding 2.2:1 to provide the cavern storage pres- Lhe maximum allowable inlet temperature sure of 35 BAR. On discharge, air from to the booster compressor, the air is the cavern passes through the TES to the cooled by passing it through a pipe which central receiver. The central receiver runs through the watershaft to the surface. is a power tower design in which 35 acres The effect of this natural heat exchanger of reflecting mirrors concentrate sunlight has been calculated assuming a water tem- on a tower mounted heat exchanger. This perature of 68°F (528°R). With a maximum central receiver heat exchanger raises charging TES outlet temperature of 820°F the temperature of the air from 820°F (1280°R), the maximum booster compressor (1280°R) to 1500°F (1960°R) before inlet temperature would be 240°F (700°R). expansion in the turbine. There is no This is a sufficiently low inlet tempera- storage for the solar thermal system and ture to prevent damage to the booster a backup fuel fired combustion chamber is compressor. The amount of heat lost to postulated in the economic analysis to the water on each cycle is small and is allow for solar unavailability. accounted for in the TES efficiency rating of 90%. The compressor and turbine chains are connected through clutches to the motor The way the system has been config- generator such that either chain may be ured, it is not possible to utilize the appropriately connected on charge or heat of compression from the booster discharge. compressor. With an adiabatic efficiency of 75%, the thermal energy is 75.46 Btu/ lb or 13% of the turbine heat rate on discharge. The air in the cavern goes directly to the TES on discharge. If it were not cooled before storage, the lower 451 end of the bed would equilibrate at too 'or a turbine expansion ratio of 32:1, high a temperature to act as an effective an isentropic efficiency of 90%, and a intercooler on charging. Storing the air generator efficiency of 98%, the mass flow at L-gh temperatues also increases the through the turbine is 7.85 lbs/kWh. The volume required and thus would increase solar thermal heat rate is 2260 Btu/kWh, the capital cost. If the pipe from the which is equivalent to a "thermal effi- booster compressor to the cavern is run ciency" of 151%. This "thermal efficiency" through the watershaft to form a counter- is useful only in scaling the solar thermal flow heat exchanger, the air in the cavern portion of the system; that is, the turbine will be at an average temperature of 80°F generator efficiency in the Boeing design (54O°R). At this pressure and temperature, was 42%. With this hybrid, 3.6 times as the cavern volume required is 3.0 ft3/kWh much energy is produced with the same (discharge). scale central receiver and heliostac field. Alternately, the cost of the components in 2.2 CAES DISCHARGE CONSIDERATIONS the solar thermal system expressed in dollars per kWh are reduced by a factor On discharge, air leaving the cavern of 3.6. is piped directly to the lower end of the TES. The air temperature is raised from 3.0 ECONOMICS 80°F (540°R) to 820°F (1280cR) in passing through the packed bed. The discharge 3.1 CAPITAL COSTS temperature of the packed bed over the entire discharge interval is shown in The capital costs are based on a 100 Figure 2, The average TES discharge MWe, 12 hr storage system. The costs for temperature is 78O°F (1240°R). This air the solar thermal system have been taken is then piped to the solar central directly from the Eoeing cost estimate receiver (or a combustion chamber) where with appropriate scaling as described the temperature is raised to 1500°F above. A cost escalation of 8.5% per year (1960°R) before entering the expansion was used, and all estimates are expressed turbine. The solar thermal heat rate is in 1980 dollars. 288 Btu/lb of air. The central receiver chosen as a model for performance and cost The capital costs of the system fall was a Boeing design (6,7). The Boeing into three general categories: mining system is designed to operate at 34 BAR in and construction; solar central receiver a closed Brayton cycle. The basic design and heliostats; and rotating machinery of the system with some scaling as de- and balance of plant. scribed below has been taken intact with the closed cycle turbomachinery replaced 3.1.1 COST SUMMARY MINING AND CONSTRUCTION by the 32 BAR expansion turbine described below. $/kW Geology Land Site Improvements 7.00 TES DISCHARGE TEMPERATURE (*R) vsTIME FOR SUCCESSIVE Cavern Excavation @ $1.0/ft DISCHARGE PERIODS ON A DAILY CYCLE BASIS (12 hr) 36.00 Water Shaft 1..80 Air Shafts 3..40 Valves & Misc. Fittings 5..00 Thermal Energy Storage 35.00 5 HOI Pond (land & sealing) 2.80 $91.00 © DISCHARGE PERIOD No 1 Mining & Construction Subtotal = © DISCHARGE PERIOD No.5 STEADY STATE CONDITION 3.1.2 COST SUMMARY SOLAR THERMAL $/kW Land & Site Improvements 0.41 Structures & Facilities 20.73 Heliostats 121.92 Central Receiver 42.67 DISCHARGE TIME I HOURS) Fuel Storage 27.63 Figure 2 Solar Thermal Subtotal = S213.36 452 3.1.3 COST SUMMARY-ROTATING MACHINERY & 3.5 FUEL COST BALANCE OF PLANT For purposes of estimating delivered $/kW energy cost, a solar availability factor Compressors, Turbines, Clutches, of 62.5% is used. The remaining 37.5% of etc. 160.00 thermal energy would be supplied by oil or Generators, Transformers, natural gas. On an annual basis, the Switchyard 16.14 average fuel heat rate would be 0.375 x Balance of Plant 13.25 2290 Btu/kWh or 860 Btu/kWh. With an estimated 1980 fuel cost of $3.55/MBtu, the Rotating Machinery Subtotal $189.39 incremental fuel cost is 3.0 mills/kWh. 3.1.4 TOTAL SYSTEM COST 9/K.W 3.6 OPERATING AND MAINTENANCE COST MINING & CONSTRUCTION 91.00 Operating and maintenance cost has SOLAR THERMAL 213.36 been set at 3 mills/kWh which is based ROTATING MACHINERY & BALANCE on operating experience with conventional OF PLANT 189.39 gas turbines. CONTINGENCY @ 15% 77.14 INTEREST DURING CONSTRUCTION 88.71 3.7 ECONOMIC SCREENING CURVES TOTAL COST - 1980 $659.60 In order to decide the relative value 3.2 AVAILABILITY