Cold Air Distribution in Office Buildings: Technology Assessment for Califonia

UC Berkeley HVAC Systems

Title Cold air distribution in office buildings: technology assessment for califonia

Permalink https://escholarship.org/uc/item/6qx5k7pr

Authors Bauman, Fred Borgers, T. LaBerge, P. et al.

Publication Date 1993

Peer reviewed

eScholarship.org Powered by the California Digital Library University of California 3702 COLD AiR DiSTRiBUTiON iN OFFICE BUiLDiNGS: TECHNOLOGY ASSESSMENT FOR CALiFORNiA

F.S. Bauman, P.E. T. Borgers, Ph.D. P. LaBerge A.J. Gadgil, Ph.D. Member ASHRAE

ABSTRACT adopt TESsystems in their designs, manyutilities have introduced financial incentive programs. Cold air distribu- This paper presents the results of a study to assess the tion (CoAD)technology in cornmercial buildings has arisen current state of practice atwl the energy atwl operating cost because of the building space and cost savings that can be implications of cold air distribution in California and to realized by combiningit with ice storage. By distributing identify the key research needs for the continued develop- lower temperature air (40°F to 50°F [4.4°C to 10°C]) ment of this technology in newcommercial buildings in the throughout the building, a CoADsystem can take greater state. Whole-building energy simulations were made to advantageof the chilled water (typically 34°F to 36 °F [I °C compare the energy performance of a prototypical office to 2°C]) produced by the ice storage system. The colder building in three California climates using conventional and temperature allows primary supply air volumes to be cold air distribution, with and without ice storage, to show reduced, compared to a conventional 55°F (13°C) supply the impactsof load shifting, energyuse, and utility costs for air design, while still satisfying the building’s cooling load. three typical utility rate structures. The merits of economiz- Consequently, fans and ducts can be downsized, reducing ers and fan..powered mixing boxes were also studied when first costs and operating costs and often saving valuable used in conjunction with cold air delivery. A survey was floor area and vertical height. Since the reduction in fan cotgtucted to assess the perceived strengths and limitations energy use also occurs primarily during on-peak hours, this of this technology, perceived barriers to its widespreaduse, further reduces peak electricity demand. attd user experience. The survey was based on interviews A considerable anaount of research has been done on with consulting engineers, equipment manufacturers, the topics of cool storage and cold air distribution. Dorgan researchers, utility representatives, and other users of coM and Elleson (1988) present a comprehensive design guide air distribution technology. Selected fimtings from the on CoADsystems based on current practice and available industry survey are also discussed. research at that time. Interest in obtaining performancedata Cold air distribution (CoAD)is found to always reduce on operational TES/CoADsystems has led to a few field fan energy use in comparisonto conventional 55°F (13°C) studies (Dorgan and Elleson 1987; Merten et al. 1989; air distribution systems whenconditioned air is delivered Dorgan et al. 1990; Landry and Noble 1991). Whole-build- directly to the space O~ofan-powered mixing boxes). Total ing energy simulation studies have investigated energy use buiMing energy use for ice storage/CoAD systems was and operating costs for an ice TES/CoADsystem in always higher than a well-designed conventional system but comparisonto a conventional system for six U.S. climates significantly lower than a commonlyinstalled packaged (Hittle and Bhansali 1990), and to analyze the energy system. Whena favorable utility rate structure was applied, penalties associated with reduced economizer use with a the load-shying benefits of ice storage/CoAD systems CoADsystem (Catanese 1991). The importance of utility produced the lowest annual operating costs of all system- rate structures in motivating the application of thermal plant cot~gurations studied. energy storage for off-peak cooling has been recognized by many members of the building industry and has been INTRODUCTION discussed by Knebel (1990) and MacCracken(1990). recent years, several workshops and seminars have been Space cooling in commercialbuildings represents the held to disseminate information on TES/CoADtechnology largest single category of electricity demandoccurring to a larger audience (EPRI 1987, 1990a, 1991a; ASHRAE during a utility’s summeron-peak hours. Ice storage is one 1990a). Other available publications provide updated form of thermal energy storage (TES), or off-peak air summaries of important issues in TES/CoADtechnology conditioning, that allows most of a building’s cooling (EPRI 1990b, 1991b; Elleson 1991). energy requirements to be shifted to off-peak periods, thus Recent research on cold air distribution has focused on reducing the need for a utility to increase generating the roomair diffusion problemin terms of maintaining both capacity. To encourage building owners and developers to acceptable thermal comfort and indoor air quality. Berglund Fred S. Baumanis a research specialist and Paul LaBergeis a graduate student researcher for the Center for EnvironmentalDesign Research, University of California, Berkeley, CA.To~n Borgers is a professor in the Departmentof Chemistry, HumboldtState University, Arcata, CA.Ashok J. Gadgilis a staff scientist III at the LawrenceBerkeley Laboratory, Berkeley, CA.