FACTOR to a utility of any amount of plant ex- pansion, one of the most useful tools for Due to both planned and forced first cut evaluation is the economic outages, a power generation facility is screening curve. Alternative generating not available 100% of the time. However, options may be plotted in terms of cost gas turbine plants have very short outage per year per unit capacity versus genera- periods, i.e., complete overhaul is quick ting hours per year. If a plant were to compared to a nuclear or coal plant. be used for reserve capacity and operated Based on the gas turbine experience, the only several hundred hours a year, the availability factor has been set at 85%. capital costs are more significant than the fuel costs and the choice is clearly 3.3 FIXED ANNUAL CHARGE RATE the simple open cycle gas turbine peaking plant. On the other hand, if the baseload The fixed annual charge rate is based is expanded, i.e., generation time is six on investor return on capital, life of to eight thousand hours per year, the fuel the plant and annual tax rate. For this and operating costs would dominate and the analysis the fixed annual charge rate is choice would be a nuclear or coal plant. 18%. The annual capital cost is: Passive storage systems do not have Annual Capital Cost-Total Capital Cost fuel costs, but do have incremental opera- Availability Factor ting and maintenance costs as well as x Fixed Annual Charge Rate. charging energy costs. Conventional fuel fired CAES has both an incremental 3.4 CHARGING ENERGY COST charging cost and a fuel cost and this is indicated on the screening curves by the In order to charge the storage, off- higher slope these systems have over the peak energy from a baseload facility is passive or adiabatic systems. used. Generally, the capital cost of a baseload plant is charged against energy 4.0 SUMMARY AND CONCLUSION delivered to load and storage is charged only the incremental cos?- of fuel, opera- Because of the high capital costs of tion and maintenance. For a nuclear the solar/CAES option, it does not seem plant (8), the incremental cost is 7 economically competitive with any of the mills/kWh charging energy and for a coal UPHS (Underground Pumped Hydro Storage) plant, 15 mills/kWh. Each kWh of dis- or other CAES systems. charge requires 0.9 kWh of charge. There- fore, the incremental pumping cost for nuclear is 6.3 mills/kWh and 13.5 mills/kWh for coal baseload. 453 The final busbar power cost per year REFERENCES versus generation hours per year is shown in Figures 3 & 4 for nuclear and coal 1. Mattick, W., et. al., "Huntorf - The pumping. Also shown are several other World's First 290 mW Gas Turbine Air storage systems for economic comparison. Storage Peaking Plant." The busbar energy cost may be derived by dividing the ordinate by the abscissa. 2. Koutz, S. L., "Energy Storage System With nuclear pumping the hybrid solar/CAES and Method," U.S. Patent No. 3,677,008, option is marginally competitive with the July 1972. open cycle gas turbine with a generation time of 3000 hours per year. 3. Stephens, T., "Adiabatic Compressed Air Energy Storage Systems," Proceedings POWER GENERATION COST COMPARISON of the Workshop on Compressed Air 12-32 BAR CAES SINGLE OR NO TES Energy Storage Systems. ERDA-76-124. NUCLEAR PUMPING O 7MILLS/kWh 4. Glendenning, I., "Advanced Compressed Air Storage - An Appraisal," Com- / ^~~— pressed Air Energy Storage Technology ^ 200 • A. __-—•/• t-——" i ual IAR SOLA! HVtRID SINGLE TES Symposium. Asilomar Conference R ADrAiATK CA£S SINGLE TCS Grounds, Pacific Grove, California, ft >» K •AR AQUIFER NO TCS 15-17 May 1978. o C&~"' u 4400 ft 2 OIOPT0J9 •AR SOLUTION MINED NO TES 5. Hamilton, N. I., "Packed Beds for i Thermal Energy Storage in an Under- O 100 R AD1ABATIC SINGLE TE5 |AIAN0ONCDMINE) ground Compressed Air Energy Storage I D tO hr WHKNrGMT PUMPING System," AS/ISES, Denver, Colorado, 3,. O 6 hr WEENMGHT PUMPING August 1978. 6. Gintz, J. R., "Closed Cycle, High 2000 4000 4000 Temperature Central Receiver Concept GENEIATION TIME HI/YEAH for Solar Electric Power," EPRI ER-183, Project 377-1, February 1976. Figure 3 7. Gintz, J. R., "Advanced Thermal Energy Storage Concept Definition Study for Solar Brayton Power Plants," Final POWER GENERATION COST COMPARISON 12-32 BAR CAES SINGLE OR NO TES Technical Report, Volume I, ERDA COAL PUMPING 9 IS MILLS /kWhr Contract EY-76-C-03-1300, December 1976. 8. Giramonti, A.J., "Preliminary Feasi- 32 Ml SOLAt H NOONEOMmE) • 10 kr WEIKNKSHI PUKWNG O t tr WEEINIOHT PUMPMC 2000 400D «ooo GENEIATFON TIME Hlt/tEA* Figure 4 454 APPENDICES 455 THE APPLICATION OF FLYWHEEL ENERGY STORAGE TECHNOLOGY TO SOLAR PHOTOVOLTAIC POWER SYSTEMS Alan Millner Massachusetts Institute of Technology Lincoln Laboratory Lexington, MA 02173 ABSTRACT INTRODUCTION Solar photovoltaic (PV) electric load in order to maximize the electric power systems presently being developed power extracted from the array, and invariably use electric storage batteries this function (commonly referred to as when on-site energy storage is required. maximum power tracking) can also be Moreover, studies of future PV power provided by the flywheel. Figure 1 systems assume continued use of batteries shows the system block diagram compar- for on-site storage, albeit with more ison of a battery system and inverter, advanced, efficient and less expensive a conventional flywheel with DC input battery designs. This preeminence is due and output followed by a DC-to-AC at least in part to the generally held inverter and a combined flywheel conviction that no other on-site storage storage and power conditioning system. system can compete economically with The simplicity of the last block batteries for PV usage. However,studies diagram reflects the real cost savings performed during the past year at MIT/ possible with this implementation. Lincoln Laboratory show that flywheel These simplifications hinge on the use energy storage will be technically and of an efficient, low-drag motor- economically competitive with either generator, such as one recently present-day or advanced storage batteries designed by MIT/LL for a spacecraft if the flywheel storage system is pro- application. The system also utilizes perly configured. This conclusion was low-drag high-speed bearings, such as reached after comparing battery and a magnetic bearing recently designed, flywheel storage in a systems context, built and tested for the same program. whereby their influence on other sub- systems (such as inverters) was deter- The costs for a flywheel system mined. with the capabilities enumerated above were estimated for a PV residential ap- The essence of the proposed approach plication and were compared with costs is the utilization of the flywheel sub- for a more conventional system contain- system for more than the energy storage ing batteries, inverter and maximum function. A PV power system usually power tracker. Two different scenarios requires an inverter to convert the were considered: 1) present-day bat- low-voltage DC output from the solar teries and costs and present-day tech- arrays to a (usually) higher voltage nologies and costs for flywheels with AC waveform, and this operation can production quantities assumed in both be performed by the flywheel unit by cases, and 2) technologies and costs use of a DC drive motor and a permanent extrapolated to the 1986 time frame magnet alternator. Also, it is usually for both the flywheel and battery sys- necessary to provide a good impedance tems. In both scenarios it was found match between the PV array and the that because of savings resulting 457 — I BATTERY SYSTEM 2. FLYWHEEL STORAGE PLUS INVERTER I FLYWH=SL I iRRAv] » DC M'G ElECTRONICS »-|lNVEBTER |—»•! 10AD 1 3. FLYWHEEL STORAGE AND CONDITIONING HYWHEEL ARRAY M MOTOK-GENERATO8 ElECIRONICS 1 » 1OAO Figure 1. Solar PV System Comparison from multiple usage of the flywheel usually requires a solar cell array, an components, such a system would be energy storage device, a maximum power economically competitive with the more tracker, and a DC-AC conversion device. conventional approach utilizing storage This proposal is concerned with the batteries, a separate inverter and a development of a flywheel system which maximum power tracker. The cost com- would perform all of these functions. parisons are given in Fig. 5. Such a device would have a DC By combining these elements with motor, an energy storage flywheel and a power conditioning design developed a 60-Hz AC generator. It would be at MIT/LL and an improved rotor procured supported on magnetic bearings in a from one of the contractors in DOE's vacuum housing which would do double storage program, a subscale (approxi- duty as safety containment for the mately 1:10) model of a residential flywheel rotor. The motor-generator solar photovoltaic flywheel energy would in fact be the same device with storage unit will be assembled. In separate input and output electronics. addition, scaling laws will be derived The electronics are shown schematically for storage coupled to solar PV in Figures 2 and 3 and are very simple systems in the power range 10 to 100 and inexpensive due to the use of the kW peak, and in the storage range 25 permanent magnet motor-generator kWh to 5 Mwh. A detailed design of concept. a full-sized 10 kW/25 kWh single- residence storage unit will be made All solar/electric power would go and analyzed for cost and worth to through the DC motor to spin up the ro- the user. Similar but less detailed tor. The motor would be a DC brushless, analyses will be performed for a ironless armature type design which 100 kW/5M¥h load center storage unit. would be controlled as a maximum power tracker for the solar array. This is SYSTEM DESCRIPTION important because the varying electri- cal output of a solar array is gener- A solar photovoltaic installation ally mismatched to the characteristics 458 C+ A- B- STABTING fc COMMUTATOI C0NVE6IEB Pig. 2: Flywheel Power Conditioning Input Schematic Fig. 3: Flywheel Power Conditioning Output Schematic of the storage system and load, reliability. Also, the higher speeds causing inefficient operation. The possible with such an assembly flywheel rotor could be an advanced (perhaps 20 thousand rpm) will allow design of one (or more) of the types smaller rotor size and enhance the presently being tested for DOE by a quality of available AC power. Mech- number of organizations. 'The elec- anical touchdown bearings would be trical output would be accomplished for cold start/stop conditions. No with a permanent-magnet brushless rotating seals would be required. generator and a silicon controlled rectifier (SCR) cycloconverter. The vacuum chamber ensures a long The generator would be sized to energy storage time (days or weeks) be- accept the large surge demands of fore aerodynamic losses become a prob- many candidate loads. lem. With bare-filament rotors, the vacuum system will provide sufficient The entire rotating unit would safety confinement. If other designs be supported on a DC magnetic bearing are enployed, the unit can be placed or hybrid magnetic/ball bearing. underground to provide safety confine- This can be powered from windings ment. The system would be run over a on the motor, allowing fail-safe, 2:1 speed range with the output held at spin-down operation. The resulting 60-Hz, generated independently or syn- rotating unit would have no brushes chronized to an external line. This co- or physical contact with the rotor, responds to a 75 percent depth of dis- allowing very long life and high charge for the energy storage function. 