ASHRAETransactions: Research 109 (1991) reports that the reduced humiditylevels occurring rapid growth in new office construction: San Jose, Fresno, buildings with cold air distribution can actually provide and San Bernardino. improved perceptions of comfort and air freshness com- In a second part of the study, a survey was conducted pared to those experienced at the same temperature in a by interviewing consulting engineers, equipment manufac- space conditioned with conventional 55 °F air. He quantified turers, researchers, utility representatives, and other users the benefits of this relationship by recommendingthat zone of cold air distribution technology. The information gath- dry-bulb temperature be adjusted upward by 1 °F for each ered through the survey was used to assess the current state 10°F reduction in supply air dewpoint temperature. of practice in California by producing a list of current Traditionally, for CoADcooling applications, fan- California projects involving cold air distribution and a powered mixing boxes (FPMB)have been used to raise the discussion of the factors influencing the future development supply temperature and flow rate and to ensure adequate of cold air distribution. diffuser performance. However, more recent research and The results of the whole-building energy simulations developmenthave focused on supplying cold air directly to are presented and discussed in detail below. Only selected the zones, eliminating the electricity use and capital and findings from the industry survey are included in this paper. maintenance costs of FPMBs.Concern over the perfor- For a full discussion of the survey, refer to Baumanet al. manceof diffusers supplying air directly to the space under (1992). low-temperature, low-flow conditions has prompted a numberof experimental and numerical studies. Gadgil et al. DESCRiPTiON OF PROTOTYPEBUiLDiNG (1991) describe detailed laboratory tests of a commercially available linear diffuser demonstrating acceptable perfor- Huanget al. (1991) have investigated the characteristics mancewith a 30°F (17°C) supply/roo~n temperature differ- of conmiercial office buildings in California and have ence and a supply volume of 0.3 cfm/ft 2 (1.5 L/s.m2). separated them into two categories: old and new. This study Anderson et al. (1991) describe the development of found that recently constructed buildings in mild California innovative experimental technique for visualizing the climates such as Los Angeles frequently use 76 to 80 airflow from cold air diffusers. Kirkpatrick et al. (1991) kBtu/ft2 (870 to 910 MJ/m2) pet" year in fuel and electricity, present a simple analytical model of cold air jet perfor- an improvementover manybuildings of the older stock that mance. Miller (1991) presents a design methodology for use roughly 130 kBtu/ft 2 (1480 MJ/m2). Much of the selecting cold air diffusers based on previously completed description of the building shell, scheduling, and internal laboratory tests and the ASHRAEAir Diffusion Perfor- energyuse in the definition of the prototypical newCalifor- mance Index (ASHRAE1990b). nia commercialbuilding used as input to DOE-2.1Ein this "[’he purposeof this paper is to present the results of a study is based upon their" work. study to assess the current state of practice and the energy "[’he prototypebuilding is defined as havingthree floors and operating cost implications of cold air distribution in of 60,000 ft 2 (5,570 2) each and has s teel a nd spandrel California and to identify the key research needs for the glass R7.5 walls. Thirty percent of the gross wall area is continued adoption of this technology in new commercial double-panedglass, having a normal transmittance of 62 %. buildings in the state. The roof is steel and metal decking under tar and gravel Whole-building energy simulations using the DOE-2.1E and is insulated to R15.8. The interior floors are carpeted computer program were performed to investigate the energy four-inch concrete, and interior walls are steel stud and perfo~xnance and operating costs of a prototypical new gypsum board. California office building using cold air distribution (42°F For simulation purposes, each floor is divided into the [5.6°C]) in comparison to the same building with two usual single core zone and four perimeter zones consistent different conventional55 °F (13 °C) air" distribution systems with the assumed uniformity of building use. Between (packaged system and component-assembled system). floors and below the roof are three foot-high plenums for emphasizethe energy-saving potential of cold air distribu- utilities with bases of lay-in acoustical tiles, producingan tion, with the exception of a separate series of simulations 8.5-ft (2.6-m)floor-to-ceiling height. Then’haltransfer takes investigating fan-powered mixing boxes, all simulations place between core and perimeter zones and between zones assumedthe direct supply of conditioned air (42°F or 55 °F) arid plenums above. to the space without the use of fan-poweredmixing boxes. Scheduleduse of the building is the standard five-day Simulations explored the energy use and operating costs for work week with hours of 8 a.m. to 5 p.m. Limited use is cold air distribution (1) with four different ice storage assumedduring the weekendsand evenings. Full occupancy capacities (one without storage, one with partial ice storage, is assumed on work days with one occupant for every 200 one with full ice storage, and one with weeklyice storage), square feet of gross floor area. "[’he lighting load at full (2) in comparison with economizer" use, (3) for different occupancyis 1.67 W/ft2 and2. equipmentloads are 0.8 W/ft fan-powered mixing box designs and control strategies for In addition maximum,domestic hot water use of 73,000 direct supply of cold air, and (4) for three different utility Btu/h (21 kW) and elevators that draw 195 Btu/h (57 rate structures. Most of the simulations were repeated for when in full use are assumed. Elevator use coincides three California climates, representing areas of potentially closely with occupancyof the building.

110 ASHRAETransactions: Research Infiltration is assumedto be approximately0.35 ach in a gas-fired domestic water heater. In the absence of an the perimeter zones and 0.25 ach in the core zones when evaporative condenser model in DOE-2, an oversized the building is unoccupiedand therefore not pressurized. cooling tower with two-speed fans is used to approximate To provide a broader basis for comparison, the simula- its performance. All simulations of component-assembled tions modeledtwo different design approaches to a conven- configurations use the samechiller performanceparameters, tional 55 °F (13 °C) variable-air-volume (VAV)system. which approximate an efficient positive displacement one case, two packaged VAVunits provide cooling by machine. Efforts were madeto obtain industrial-rated screw direct expansion and provide heating by hot water to chiller performance maps from manufacturers; however, terminal reheat coils. One unit serves primarily the core full curves were not made available. Therefore, these zones and the other primarily the perimeter zones. Theyare machinesare simulated at this juncture only in an approxi- equipped with an economizer option and are air cooled. mate manner, but the approximations match the manufactur- The second conventional 55 °F configuration, as well as er’s measurementdata whenever available. Positive dis- all cold air (42°F [5.6°C]) distribution systemsstudied, use placement machines, such as screw chillers, are more two simple VAVsystems with terminal reheat. The cooling tolerant of the higher differential pressures encountered coil is served by chilled water or a 25 % ethylene glycol during ice-making. At this time we are aware of only a few water solution, and both primary heating and terminal coils centrifugal machines that are recommendedfor ice-making are served by hot water. One system serves primarily applications (this is confirmed by our industry survey perimeter zones and the other primarily the core regions of results). For the component-assembledconfigurations, the the building. Supply and return fan sizes are determined circulation pumpsare variable speed and operate against the from building cooling loads and the minimumsupply air same head pressures in the two non-storage configurations temperature. and against an increased head in the case of the storage Selection of 42°F supply air temperature for the cold configurations. To decrease pumpingcosts of cool storage air distribution systemssimulated in this study was based on and retrieval, we use the seven-foot pressure drop charac- the results of our industry survey; 42°F represented the teristic of the commercially available ice-ball system lowest design temperature in use by consulting engineers (manufacturer’sliterature). All chiller capacities are chosen practicing in California. It should be noted that 44°F to to just meet the peak cooling demand(i.e., no reserve 46°F supply air temperature may be a more economical cooling margin) with the intent of reducing electrical costs. choice, dependingon climate and building characteristics. Someconfigurations showslight overloading during peak Primarysupply air static pressure is 4.0 inches of water demand periods, which in the case of the no storage and the return fan static pressure is 1.5 inches of water. configurations may occur during early morning pull-down Pressure is controlled by fan speed. Cooling setpoint for periods. In no case is the cooling system overloaded more zone thermostats is 76.2°F (24.6°C) whenever the mini- than 11 hours during the year. Elaboration of each configu- mumsupply air temperature is 55°F and adjusted upward ration follows. to 77.1°F (25. l°C) if 42°F minimumsupply air is speci- Case 1--55°F Packaged: A configuration that fied. This upward adjustment of the thermostat cooling represents approximately two-thirds of installed cooling setpoint is a conservative estimate of the occupant comfort capacity in the U.S. (Piepsch 1991) and is characterized benefits associated with the reduced humidities obtained relatively poor energy performance is the packaged, or with 42°F supply air temperature (Berglund 1991). The unitary, HVACunit. In the present study, this (rooftop) heating setpoint is 72.0°F(22.2°C) in all simulations in this unit is simulated to have economizer capability, using study. reciprocal compressors, direct-expansion cooling, and dry The outside air intake dampers are fully closed when cooling towers to provide 55°F supply air. No attempt is the building fans are off and during warm-upor sunmaer made to refine the performance of any component for pull-down periods. The minimumsupply air rate is 0.30 energy efficiency. cfm/ft2, whichis lower than normalpractice but is found to Case 2--55°F Base Case: In this, the base case, two provide adequate diffuser performance even at low supply chillers are simulated to provide 55°F supply air. One air temperatures(Gadgil et al. 1991). The outside air supply chiller is sized to carry one-third and the second two-thirds during occupancyis at least 19 cfm (9 L/s) per person. The of design loads and are operated throughout the year as VAVboxes can be throttled to 20%in the core zones and close to capacity as possible by hourly scheduling. This to 30%in the perimeter zones. Enthalpy-controlled econo- configuration was suggested by one of our utility contacts mizer operation is permitted wheneverambient air enthalpy and achieves lower electricity use than the often used is less than that of the return air. installation of two chillers of equal capacity. One preassembled (packaged) VAVsystem-plant Case 3--42°F without Storage: Again, two chillers configuration and five component-assembled VAVsystem- are simulated, sized as above to provide 42°F supply air on plant configurations were chosen for comparison and demandwith direct injection diffusers. investigation. The primary heating equipment for all Case 4--42°F with Half Storage: Ice storage is component-assembledconfigurations includes a gas-fired hot added, which is sized to meet half the cooling design day water boiler serving the main heating and terminal coils and load and to be charged during off-peak hours by a dedicated