459 An attractive practical appli- with utility power drawn only when the cation for this device is to utilize energy reverse in the flywheel is low. it with a PV-powered single-family It can be scaled up to larger sizes as- residence operating in either a stand- sociated with multiple dwelling units alone mode or coupled to a utility, or commercial PV applications. TASX 1 Build one 1/10 scale model lyitetn with a 1 kwh purchased rotor optimized for PV storage. Refine scaling laws. TASK 2 Design a residence-sized system and a 100-kH load center system and analyze for cost and worth to the user. TASK 3 Build one full-scale, residence-sized system and interface it to a solar array and residential load. Fig. 4: PV Flywheel Storage Component Project Battery System Flywheel System law High Low High Storage $/kWh 90 ISO 75 145 DC input SAW 50 150 85 125 AC output SAW 50 300 70 150 Enclosure $/kWh 32 50 24 32 Residence Total $3,852 $8,400 $3,705 $6,875 (Based on 25 kWh + 6 kHDC + 10 kWAC) Fig. 5: Comparison of Prices FLYWHEEL BASED SYSTEM BATTERY BASED SYSTEM DC Motor Electronics 95* Max Power Tracker 96% DC Motor 96% Batteries 80% AC Generator 95% Inverters 85% Drag Losses 95% Gen. Electronics Losses 92% 73.3% TOTAL 65.3% Fig. 6: Comparison of Efficiencies of Flywheel Energy Storage and Conversion System versus That of a Battery Inverter and Max Power Tracker System 460 Motor/generator Rotor Bonom bearing MECHANICAL AND MAGNETIC ENERGY STORAGE TECHNOLOGY MEETING The Mimslyn Motor Inn Luray, Virginia October 24-26, 1978 Tuesday, October 24 8:00-9:00 a.m. Registration 9:00-9:30 a.m. Introduction Session I: FLYWHEELS 9:00 a.m.-12:20 p.m. Chairman - Thomas M. Barlow, Lawrence Livermore Laboratory 9:30 a.m. Electric and Hybrid vehicle Applications (EHV-MEST) T. M. Barlow, Lawrence Livermore Laboratory 9:40 a.m. The t of Mechanical-Energy-Storage-Devicage- e Addition on the Performance of Electrictr! c vehicles Robert F. McAlevy, III, Robert F. McAlevy & Assoc. 9:50 a.m. Advanced Flywheel Energy Storage Unit for a High Power Energy Source for Vehicular Use Art E. Raynard, Garrett-AiResearch 10:20 a.m. Regenerative Flywheel Energy Storagrage System Edward L. Lustenader,, General Electriectric Company 10:40 a.m. BREAK 11:00 a.m. low Cost Flywheel Demonstration D. W. rst, Johns Hopkins University 11:20 a.m. Materials Program for Fiber Composite Flywheels James A. Rinde, Lawrence Livernore Laboratory 463 Session II: FLYWHEELS 1:45 p.m. - 3:00 p.m. Chairman - Robert 0. Woods, Sandia Laboratories 1:45 p.m. Overview of Component Development Robert O. woods, sandia Laboratories 2:00 p.m. Sandia Composite-Run Flywheel Development £. David Reedy, sandia Laboratories 2:15 p.m. Structural Modeling of a Thick-Rim Rotor A. Keith Miller, sandia Laboratories 2:45 p.m. Aerodynamic Heating of High-Speed Flywheels in Low-Density Environments Mel Baer, sandia Laboratories Afternoon Free 5:30 p.m. Reception 6:45 p.m. Dinner Biergy—rA Bark Service Perspective R. William Hottmeyerr Shenandoan National Park Session III: FLYWHEELS 8:15 p.m. - 10:00 p.m. Chairman - Robert O. Woods, Sandia Laboratories 8:15 p.m. The Application of Fluid Film Bearings and a Passive Magnetic suspension to Energy storage Flywheels M. w. Eusepi, Mechanical Technology, Inc. 8:30 p.m. Low-Loss Ball Bearings for Flywheel Applications David B. Eisenhaure, Ttie cnarles stark Draper Laboratories 8:45 p.m. Seal Studies for Advanced Flywheel Systems I. Anwar, The Franklin institute 9:00 p.m. A Composite Flywheel for vehicle Use Francis c. Younger, William H. urooecK & Assoc. 9:15 p.m. Progress in Composite Flywheel Development P. ward Hill, Hercules, Inc. 9:30 p.m. Advance Composite Flywheel for Vehicle Application Donald E. Davis, Rockwell international 9:45 p.m. High-Energy-Density Flywheel D. L. satchweii, Garrett-AiResearch 10:00 p.m. Adjourn 464 Wednesday, October 25 8:00 a.m. Registration Session IV: SOLAR MECHANICAL 9:00 a.m. - 11:45 a.m. Chairman - Henry H. Dodd, Sandia Laboratories 9:00 a.m. Solar Mechanical Energy Storage Project B. c. Caskey, Sandia Laboratories 9:30 a.m. The Band Type Variable Inertia Flywheel and Fixed Ratio Power Recirculation Applied to it David G. Ullman, Union College 10:00 a.m. Cellulosic Flywheels Arthur G. Erdman, Univ. of Minnesota 10:15 a.m. A Concept for Suppression of Nonsynchronous Whirl "in" Flexible Flywheels John M. Vance, Texas ASM University 10:30 a.m. Break Flywheel Energy Storage Systems 10:45 a.m. Conceptual Design of a Flywheel Energy Storage System Francis C. Younger, William M. Brobeck & Assoc. 11:15 a.m. Residential Flywheel with Turbine Supply Theodore W. Place, Garrett-AiResearch 12:00 p.m. Lunch Session V: SUPERCONDUCTING MAGNETIC ENERGY STORAGE 1:15 p.m. - 3:15 p.m. Chairman - John D. Rogers, Jr., Los Alamos Scientific Laboratory 1:15 p.m. Superconductive Diurnal Energy Storage Studies R. W. Boom, Univ. ot Wisconsin 1:35 p.m. Recent Component Development Studies for Super- conductive Magnetic Energy storage S. van Sciver, Univ. of Wisconsin 1:55 p.m. Power System Stability Using Superconducting """Magnetic Energy Storage Dynamic Characteristics of the BPA System Lee Cresap, Bonneville Power Corporation 2:10 p.m. Hybrid Computer Study of a SMES Unit for Damping Power System Oscillations Paul Krause, Purdue University 465 Superconducting Magnetic Energy Storage (Cont'd) 2:25 p.m. Superconducting M&Ljnetic £fcergy Storage John D. Rogers, Jr., Los Alamos Scientific Laboratory 3:00 p.m. Superconducting Magnetic Energy Storage for Power System Stability ApplicaEIons Carl Chowaniec, Westinghouse Electric Corp. 7:00 p.m. Dinner Session VI: UNDERGROUND PUMPED HYDROELECTRIC STORAGE 8:30 p.m. - 10:00 p.m. Chairman - George T. Kartsounes, Argonne National Laboratory 8:30 p.m. Underground Pumped Hydro Storage: An Overview Shiu-Wing Tam, Argonne National Laboratory George T. Kartsounes, Argonne National Laboratory 9:10 p.m. Evaluation of One and Two Stage High Head Pump/Turbine for underground Power Stations John Degnan, Allis-Chalmers Corp. 9:40 p.m. Multistage Turbine-Pump with Controlled Flow Rate Alexander Goknman, Univ. of Miami 10:00 p.m. Adjourn Thursday, October 26 Session VII: COMPRESSED AIR ENERGY. STORAGE 9:00 a.m. - 11:40 a.m. Chairman - Walter v. Loscutoff, Battelle Pacific Northwest Laboratories 9:00 a.m. CAES Program Overview Walter V. Loscutoff, Battelle Pacific Northwest Laboratories 9:15 a.m. Fluid Flow and Thermal Analysis for CAES in.Porous Rock Reservoirs L. E. wiles, Battelle Pacific Northwest Laboratories 9:35 a.m. Thermo Mechanical Stress Analysis of Porous Rock Reservoirs J. R. Friley, Battelle Pacific Northwest Laboratories 9:55 a.m. Potential Air/Water/Rock Interactions in a Porous Media CAES Reservoir J. A. stottlemyre, Battelle Pacific Northwest Laboratories 10:15 a.m. Break 466 Compressed Air Energy Storage (Cont'd.) 10:30 a.m. Preliminary Design and Stability Criteria for CflES Hard Rock Caverns P. F. Gnirk, RE/Spec inc. 10:50 a.m. Preliminary Long-Terro Stability Criteria for CAES Caverns in Salt Danes R. L. Thorns, Louisiana State University 11:10 a.m. Fabric Analysis of Rock Subjected to Cycling with Heated, Compressed Air H• J. Pincus, Univ. of Wisconsin 11:30 a.m. Numerical Modeling of Behavior of Caverns in Salt for CAES S. Serata, Serata Geomechanics Inc. 11:50 a.m. The Design Optimization of Aquifer Reservoir-Based CAES Friclerick W. Ahrens, Argonne National Laboratory 12:10 p.m. Lunch Session VIII: COMPRESSED AIR ENERGY STORAGE 1:00 p.m. - 2:50 p.m. Chairman - Walter V. Loscutoff, Battelle Pacific Northwest Laboratories 1:00 p.m. Advanced CAES Systems Studies Walter v. Loscutoff, Battelle Pacific Northwest Laboratories M. A. McKinnon, Battelle Pacific Northwest Laboratories 1:20 p.m. Solar Thermal Augmentation of CAES Gerrard T. Fiynn, MIT 1:40 p.m. Evaluation of TurboMachinery for Compressed Air Energy storage Plants George T. Kartsounes, Argonne National Laboratory Choong S. Kim, Argonne National Laboratory 2:00 p.m. Evaluation of the Use of Reciprocating Engines in Compressed Air Energy Storage Plants George T. Kartsounes, Argonne National Laboratory James G. Daly, Argonne National Laboratory 2:40 p.m. Concluding Remarks Thomas M. Barlow 2:50 p.m. Adjournment 467 FIRST ANNUAL MECHANICAL AND MAGNETIC ENERGY STORAGE TECHNOLOGY MEETING The Mimslyn Inn Luray, Virginia October 24-26, 1978 Attendees List ACKERMAN, Sam L. BARLOW, Thomas M. Director, Magnet Systems Program Manager General Dynamics Lawrence Liverrnore Laboratory P.O. Box 80847 7000 East Avenue San Diego, CA 92138 Livermore, CA 94550 ADOLFSON, William F. BEACH, Raymond F. Senior Scientist Project Manager Booz, Allen & Hamilton, Inc. NASA-Lewis Research Center 4330 East-West Highway 21000 Brookpark Road Bethesda, MD 20014 Cleveland, OH 44135 AHRENS, F. BEACHLEY, Norman Mechanical Engineer Professor Argonne National Laboratory University of Wisconsin 9700 South Cass Avenue 1513 University Avenue Argonne,IL 60439 Madison, WI 53706 ALLEN,Bob Beck, Curt Battelle Pacific Northwest Battelle Pacific Northwest Laboratories Laboratories c/o G.C. Chang P.O. Box 999 U.S. Department of Energy Richmond, WA 99352 600 E Street, NW, Room 416 Washington, D.C. 20545 BERMAN, Irwin Principal Engineer ANWAR, I. Commonwealth Edison Company Senior Staff Engineer P.O. Box 767 Franklin Institute 20th and Parkway Chicago, IL 60690 Philadelphia, PA 19119 BERVIG, D.R. BAER, M.R. Project Engineer Member, Technical Staff Black & Veatch Sandia Laboratories P.O. Box 8405 Kirtland Air Force Base Kansas City, MO 64114 Albuquerque, NM 87185 BLAY, Dominique BAKER, Merl Engineer Coordinator of Energy Conservation French Atomic Energy Commission Oak Ridge National Laboratory c/o French Embassy BoxX 1730 Rhode Island Avenue, NW Oak Ridge, TN 37830 Room 1217 Washington, D.C. 20036 469 BLOMQU1ST, Carl CAMPBELL, James Chemical Engineer Program Manager, Energy Argonne National Laboratory Propulsion Technician 9700 South Cass A'. rnue U.S. Department of Transportation Argonne, IL 60439 2100 Second Street, t ' Washington, D.C. 20590 BLOOM, Harold L. Project Engineer CASKEY, Bill C. Energy Systems Program Department Staff Member General Electric Company Sandia Laboratories Building 36, Room 421 Division 5716 Scheneetady, NY 12345 P.O. Box 580 Albuquerque, NM 87185 BOG ART, Locke Planner, Technological Transfer CHANG, George C. U.S. Department Of Energy Chief, APM Branch 600 E Street, NW Division of Energy Storage Systems Washington, D.C. 20545 Office of Energy Technology U.S. Department of Energy BOOM, Roger 600 E Street, NW Professor of Nuclear and Metallic Washington, D.C. 20545 Engineering Engineering Experiment Station CHARLWOOD, Robin University of Wisconsin - Madison Acres American, Inc. 1500 Johnson Drive 900 Liberty Bank Building Madison, WI 53706 Buffalo, NY 14202 BORTZ, Susan E. CHOWANIEC, Carl Consultant Westinghouse Corporation Bradford National Corporation 700 Braddock Avenue 1901 L Street, NVV #301 Pittsburgh, PA 15112 Washington, D.C. 20036 CREAGAN, Robert J. BRAASCH, Richard H. Director, Technical Assessment Division Supervisor Westinghouse Corporation Sandia Laboratories Gateway Center Division 4715 Pittsburgh, PA 15222 Albuquerque, NM 87185 CRESAP, Richard Lee BRANDVOLD, Glen Analysis Engineer Department Manager Bonneville Power Administration Solar Energy Projects 1002 Northeast Holladay Sandia Laboratories Portland, OR 97208 Division 4716 Albuquerque, NM 87185 CROTHERS, William Lawrence Livermore Laboratory BROBECK, William M. 7000 East Avenue Cha'/man of the Board Livermore, CA 94550 William M. Brobeck & Associates 1235 Tenth Street CUNDY, Thomas Berkeley, CA 94710 Senior Mathematics Analyst Serata Geomechanics CALLAHAN, Gary D. 1229 8th Street RE/SPEC, Inc. Berkeley, CA 94709 P.O. Box 725 Rapid City, SD 57709 470 BLOMQUIST, Carl CAMPBELL, James Chemical Engineer Program Manager, Energy Argonne National Laboratory Propulsion Technician 9700 South Cass Avenue U.S. Department of Transportation Argonne, IL 60439 2100 Second Street, SW Washington, D.C. 20590 BLOOM, Harold L. Project Engineer CASKEY, Bill C. Energy Systems Program Department Staff Member General Electric Company Sandia Laboratories Building 36, Room 421 Division 5716 Scheneetady, NY 12345 P.O. Box 580 Albuquerque, NM 87185 BOG ART, Locke Planner, Technological Transfer CHANG, George C. U.S. Department Of Energy Chief, APM Branch 600 E Street, NW Division of Energy Storage Systems Washington, D.C. 20545 Office of Energy Technology U.S. Department of Energy BOOM, Roger 600 E Street, NW Professor of Nuclear and Metallic Washington, D.C. 20545 Engineering Engineering Experiment Station CHARLWOOD, Robin University of Wisconsin - Madison Acres American, Inc. 1500 Johnson Drive 900 