ASHRAETransactions: Research 111 chiller. The storage tank is assurned to be insulated and measure (California Energy Commission 1992). Six buried to reduce thetxnal losses to approximately 2.5 %of simulations were performed covering the three clinmtes to capacity per 24-hour period. Measured thernml storage investigate the energy and cost penalties without economizer losses have exceeded 10%of capacity per ~4-hour period, use for both 55°F and 42°F supply air. resulting in significant penalties in overall armual perfor- Anotherpossible configuration was to avoid the use and mance(industry survey results; Merten et al. 1989). benefits of low-temperatureair by supplying 55°F air from The discharge rate of the storage is driven by demand ice storage. In most cases with this system, very poor use up to a maximumthat would deplete storage by the end of is made of the investment to produce the cool storage. the business day. If additional cooling is needed, a supple- However,prelinfinary simulations were performed with this mental chiller placed upstreamof storage precools returning configuration for the three selected California climates in solution to storage. Upstreamplacement enables the chiller view of the high percentage of time economizer operation to operate at a higher suction temperature and thus be more is possible during normal occupancy. This has been noted efficient (Peters et al. 1986). It is realized that in practice earlier by Hittle and Bhansali (1990). one may require a single machine to provide both ice The configurations presented here are believed to most generation and daytime assistance. However,for purposes directly explore the advantages and disadvantages of using of tabulating chiller performance,it is more convenient to TES/CoAD. simulate two machinesand separate their functions. Supply air is providedthrough direct injection diffusers. UTILITY RATE STRUCTURES Case 5--42°F with Full Storage: Ice storage is sized to meet all the needs of the cooling design day. The tank is Table 1 showsdetails of three utility rate structures, A, assumed to be insulated such that thermal losses are B, and C, that were adapted from those existing in Califor- approximately 2.5% of total capacity per 24-hour period. nia and modified for this study. Of the three, A is more The storage is charged during off-peak hours Sunday favorable toward load-shifting efforts because demand through Thursday, with any carryover from Friday contrib- charges from 10 p.m. to 6 a.m. are 17%of peak-period uting to additional storage losses. Supply air is provided demandcharges, which apply between the hours of 10 a.m. throughdirect injection diffusers. and 5 p.m. during the summer months. Even during the Case 6--42°F with Weekly Storage: Ice storage is winter months off-peak dernand charges are 72 % of peak. sized to meet all the needs of the most demandingfive-day Utility B summerdemand charges from 9 p.m. to 8 a.m. work week with charging permitted weekday evenings and are approximately 28 % of the peak demandcharges, which weekends if necessary to bring storage to capacity by apply between 11 a.m. and 6 p.m. For comparable periods, business hours on Mondaymornings. Storage loss rates are Utility C shows off-peak to be 36% of peak demand the same as those assumed for Cases 4 and 5. No supple- charges. Neither B nor C offer lower dernand charges mental chiller assists cooling demands during occupied during winter months, when cooling is often needed, and hours throughout the week, reducing utility demandcharg- thus do not benefit ice storage systems. Energy costs for es. Althoughnot used in this case, somestrategies schedule Utility A during the summeroff-peak per’iod are 49%of the chiller to operate 24 hours per day, allowing smaller on-peak rates, those for B and C are 46% and 52%, chiller and storage capacities to be installed. Supplyair is respectively, of on=peak rates. Off-peak winter energy provided through direct injection diffusers. charges for Utility A are 52% of peak, and those of Several other configurations fbr study were considered Utilities B arid C are both 87 %of’ peak. After careful and included, augmentingthe original scope of the study, as analysis of somesimulation results it was found that rate discussed briefly below. structure B caused the monthly utility charge limit of During the industry survey (Baumanet al. 1992), $0.16/kWhto be applied during several summermonths. was learned that most cold air delivery systems use fan- Ice making is confined to periods of low demand powered mixing boxes to prevent cold drafts in the occu- charge which begin weekday evenings at 10 p.m., and is pied regions of the zone. After laboratory rneasurements permitted, if needed, to continue to 8 a.m. Ice-making showedthat direct supply of low-temperatureair" could also schedules remain the samethroughout the year, as the three be successful (Gadgil et al. 1991), we decided to perform climates investigated have ~ignificant cooling requirements four simulations using fan-powered mixing boxes to during the winter months. It was found that slight changes investigate their’ energy use and operation costs, depending in rate structures mayalter the attractiveness of TES/CoAD on their design and operational control strategy. systems significantly. Wefound during the industry survey that the use of air-side economizers may be omitted in new high-rise RESULTS construction primarily because of cost and the reduction in revenue-producing floor area required by the additional The simtdation results are discussed below for the six mechanicalequipment. Apparently it is possible to satisfy system configurations and two of the three climatic regions California building perfon-nance standards without this (San Jose and Fresno). The simulation results for San

112 ASHRAETransactions: Research TABLE1 Utility RateStructures

Utilil ,A Utility B Utility C

Winter Summer Winter Winter Oct 1-Mar31 Apt 1 - Sept30 Nov1 - Apr30 May1 - Oct 31 Novi r Apt 30 May1 - Oct 31

On-peakdemand ($/kW) 3.94 16o92 4.20 16.00 4.15 11.60

Off-peak demand($/kW) 2.85 2.85 4.20 4.50 4.15 4.15

On-peakrate (S/kWh) 0.06876 0.07671 0.0700 0.1400 0.06544 0.11277

Partial-peakrate (S/kWh) 0.04305 0,05050 0.0700 0.1000 0.06544 0.07654 Off-peakrate (S/kWh) 0.03550 0.03753 0.0610 0,0650 0.05669 0.05843

Monthlyave. rote limiter (S/kWh) 0.35 0.35 0.16 0.16 0.15379 0.15379

Ratchet none none none nolle none

Week-dayon-peak hours 5 pm- 8 pm 10am-5pm none 11 am-6pm none llam-6pm Week-daypartial-peak hours 6 am- 5 pro, 6amo 10 am, 8am~10pro 8am- 11 am, 8am- 10pm 8am- 11 am 8 pm- 10 pm 5 pm- 10 pm 6pm-gpm 6pm-9pm Off-peakhours 10pm-6am* 10prn-6am* 10 pm- 8 am* 9pm-8am* I0 pm- 8 am* 9pm-8am*

*plus all 24 hoursweekends and holidays

Bernardino are not presented here due to similarities with that for the base case; however, the electrical energy used results for Fresno. Refer to Baumanet al. (1992) for for the compressors and cooling tower fans exceeds that of complete set of simulation results. The primary emphasis the base case by 80 %. These factors are primarily responsi- for comparisonpurposes is that of ventilating and cooling ble for increasing the percentageof total building electricity system electricity use and its effect on operating costs. used for cooling to approximately 22%versus 17%for the Capital costs are not quantified as a part of this study but base case. Comparing the 42°F supply air case without maybe inferred from equipment sizes. storage to the base case, even thoughthe total fan electrical First, the six system configurations are discussed. This use decreases by 36 %, the chiller use increases by more is followed by results for economizer use and, finally, than 113 % (comparedto the base case)~ which overshadows fan-powered mixing boxes (FPMB). Since many of the savings in fan energy use. Even though the fraction of total points of interest occur repeatedly during the discussion, electricity use attributable to building fans decreases to three figures are introduced at this time. Figure 1 presents 4.4 %, downfrom the base case of 7.6 %, the electricity use annual cooling and fan electricity use for the two climates, for cooling (chiller, pumps,condenser fans) inc, eases San Jose and Fresno. Figure 2 presents a comparisonof the about 17 %of the total building electricity use, comparedto annual total building operating costs for the three selected 9 %in the base case. utility rate structures for the two climates. Figure 3 sum- The half-storage configuration shows a 60 %increase in marizes the installed chiller and storage capacities for all plant cooling load, using 72 %more electricity for cooling cases except the packaged system for San Jose and Fresno. than the 55°F base case. Overall cooling system perfor- For Case 4, the partial-storage configuration, only the ice- mance as measured by COP shows a decrease of 12% makingchiller capacity is shown.All stated chiller sizes are comparedto the 55°F base case. Frequent part loading and rated for standard ARI conditions (i.e., 44°F [6.7°C] higher daytime condenser temperatures of the supplemental leaving water temperature, 85°F [29.4°C] entering con- chiller decrease its efficiency significantly. Energyuse by densing water temperature). the condenser increases roughly 40 %over that of the base case, in part because the cooling tower is sized to meet the San Jose requirements of both simulated chillers, resulting in the simulation of a somewhatoversized fan. Cooling and fan energy use for the San Jose climate Of the total energy stored in the form of ice, approxi- amongthe six system configurations are summarized by mately 3.3% is lost through thermal gain over the year, Figure la. It is apparent that the 55°F packaged system, 96.2 %is recovered for building cooling, and the remainder case 1, is the most energy intensive amongthe six configu- is carried over to the next year on the last day of the rations, followed by the 42°F supply air system without simulation. Even though the 22 MBtu(1,830 ton-h, 6.4 storage. Fan electrical use for case 1 is nearly identical to MWh)storage is sized to meet half the design-day cooling

ASHRAETransactions: Research 113 8oo~ I8oo 700- 700j...... 600- 6001...... 500d...... 500- 400- 300 ...... ~ ...... 300-

200 ...... 200-

10 ...... 100- ~ 0 55OF 55°F 42°F 42OF 42OF 42°F 55OF 55°F 42OF 42OF 42OF 42°F S~ ~ S~ S~e Storage Storage Storage Storage

[~ Cooling Energy ~ Fan Energy ] Cooling Energy ~ Fan Energy

Figure la Annual cooling and fan electricity use: San Jose. Figure lb Annual cooling and fan electricity use: Fresno.

350

300

250...... ~ ...... 250

200- 200"

150- 150.

100- 100.

50- 50

O-

~ 42°F 1/2 Storage ~ 42°F bed[ Storage ~ 42°~z~" Storage

Figure 2a Annual total building electricity cost: San Jose. Figure 2b Annual total building electricity cost: Fresno.