Liberty Bank Building Madison, WI 53706 Buffalo, NY 14202 BORTZ, Susan E. CHOWANIEC, Carl Consultant Westinghouse Corporation Bradford National Corporation 700 Braddock Avenue 1901 L Street, NW #301 Pittsburgh, PA 15112 Washington, D.C. 20036 CREAGAN, Robert J. BRAASCH, Richard H. Director, Technical Assessment Division Supervisor Westinghouse Corporation Sandia Laboratories Gateway Center Division 4715 Pittsburgh, PA 15222 Albuquerque, NM 87185 CRESAP, Richard Lee BRANDVOLD, Glen Analysis Engineer Department Manager Bonneville Power Administration Solar Energy Projects 1002 Northeast Holladay Sandia Laboratories Portland, OR 97208 Division 4716 Albuquerque, NM 87185 CROTHERS, William Lawrence Livermore Laboratory BROBECK, William M. 7000 East Avenue Chairman of the Board Livermore, CA 94550 William M. Brobeck & Associates 1235 Tenth Street CUNDY, Thomas Berkeley, CA 94710 Senior Mathematics Analyst Serata Geomechanics CALLAHAN, Gary D. 1229 8th Street RE/SPEC, Inc. Berkeley, CA 94709 P.O. Box 725 Rapid City, SD 57709 470 DAVIS, Donald E. EVANS, Harold E. Program Manager Branch Head Rocketdyne Division NASA/GSFC Rockwell International Greenbelt, MD 20811 6633 Canoga Avenue Canoga Park, CA 91304 FARQUHAR, Oswald Professor DEGNAN, John University of Massachusetts Manager, Mechanical Developments Department of Geology Allis Chalmers Hydro Turbine Amherst, MA 01003 Division Box 712 FAUCONNIER, Jean Claude York, PA X7405 Engineer French Atomic Energy DERBY, Roger Commission Program Manager c/o French Embassy Division of Energy Storage Systems 1730 Rhode Island Avenue, NW Office of Energy Technology Room 1217 U.S. Department of Energy Washington, D.C. 20036 600 E Street, NW Washington, D.C. 20545 FLYNN, Gerald T. Staff Engineer DE VINEY, Glen Massachusetts Institute of Section Engineer Technology Commonwealth Edison Company Lincoln Laboratory P.O. Box 767 Box 73 Chicago, IL 60690 Lexington, MA 02173 DODD, Henry M. FOSSUM, Arlo F. Division Supervisor RE/SPEC, Inc. Sandia Laboratories P.O. Box 725 Division 5743 Rapid, City, SD 57709 P.O. Box 5800 Albuquerque, NM 87185 FRANK, Andrew Professor EISENHAURE, David University of Wisconsin Charles Stark Draper Laboratory Department of Electrical and Mail Station 37 Computer Engineering 555 Technology Square Madison, WI 53706 Cambridge, MA 02139 FRIGO, Art EMIGH, C. Robert Mechanical Engineer Associate Division Leader Argonne National Laboratory Energy Technology 9700 South Cass Avenue Las Alamos Scientific Laboratory Argonne, IL 60439 P.O. Box 1663 Los Alamos, NM 87545 FRILEY, John R. Research Engineer ERDMAN, Arthur G. Battelle Pacific Northwest Associate Professor Laboratories University of Minnesota P.O. Box 999 Mechanical Engineering Department Richland, WA 99352 Minneapolis, MN 55455 GAHIMER, John EUSEPI, Martin W. Program Manager Program Engineer Division of Energy Storage Systems Mechanical Technology, Inc. Office of Energy Technology 968 Albany-Shaker Road U.S. Department of Energy Latham, NY 12110 600 E Street, NW Washington, D.C. 20585 471 GIRAMONTI, A.J. HERBERMANN, Richard Senior Engineer Chief Engineer, Fusion Division United Technologies Grumman Aerospace Corporation Research Center Mail Stop B-27-35 Silver Lane Bethpage, NY 11714 East Hartford, CT 06108 HILL, P. Ward GLASER, Dick Superintendent, Advanced Kelsey Hayes Technology 2500 Green Road Hercules Inc. Ann Arbor, MI 48105 P.O. Box 210 Cumberland, MD 21502 GNIRK, Paul F. RE/SPEC, Inc. HOLLIDAY, Robert P.O. Box 725 Program Manager Rapid City, SD 57709 Division of Energy Storage Systems Office of Energy Technology GOKHMAN, Alexander U.S. Department of Energy Associate Professor 600 E Street, NW Department of Mechanical Engineering Washington, D.C. 20545 University of Miami Coral Gables, FL 33124 HOPPIE, Lyle O. EATON-ERC GRASSBERGER, Robert Box 766 Program Manager Southfield, MI 48037 BDM Corporation 2600 Yale, SE HURLEY, James D. Albuquerque, NM 87106 Project Development Manager General Electric HAGEN, David Building 5, Room 425 Research Associate Sehenectady, NY 12345 University of Minnesota 111 Church Street, SE JOKL, A.L. Minneapolis, MN 55455 Supervisor, Physical Scientist MERADCOM HAMPSON, Chris Attn: DRDME-EA Senior Research Scientist Fort Belvoir, VA 22060 International Research and Technology Corporation KAPNER, Mark 7655 Old Springhouse Road Engineer McLean, VA 22102 Hittman Associates 9190 Redbranch Road HAYES, Edward Columbia, MD 21045 Vice President Kelsey Hayes KARTSOUNES, George 38481 Hurron River Drive Mechanical Engineer Romulus, MI 48174 Argonne National Laboratory 9700 South Cass Avenue HECK, Francis M. Argonne, IL 60439 Manager, Systems Engineering Westinghouse Electric Corporation KATZ, Donald P.O. Box 10864 University of Michigan Pittsburgh, PA 15236 2011 Washington Avenue Ann Arbor, MI 48104 472 KOSTAL, Kenneth MARSHALL, H.K. Senior Project Engineer Associate President Sargent and Lundy Kinergy Research & Development 55 East Monroe P.O. Box 1128 Chicago, IL 60525 Wake Forest, NC 27587 KRAUSE, Paul MILLER, A. Keith Electrical Engineering Member, Technical Staff Department Sandia Laboratories Purdue University Division 5521 West Lafayette, IN 47907 Albuquerque, NM 87185 KULKARN1, S.V. MILNER, Alan Lawrence Livermore Laboratory Staff Member P.O. Box 808, L-338 Massachusetts Institute Livermore, CA 94550 of Technology Lincoln Laboratory LARRECQ, A.J. P.O. Box 73 President Lexington, MA 02173 Power Generators, Inc. 94 Stokes Avenue MC ALEVY, Roberty F. Ill Trenton, NJ 08638 Senior Associate Robert F. McAlevy III & LEMMENS, Joseph R. Associates Kinergy Research & Development 1204 Bloomfield Street P.O. Box 1128 Hoboken, NJ 07030 Wake Forest, NC 27587 MC COY, H. E. LOSCUTOFF, W.V. Oak Ridge National Laboratory Project Manager 132 Balboa Circle Battelle Pacific Northwest Oak Ridge, TN 37830 Laboratories P.O. Box 999 MC DONALD, Alan T. Richland, WA 99352 Professor School of Mechanical Engineering LUSTENADER, E.L. Purdue University Manager West Lafayette, IN 47907 General Electric Company 1 River Road MC SPADDEN, William R. Sehenectady, NY 12345 Senior Research Engineer