2.00 4O

1.75" 1.751...... ~: 1.50- 30 1o50"

1.25- ,_~ c, 1.00- 20 20 ~ o o ~ 0.75-

o 0.50- ’10 .... 10 ~

0.25" 0~25-

0.00- 0 ~ 0.00- 42OF ~eekly Case Storage Sbrage Storage Storage Case Storage Storage Storage Storage I~ Chiller ~ Storage I I[ ~ Chiller ~ Storage I Figure 3a Installed chiller and storage capacities: San Jose. Figure 3b Installed chiller atwl storage capacities: Fresno.

114 ASHRAETransactions: Research needs, about 83 %of the total annual cooling load is met by Figure 2b is a summaryof the utility electricity costs storage. Storage is capable of meeting nearly all cooling using the three rate schedules whenapplied to the simula- needs during the winter periods as hourly reports showno tion results for the Fresno climate. The packaged unit supplemental chiller use during the months of December increases peak demand by approximately 27%. The half- and January. storage case shaves peak demandby 17 %; full and weekly Even though the full-storage case causes the plant storage reduces it by about 33 %. Whenthe advantages of cooling load to increase by 65 % over the base case and using off-peak energy are included, utility rate schedule A chiller electricity use to increase by 104%, the lower produces a cost savings of 11%for the half-storage case off-hour rates and a 30 %reduction of peak demandactually and 17 %for both the full and weekly storage cases. The lower the annual electricity costs. Twodedicated ice-making rate structure of utility B changes operating costs of the chillers of 5.0 MBtu/h(417 tons, 1.5 MW)total capacity storage options only slightly from the base case. Utility rate are necessary to meet the larger loads. It is also worth C offers a 5% cost savings for half storage and a 10% mentioning that the lower nighttime ambient temperatures savings for full and weekly storage. The packaged system reduce condenser fan kWhfor the full-storage case to 12% causes cost increases of 19%with Utility A and 14%with below the base case use. both Utilities B and C. Case six, using five-day storage, compared with the Figure 4 displays monthly chiller COPin response to base case uses 34 %less fan energy, but a 75 %increase in varying operating conditions. Chiller COPexcludes the plant cooling load increases total electricity used for cooling auxiliary energy consumption of cooling tower fans, by 85%, with chiller use alone increasing by 115%. A condenser water pumps, and cold loop pumps. Chiller COP storage capacity of 105 MBtu(8,750 ton-hr, 30.8 MWh) for the packagedunits is not included. In the figure, six required to satisfy 100 %of the cooling demandsbut loses chiller COPsare shown, as the performancesof two chillers approximately 9 % of generated cooling capacity annually. are modeledseparately for the half-storage configuration. The significance of utility rate structures to offer As discussed previously, one is the main chiller, dedicated encouragementto shifting building electricity use to off- to ice making (42°F 1/2 Sto Main), and the other is the peak periods-reducing daytime peak demand--is seen in supplemental chiller, used for producing chilled water Figure 2a. In this figure, the annual electricity costs are (42°F 1/2 Sto Supp). The same trends are observed (and shownfor three rates if applied to the six configurations in not shownfor brevity) for the five configurations in the the San Jose climate. Using utility A rate structure, the half other two climates studied. As mentionedpreviously for the storage case results in an 8.7 %decrease in electricity cost, half-storage configuration, frequent part loading of the full storage in a 13.5%decrease, and weekly storage in a supplementalchiller reduces its performanceto nearly that 13 % decrease below the base case. The packaged system of ice-making levels. During the months of January and raises electricity costs by 11.6 %and the 42°F no-storage December,the supplementalchiller does not operate; during case raises them by 5.3% over those of the base case. February and November,only infrequent and light loads are Utility rate B produces a 2 %increase for half storage and encountered. This is one of the pri~nary reasons whyoverall a 1% decrease for both full and weekly storage. The cooling COPfor the half-storage configuration is only packaged unit raises electricity costs 7.5%, and the 42°F slightly better than the full-storage case. The half-storage no-storage raises them 6.5 %above those of the base case. ice-making chiller is simulated to have a higher average Electricity costs based on utility rate C showtrends similar COPthan its counterparts in the full and weekly storage to those for utility rate B.

8 Fresno

Fresno represents an inland region of greater dry-bulb temperature swings and generally lower relative humidities than San Jose. Figure lb presents a summaryof the cooling and fan energy use for the six cases for the Fresno climate. As in the coastal climate, the base case uses the least electricity and the packagedunit the most. The penalty for using dry coolers appears to be convincing, as the packaged unit uses 97 % more electricity for cooling than the base

case. The second highest electricity use is the 42°F no-sto- J F M A M J J A S O N D rage case. Annualtotal building electricity use by the packaged system is 12.1% above the base case, the 42°F "-’=- 550F w[o Storage --+-- 4~°F ~[o Storage ~ 4~°F 1[~ Sto Main no-storage is 4.4% above, half storage is 1.6% above, full ~ 42°F i/2 Sto Supp-++- 42°F FulI Storage ~ 4~2°F Weekly Storage storage is 2.3 % above, and weekly storage is 2.9% above the base case. Figure 4 Chiller monthly COP: Fresno.

ASHRAETransactions: Research 115 cases becauseheat can be rejected fi’om a cooling tower that is sized for both the supplementaland ice-making machines. The advantageof only nighttime chiller operation, as in the case of full storage, reduces fl~e condenser fan energy use 0.9--- below that of the base case, and together with the cooler condenser temperatures, reduces chiller head pressures. 0.8- --~ Chiller performance for the weekly storage case is very similar to that of the full-storage case. Theratio of the total heat dissipated by the condenserto the total cooling load is 0.6- greatest in the case of the 42°Fdelivery air without storage. This is indicated by the consistently lower COPthroughout the year with noticeable part loading during January and 0.4 December.This effect would be more severe if two chillers of equal capacity were installed. Storage -+-- 42°F ~ Storage ~ 42°~’ WeeklyStorage Figure 5 showsthe fraction of stored energy recovered by the monthfor well-insulated and properly sized storage in the Fresno climate for the three storage configurations. Figure 5 Storage energy recovery ratio, monthly: Fres- ]’he low fraction for the monthof January is a simulation no. artifact that has been brought to the attention of DOE2.1E authors. It is noted that half storage yields roughly 5% economizeruse based on the following two strategies. (1) greater fractions of recoverable energy than full storage In coastal climates, ambient air is used only when the during the cooler monthsas a result of greater drawdown. dry-bulb temperature is more than 10°F below the building As detnand for cooling increases during the warmer return tetnperature. (2) In the drier inland climates ambient months, these differences becomemuch less as both are air is used only whenthe dry bulb temperature is morethan used to near design capacity. Weeklystorage losses are 5°F below the return air temperature. Admittanceof higher higher because of the larger surface area and low fractional ambient temperatures for economizer use commonlycauses drawdownduring winter months. Half storage provides an muchgreater latent loads. Wheneconomizer operation is annual storage energy recovery ratio of 95 %, full storage enthalpy controlled, energy consumptionis further reduced provides a ratio of 92 %, and weekly storage provides 89 %. due to improved control of both latent and sensible loads Half storage is able to satisfy 76 % of annual cooling load imposedon the cooling coil. It is recognized that enthalpy- requirements, and full and weekly storage satisfy 100%.It controlled systems need more tnaintenance than dry-bulb can also be inferred that multiple storage modules would controllers. even the profiles considerably and reduce lost chiller work. Figure 6 is a comparison of important componentsof At the present time, simulation of more than one cold the cooling energy use for the two climates, San Jose and storage tank is not possible. Fresno, using 55°F and 42°F supply air with and without economizeruse. Figure 6a shows that failure to use econo- Economizers mizers in San Jose with 55°F supply air increases the plant cooling load by 78 %, the total electricity used for cooling "l’he economic benefit of an air-side economizer is by 65 %, and fan electricity by 6 %. The slight reduction in expected to be sensitive to climate and the cooling coil fan energy use with an economizer results from a lower temperature. Additional sirnulations were run with econo- load on the cooling coil and a reduced supply temperature mizer use enabled and disabled for the no-storage cases of permitted by an approximate four-degree throttling range of 55°F and 42°F supply air so benefits could be more easily the mixed-air temperature controller. "l’he annual building compared. All economizer simulations are with enthalpy electricity use increases by 7%as a result of disabling control, whereuse of outside air is rejected if enthalpy is economizers,causing utility electrical costs to increase by greater than that of the return air. Because of reduced 5 %for utility A and 6 %for utilities B and C. building air humidity ratios whenlow temperature supply If 55°F air rather than 42°F air is supplied from ice air is used, the ambient enthalpies that are greater than storage in the coastal climate of San Jose, the opportunity about 25 Btu/lbm (58 kJ/kg) dry air must be rejected when to use economizers is greatly increased. The extent of 42°F supply air is used, and ambient enthalpies that are economizer use affects the number of hours the night- greater than about 28.8 Btu/lbm (67 kJ/kg) dry air must scheduled chillers tnust operate to supply ice. A simulation rejected when55°F supply air is used. was made to investigate this configuration, and in this A series of simulations were performed in each of the preliminary analysis, the energy used for cooling is about three climates studied, with the economizeruse controlled 20%greater than that of the base case. If the fan energy only by dry-bulb temperature in the case of 42°F supply required is added to the total cooling energy, the total air. Energy consumption is significantly reduced with annual electricity use for ventilation and cooling is lower

116 ASHRAE Transactions: Research 700- 700 600’ 600’ 500 500 400 400- 300 300- 200" 100-

55°F 55OF 42OF 42°F 55°F 55OF 42OF 42OF BaseCase BaseCase w/oStorage w/oStorage BaseCase BaseCase w/oStorage ~/o ~ w/Econ w/o ~. w/~n w/oEcon w/Econ w/o Econ w/Econ w/o Econ [~ Cooling Energy [~ Fan Energy

Figure 6a Effect of economizer operation on annual Figure 6b Effect of economizer operation on annual cooling and fan electricity use: San Jose. cooling and fan electricity use: Fresno.

than any of the four 42°F supply air configurations. The use is possible in 93 %of the normal occupancyperiods if total building electricity use increases only about 1.8 %over 55°F supply air is specified and in 58 %of the sameperiods that of the base case, the peak demanddecreases by 24 %, if 42°F supply air is used. and the electrical operating costs decrease about 11%for Fresno’s climate shows a smaller cost penalty for not utility A, are unchangedfor utility B, and decrease nearly using the economizerthan does the cooler coastal San Jose 6 %for utility C. area. Figure 6b shows the effects of using an economizer Penalties are less severe for not using an economizerin on annual and fan energy use in Fresno. For the 55°F case configurations with 42°F supply air. In the California without economizer, fan energy use remains nearly con- climates studied, the hours of availability of sufficiently low stant, the total plant cooling load increases by 33%, and ambiententhalpies (less than about 25 Btu/lb dry air) during total electricity use for cooling increases by 32%. Total periods requiring building cooling are greatly diminished building kWhincreases by 4.3 %causing electricity costs to from the case of 55°F supply air (which requires ambient rise by 2.4 %for utility A, 3.3 %for utility B, and 3.1%for enthalpies to be less than about 28.8 Btu/lb dry air). Thus, utility C. Use of 42°F supply air lowers building relative for the San Jose region for the 42°F supply air configura- humidity levels to roughly 30% and, coupled with the tion, the increase in total coolingload and in total electricity usually muchhigher daytime temperatures of the region, use for cooling owing to not using a economizer is 17%. reduces the penalties for operation without economizer.The Building fan electricity consumption increases by 4.5%, total fan electricity use decreases7.6 %and the total cooling havinglittle impacton the total building electricity use, and load and electricity use for cooling increase by about 13 %. operating costs increase by approximately 2% for all The total building electricity use increases by 2.6% and utilities. utility electrical costs by 1.3 %for utility A to 1.8 %for Figure 7 shows the frequency distribution of ambient both utilities B and C. The preliminary simulation with enthalpy for the San Jose climate. Fromthe figure, one can 55°F air from full ice storage, in place of the standard deduce the maximumnumber of hours during the year 42°Fair, results in total cooling and fan electricity use that economizeruse is possible for the two cases of 42°F supply is approximately 9.6% greater than the 55°F base case. air and 55°F supply air. Since weather patterns are not However,the operating costs are 13 %lower than the base significantly affected by the day of the week, the ratios of case for utility A, almost identical with base case costs with maximumeconomizer availability (presented inside the box utility B, and 7 %lower than base case costs for utility C. in Figure 7) are calculated by dividing the number of Figure 8 shows that the fraction of business hours when business hours for which economizeruse is possible by the economizer use is possible using 55°F air is 79%and with total business hours of the year, using all seven days per 42°F air is 58%. week. The ratios do include hours during which the The 42°F supply cases without economizer use for San building does not need cooling. Betweenthe horizontal lines Bernardino showed an increase in total building cooling at 28.8 and 25.0 Btu/lbm dry air are the additional number load and electrical use for cooling of 13 %and an increase ot/ business hours throughout the year that permit use of of 2.7 %in electricity use overall, resulting in cost increases economizers when55°F supply air is used instead of 42°F of 1.5 to 1.8%. It is interesting to compare performance supply air. For the San Jose area, thhs implies economizer differences for San Bernardino between the full-storage