Battelle Pacific Northwest MAILLOT, Bernard Laboratory Engineer P.O. Box 999 French Atomic Energy Richland, WA 99352 Commission c/o French Embassy NASH-WEBBER, James L. 1730 Rhode Island Avenue, NW Facilities Project Manager Room 1217 Massachusetts Institute of Technology Washington, D.C. 20036 Energy Laboratory P.O. Box 69 MANAKER, Arnold M. Cambridge, MA 02139 Mechanical Engineer Tennessee Valley Authority NEWHOUSE, Norman L. 1360 Commerce Union Bank Design Engineer Building Brunswick Corporation Chattanooga, TN 37401 4300 Industrial Avenue Lincoln, NE 68504 473 NICOL, James PINCUS, Howard J. Vice President Professor A.D. Little Department of Geological Acorn Park Sciences Cambridge, MA 02140 University of Wisconsin - Milwaukee NORMAN, Thomas A. Milwaukee, WI 53201 Program Engineer U.S. Postal Service PLACE, Theodore W. 11711 Parklawn Drive Project Engineer Rockville, MD 20852 Garrett AiResearch Corporation 2525 West 190th Street OMOHUNDRO, Laura L. Torrance, CA 90505 Vice President Kinergy Research & Development POST, Dave P.O. Box 1128 Union Carbide Corporation Wake Forest, NC 27587 Y-12 Plant Oak Ridge, TN 37830 PARADIS, L.R., Jr. Principal Engineer RABENHORST, David W. Raytheon Company Flywheels Program Manager Hartwell Road Applied Physics Laboratory Bedford, MA 01730 John Hopkins University Johns Hopkins Road PARSONS, Lawrence F. Laurel, MD 20810 Civil Engineer Department of the Interior RAYNARD, Arthur E. Bureau of Reclamation Senior Project Engineer Washington, D.C. 20240 Garrett AiResearch Corporation 2525 West 190th Street PATRICK, A.J. Torrance, CA 90505 Manager, Energy Development Avco Systems Division REDDICK, William C. 201 Lowell Street Scientific Advisor Wilmington, MA 01887 U.S. Department Of Energy 20 Massachusetts Avenue, NW PAX, Charles Washington, D.C. 20545 U.S. Department of Energy 20 Massachusetts Avenue REEDY, E. David, Jr. Washington, D.C. 20545 Member, Technical Staff Sandia Laboratories PERRAM, Robert Division 5844 Program Manager KAFB East Aerospace Corporation Albuquerque, NM 87185 20030 Century Boulevard Germantown, MD 20767 RINDE, J.A. Chemist PEZDIRTZ, George F. Lawrence Livermore Laboratory Director, Division of P.O. Box 808 L-338 Energy Storage Systems Livermore, CA 94550 Office of Energy Technology U.S. Department of Energy ROGERS, John D. 600 E Street, NW Los Alamos Scientific Laboratory Washington, D.C. 20545 Org. CTR-9, Mail 464 Los Alamos, NM 87545 474 NICOL, James PINCUS, Howard J. Vice President Professor A.D. Little Department of Geological Acorn Park Sciences Cambridge, MA 02140 University of Wisconsin - Milwaukee NORMAN, Thomas A. Milwaukee, WI 53201 Program Engineer U.S. Postal Service PLACE, Theodore W. 11711 Parklawn Drive Project Engineer Rockville, MD 20852 Garrett AiResearch Corporation 2525 West 190th Street OMOHUNDRO, Laura L. Torrance, CA 90505 Vice President Kinergy Research & Development POST, Dave P.O. Box 1128 Union Carbide Corporation Wake Forest, NC 27587 Y-12 Plant Oak Ridge, TN 37830 PARADIS, L.R., Jr. Principal Engineer RABENHORST, David W. Raytheon Company Flywheels Program Manager Hartwell Road Apr lied Physics Laboratory Bedford, MA 01730 John Hopkins University Johns Hopkins Road PARSONS, Lawrence F. Laurel, MD 20810 Civil Engineer Department of the Interior RAYNARD, Arthur E. Bureau of Reclamation Senior Project Engineer Washington, D.C. 20240 Garrett AiResearch Corporation 2525 West 190th Street PATRICK, A.J. Torrance, CA 90505 Manager, Energy Development Avco Systems Division REDDICK, William C. 201 Lowell Street Scientific Advisor Wilmington, MA 01887 U.S. Department Of Energy 20 Massachusetts Avenue, NW PAX, Charles Washington, D.C. 20545 U.S. Department of Energy 20 Massachusetts Avenue REEDY, E. David, Jr. Washington, D.C. 20545 Member, Technical Staff Sandi£ Laboratories PERRAM, Robert Division 5844 Program Manager KAFB East Aerospace Corporation Albuquerque, NM 87185 20030 Century Boulevard Germantown, MD 20767 RINDE, J.A. Chemist PEZDIRTZ, George F. Lawrence Livermore Laboratory Director, Division of P.O. Box 808 L-338 Energy Storage Systems Livermore, CA 94550 Office of Energy Technology U.S. Department of Energy ROGERS, John D. 600 E Street, NW Los Alamos Scientific Laboratory Washington, D.C. 20545 Org. CTR-9, Mail 464 Los Alamos, NM 87545 474 ROSE, Shelley SMITH, B.B. Secretary Union Carbide Corporation Lawrence Livermore Laboratory Y-12 Plant 7000 East Avenue Oak Ridge, TN 37830 Livermore, CA 94550 SPISAK, A.J. ROSNER, Carl Manager, Advanced Program Inter-Magnetics General Westinghouse Corporation Corporation 700 Braddock Avenue P.O. Box 566 Pittsburgh, PA 15112 Guilderland, NV 12084 STILLER, Paul SAPOWITH, Alan D. Research Project Engineer Section Chief Westinghouse Corporation* Avco Systems Division 700 Braddock Avenue 201 Lowell Street Pittsburgh, PA 15112 Wilmington, MA 01887 STONE, Richard G. SATCHWELL, David L. Associate Department Head Senior Development Engineer Lawrence Livermore Laboratory Garrett AiResearch Corporation P.O. Box 808-L-123 2525 West 190th Street Livermore, CA 94550 Torrance, CA 90505 STOTTLEMYRE, Jim SCHILDKNECHT, Harold Research Scientist Member, Technical Staff Battelle Pacific Northwest Sandia Laboratories Laboratories Org. 2324 P.O. Box 999 KAFB East Richland, WA 99352 Albuquerque, NM 87185 SULLIVAN, Cornelius M. SCHLIEBEN, Ernest W. Staff Member President Massachusetts Institute of Von Research Technology 40 West Lafayette Street Lincoln Laboratory Trenton, NJ 08608 Building D-250 P.O. Box 73 SCHMIDT, Franklin R. Lexington, MA 02173 Consultant Bradford National Corporation SURABIAN, Greg #2 Research Place Consultant Rockville, MD 20850 Bradford National Corporation 1901 L Street NW, #301 SCHWARTZ Martin W. Washington, D.C. 20036 Engineer Lawrence Livermore Lrboratory SWISHER, James H. P.O. Box 808 Assistant Director for Livermore, CA 94550 Physical Storage Systems Division of Energy Storage Systems SERATA, Shosei Office of Energy Technology President U.S. Department of Energy Serata Geomechanics 600 E Street, NW 1229 8th Street Washington, D.C. 20545 Berkeley, CA 94709 