ASHRAETransactions: Research 117 40 (3051 Poss~ble/3285 Total) SA= 55 F h<=28

(2023 Posslble/3285 Total) S,A.= 42 F h<=25.0 .-~ 35 - 0

h=2 ].8

-*- 25

.12

Sat Jo ~e

5 0 2O ~o 60 ~o ,oo ~2o t~o ,6o ~o 2oo 10 ~o ~o ~o 9o ,,o ~3o ,5o ~7o ~9o Frequency (hours)

F/gure 7 Annualmaximum economizer use, enthalpy controlled: San Jose.

Figure 8 Annual maximumeconomizer use, enthalpy controlled: Fresno. low-temperaturecase and the 55°F no-storage case when Fan-Powered Mixing Boxes (FPMB) neither uses economizers.The total building kWhdecreases by a negligible 0.2%for the storage case, but the peak Four simulations were conducted in which system demanddecreases by 31%,and the annual electricity costs terminal fan-poweredmixing boxes were addedto investi- decrease by 18%for A, 6.8%for B, and 12.4%for C. If gate their energy use and explore control strategies when similar comparisonsare madefor the coastal climates, the they are used with cold air delivery systems. FPMBsare penalty for not using the econotnizerwith 55°Fair is even helpful in maintainingairflow in conditionedzones, espe- greater. cially when the system fans are operating at or near

118 ASHRAETransactions: Research minimumflows. These units induce plenum air, which may room air with the cold supply air whenever the supply be return air from other zones in the building, and reintro- reaches small and mediumflow rates. At these flow rates, duce it to the zones they serve along with whatever primary it is often suspected that the room air maynot mix well air is supplied. Because fractional horsepower motors are with supply air and the roomoccupants maybe subjected to used, these units are inefficient compared to the main uncomfortabledrafts at times whenthe diffusers would still supply motor-fan combinations. They can provide an be at less than design flows. With this control strategy, effective first step in meeting heating demands, as the electricity use by the chiller decreased by approximately plenum air is frequently several degrees warmer than the 18 %. Total fan energy remained very close to the non- conditioned space below and the added heat from the unit FPMBcase. This intriguing simulation result from DOE- motor itself elevates the reintroduced air temperature. 2.1E seemsto indicate that zones requiring cooling (interior Concern over higher operating costs is balanced by the zones) maybenefit from the use of return air from those designer’s greater confidence of occupant comfort when zones requiring heating (exterior zones). This results in FPMBsare installed in buildings using cold air. This reduction in primary fan delivery as well as the decrease in provided the justification for exploring this dimension of cooling energy use. While we believe this energy-conserv- practice with four simulations. ing effect is real, we are currently unable to verify the Because manydesigners use these units as air "blend- magnitude of the predicted cooling energy reductions. ers," with installations serving essentially all conditioned Authors of DOE-2.1Ehave been notified, and this effect is zones rather than only the perimeter zones, the building under investigation. During our industry survey, we did not zones were, for simulation purposes, reconfigured so that encounter any of the above FPMBcontrol strategies in use. all zones are equipped with FPMBs.The simulations were The fourth simulation used a series FPMBthat operated based on the San Jose climate with full ice storage configu- whenever the main building fans were on (100% of the ration. Thus, no cooling equipment is in use during FPMB time). Series units must handle primary air as well as the operation except for circulation pumps supplying the induced air and for this preliminary investigation were sized cooling coil. to handle 110 %of the primary airflow. Althoughthis is not Four graduated control strategies of FPMBuse were an energy.-efficient strategy, a significant number of explored for their impact on total building fan energy. The designers workingwith cold air distribution use continuous- first three simulations used a parallel FPMBconfiguration, ly operating series or parallel mixing boxes to ensure which does not handle primary airflows but supplements successful diffuser performanceat all times. primary air to diffusers with plenum air that mayor may A manufacturer’s performance characteristics for an not be passed across an active heating coil. All parallel efficient fan and motor were used for the simulations. units were sized to handle 80%of the maximumprimary Figure 9 presents the simulation results in terms of fan airflow to the zone served. The zone temperature served by energy (MWh)and fan-attributable fraction of total building each unit usually controls the fan motor. electricity consumption.Figure 10 presents the effects of In the first simulation, the FPMBunits were activated FPMBoperation on total annual building electricity costs in if the zone temperature fell belowthe deadbandmidpoint to terms of the percentage change from the case with no assist in heating the zone, at first by merely reintroducing FPMB.The lowest temperature setpoint simulation shows warmer plenum air (which also has the fan motor heat the least impact, especially in California climates whichdo added). If the temperature continued to fall, the terminal not have the extended heating periods other parts of the coil would carry hot water for active heating of the zone. country do. The total fan energy use increased 1.6 %raising This was approximately equivalent to operating the mixing the total building electrical use and utility costs by less than fan only as a boost to supply airflows during the heating 0.15 % with no noticeable impact on building peak demand mode. This FPMBcontrol strategy enhances diffuser since these units would not be operating during peak performance only under heating conditions. demandperiods. Results from the second simulation show In the second simulation, the FPMBunits were activat- that building fan use increases by 12.4%over the no-FPMB ed near the top of the deadbandto assist air motion within case, and total building electricity use and utility costs the zone whenthe primary flow rates to the zone are near increase by about 0.5 %without raising the peak demand. or at their minimums.As the zone temperature falls (i.e., This indicates that the majority of time the building is in the heating mode),its operating strategy duplicated that conditioned, zone temperatures are muchmore likely to be of the first simulation. This control setting did not affect on the high side of the deadband midpoint. WhenFPMB peak demand charges, which are generated at times of operation extends into the active cooling temperature maximumcooling loads. A slight reduction in heating ranges, as in the third simulation, total fan energy remained energy was observed as expected, although annual heating at the same level as the previous case, with a negligible costs did not exceed $5,000 in any of the simulations. increase in peak demandover the no-FPMBcase. For this For the third simulation, the same FPMBunits were case, utility A costs were essentially unchanged,and utility activated slightly above the temperature at which active B and C declined by 0.9 %. The cost reduction originates cooling begins. This strategy ensures the mixing of warmer primarily in the reduced chiller energy, which, as men-