TAM, S.W. Assistant Metallurgical Engineer Argonne National Laboratory 9700 South Cass Avenue Argonne, IL 60439 475 THOMPSON, Jane W EINSTEIN, Kenneth D. Consultant Bradford National Corporation Booz, Allen & Hamilton, Inc. #2 Research Place Energy & Environment Division Rockville, MD 20850 4330 East-West Highway Bethesda, MD 20014 THOMPSON, Philip Program Manager WEISS, Joel A. Division of Energy Storage Systems Physicist Office of Energy Technology Office of Solar Programs U.S. Department of Energy U.S. Department of Energy 600 E Street, NW 600 E Street, NW Washington, D.C. 20545 Washington, D.C. 20545 THOMS, Robert L. WHITCOMB, Mike Louisiana State University Engineer Institute for Hittman Associates 9190 Redbraneh Road Environmental Studies Columbia, MD 21405 Baton Rouge, LA 70803 ULLMAN, David WHITMAN, Howard Charles Stark Draper Laboratory Assistant Professor 555 Technology Square Department of Mechanical Engineering Cambridge, MA 02139 Union College Schenectady, NY 12308 WILES, Larry Development Engineer UPTON, Joseph W. Senior Research Scientist Battelle Pacific Northwest Battelle Pacific Northwest Laboratories Laboratories P.O. Box 999 Box 999 Richland, WA 99352 Richland, VVA 99352 WILKINSON, John P.D. VAN SCIVER, S.W. Manager, Solid Mechanical Unit Scientist General Electric Company University of Wisconsin P.O. Box 43 1500 Johnson Drive Schenectady, NY 12301 Madison, Wl 53711 WILLETT, David C. Vice President VAN ZANTEN, M. Acres American, Inc. M. sc 900 Liberty Bank Building Netherlands Energy Buffalo, NY 14202 Research Foundation Westerduinweg 3 Petten, Netherlands 1755 LE WILLIAMS, John R. Geotechnical Engineer Dames & Moore VANCE, John M. 4 Militia Drive Associate Professor Lexington, MA 02173 Texas A & M Mechanical Engineering College Station, TX 77843 WISSING, Thomas J. Administrator, Government VERGARA, Rudolfo D. Research and Development Research Scientist Eaton Corporation Battelle Columbus Laboratories P.O. Box 766 505 King Avenue Southfield, MI 48037 Columbus, OH 43201 476 KOSTAL, Kenneth MARSHALL, H.K. Senior Project Engineer Associate President Sargent and Lundy Kinergy Research & Development 55 East Monroe P.O. Box 1128 Chicago, IL 60525 Wake Forest, NC 27587 KRAUSE, Paul MILLER, A. Keith Electrical Engineering Member, Technical Staff Department Sandia Laboratories Purdue University Division 5521 West Lafayette, IN 47907 Albuquerque, NM 87185 KULKARNI, S.V. MILNER, Alan Lawrence Livermore Laboratory Staff Member P.O. Box 808, L-338 Massachusetts Institute Livermore, CA 94550 of Technology Lincoln Laboratory LARRECQ, A.J. P.O. Box 73 President Lexington, MA 02173 Power Generators, Inc. 94 Stokes Avenue MC ALEVY, Roberty F. Ill Trenton, NJ 08638 Senior Associate Robert F. McAlevy III & LEMMENS, Joseph R. Associates Kinergy Research 6c Development 1204 Bloomfield Street P.O. Box 1128 Hoboken, NJ 07030 Wake Forest, NC 27587 MC COY, H. E. LOSCUTOFF, W.V. Oak Ridge National Laboratory Project Manager 132 Balboa Circle Battelle Pacific Northwest Oak Ridge, TN 37830 Laboratories P.O. Box 999 MC DONALD, Alan T. Richland, WA 99352 Professor School of Mechanical Engineering LUSTENADER, E.L. Purdue University Manager West Lafayette, IN 47907 General Electric Company 1 River Road MC SPADDEN, William R. Sehenectady, NY 12345 Senior Research Engineer Battelle Pacific Northwest MAILLOT, Bernard Laboratory Engineer P.O. Box 999 French Atomic Energy Riehland, WA 99352 Commission c/o French Embassy NASH-WEBBER, James L. 1730 Rhode Island Avenue, NW Facilities Project Manager Room 1217 Massachusetts Institute of Technology Washington, D.C. 20036 Energy Laboratory P.O. Box 69 MANAKER, Arnold M. Cambridge, MA 02139 Mechanical Engineer Tennessee Valley Authority NEWHOUSE, Norman L. 1360 Commerce Union Bank Design Engineer Building Brunswick Corporation Chattanooga, TN 37401 4300 Industrial Avenue Lincoln, NE 68504 473 ROSE, Shelley SMITH, B.B. Secretary Union Carbide Corporation Lawrence Livermore Laboratory Y-12 Plant 7000 East Avenue Oak Ridge, TN 37830 Livermore, CA 94550 SPISAK, A.J. ROSNER, Carl Manager, Advanced Program Inter-Magnetics General Westinghouse Corporation Corporation 700 Braddock Avenue P.O. Box 566 Pittsburgh, PA 15112 Guilderland, NY 12084 STILLER, Paul SAPOWITH, Alan D. Research Project Engineer Section Chief Westinghouse Corporation1 Avco Systems Division 700 Braddock Avenue 201 Lowell Street Pittsburgh, PA 15112 Wilmington, MA 01887 STONE, Richard G. SATCHWELL, David L. Associate Department Head Senior Development Engineer Lawrence Livermore Laboratory Garrett AiResearch Corporation P.O. Box 808-L-123 2525 West 190th Street Livermore, CA 94550 Torrance, CA 90505 STOTTLEMYRE, Jim SCHILDKNECHT, Harold Research Scientist Member, Technical Staff Battelle Pacific Northwest Sandia Laboratories Laboratories Org. 2324 P.O. Box 999 KAFB East Riehland, WA 99352 Albuquerque, NM 87185 SULLIVAN, Cornelius M. SCHLIEBEN, Ernest W. Staff Member President Massachusetts Institute of Von Research Technology 40 West Lafayette Street Lincoln Laboratory Trenton, NJ 08608 Building D-250 P.O. Box 73 SCHMIDT, Franklin R. Lexington, MA 02173 Consultant Bradford National Corporation SURABIAN, Greg #2 Research Place Consultant Rockville, MD 20850 Bradford National Corporation 1901 L Street NW, #301 SCHWARTZ Martin W. Washington, D.C. 20036 Engineer Lawrence Livermore Laboratory SWISHER, James H. P.O. Box 808 Assistant Director for Livermore, CA 94550 Physical Storage Systems Division of Energy Storage Systems SERATA, Shosei Office of Energy Technology President U.S. Department of Energy Serata Geomechanics 600 E Street, NW 1229 8th Street Washington, D.C. 20545 Berkeley, CA 94709 TAM, S.W. Assistant Meialurgical Engineer Argonne National Laboratory 9700 South Cass Avenue Argonne, IL 60439 475 WOODS, R.O. Member, Technical Staff Sandia Laboratories Division 4715 Albuquerque, NM 87185 YONK, Alan K. Senior Geologist Sargent and Lundy 55 East Monroe Street Chicago, IL 60603 YOUNGER, Francis C. President William M. Brobeck & Associates 1235 10th Street Berkeley, CA 94710 477 •n. nan nanm era •