ASHRAETransactions: Research 119 12%

10%"

8%"

6%"

4%"

2%"

0%-

0 "0% No Series, No ’ Parallel,’ Parallel,’ Parallel,’ Series, FPMB Heat Minimum Begin 100%on FPMB Heat Minimum Begin 100% on Assist Prim Air Cooling Assist Pdm Air Cooling [~MWh ~% ] E~Rate A ~ Rate B B Rate C ]

Figure Effects of fan-powered mixing boxes (FPMB) Figure 10 Effects of fan-powered mixing boxes (FPMB) on total fan electricity use: San Jose, full on annual total building electricity cost.. San storage, 42°F. Jose, full storage, 42 tioned, declined about 18%. If series units are operated engineers, equipment manufacturer representatives, re- whenever the primary fan is turned on, the fan energy searchers, utility and energy commissionrepresentatives, increases by roughly 150%, the peak demand increases and other users of cold air distribution technology. The about 4.2 %, and the utility costs for all three rates increase contact list was started through a few well-knownconsulting approximately 10%to 11%. In terms of fan electricity engineers and a recent workshopon cold air distribution consumptionas a fraction of total building electrical use, (EPRI 1991a). The list grew largely through word of mouth the heating-only strategy (first simulation in this series) as we attempted to identify all major cold air distribution affected fan energy percentage negligibly, and the air projects and associated users of CoADtechnology in motion assist raised the fraction from about 4.5% to 5.0 %, California. A complete list of contributors to the survey the limited cooling range to 5.1%, and the series application appears in Baumanet al. (1992). The purpose of the survey to 10.3%. was to assess the current state of practice and future The significant finding illustrated in Figures 9 and 10 directions and needs of this technology, with an emphasis is the following. If the FPMBsare operated continuously on conditions in the state of California. The results of the (strategy 4), they lead to a substantial increase in fan energy survey had three major uses. use and on-peak demandcompared to the cases of direct supply of cold air (no FPMB)or when parallel FPMI3sare 1. Support for Whole-Building Energy Simulations. operated under various zone control strategies. Strategies 1, Performance data from equipment manufacturers 2, and 3 avoid or have little impact on peak demandand helped to more accurately specify the cooling plant and elevate fan energy marginally by operating the FPMBsonly HVACequipment. Recommendations and design whenthe cooling loads are small or medium.Thus, they do approaches used by practicing engineers provided not lead to larger energy or peak-dernandpenalties while at additional guidance for identifying a system configura- the same time ensuring adequate diffuser performance at tion and operating strategy that represented an energy- low supply rates of cold air’. Simulationresults indicate that efficient design in the current building market. chiller use may actually be decreased when FPMI3sare 2. List of California Cold Air Distribution (CoAD)Pro- operated within the cooling throttling range (strategy 3). jects. If sufficient data from further research can identify Oneof the best ways to obtain information on cold air optimum control strategies for FPMBs,these results distribution from our contacts was to discuss their indicate that a satisfactory compromisecan be reached experiences with completed or ongoing projects. Based betweenthe designer’s reluctance to specify only direct cold on the information nmdeavailable to us, we compiled air supply and the unwarranted on-peak loads and energy a list of current California projects involving cold air use caused by continuously operated FPMBs. distribution (see t3aumanet al. 1992). 3. Factors Influencing the Developmentof CoMAir Distri- INDUSTRY PERSPECTIVE bution in California. The information gathered from our contacts provided A survey was conducted during the months of May a firsthand look at the reasons behind their willingness through November1991 by interviewing several practicing or reluctance to consider cold air distribution in their

120 ASHRAETransactions: Research building designs. An assessment was made of the respondents was to determine which method would be used advantages, disadvantages, and future trends of CoAD to deliver air into the conditioned space. As in any air technologyin California based on the gathered informa- distribution system, it is essential that, for both cooling and tion. heating modes, satisfactory roomair diffusion be provided to ensure acceptable thermal comfort conditions and indoor The ways that the survey results were used to develop air quality for the building occupants. In light of the a realistic plant and HVACsystem configuration have been increased public awareness of building comfort and air discussed in the previous section on simulation results. In quality issues, consulting engineers are not willing to take the following section, we discuss a few of the important any substantial risks in this area. factors influencing the developmentof this technology in The direct supply of 40°F to 50°F (4.4°C to 10°C) California. temperature air for space cooling, while using the least amount of fan energy, increases the chances of poor Factors Influencing the Development diffuser performance. Practicing engineers and researchers of ColdAir Distribution in California alike recognize that the higher buoyancyforces associated with larger room/supply temperature differences combined The discussion in this section is based on the knowl- with lower inertial forces from reduced diffuser outlet edge, experiences, and opinions of those contacted during velocities are driving the fundamentals of the flow problem the survey. Since our survey focused on users of CoAD in the direction of diffuser failure. Not surprisingly, most technology, the results are representative of a sophisticated, of the CoADprojects surveyed used fan-powered mixing highly knowledgeable, but small segment of the currently boxes to raise the air temperature and flow rate before practicing building/HVACindustry as a whole. delivering it to the space. Only two CoADprojects supplied First-Cost Cork~iderations In its current state of cold air directly to the space: one utilized new diffusers development, cold air distribution can provide important designed specifically for cold air applications and the other economicbenefits to ice storage systems. While the desir- used a 50°F supply temperature at the high end of the range ability of thermal energy storage is primarily driven by associated with low-temperature design. utilities, whoprovide incentives to commercialcustomers The major reasons (offered unanimously by the engi- for shifting energy demandfrom peak to off-peak hours, for neers interviewed) for avoiding the direct supply of low- most commercialbuildings with ice storage, the first-cost temperature air (40°F to 50°F) into the conditioned space savings of including cold air distribution are indisputable. of the building were (1) lack of availability of supply For the majority of buildings surveyed, the combination of diffusers with the manufacturer’s backing from performance CoADwith TES produced a design that was much more testing under low supply-air temperature conditions and (2) competitive in first cost than a comparable TES system lack of confidence in the capability of a low volume alone was with conventional HVACdesign. With further cold-air distribution system to provide effective ventilation research and development,cold air distribution installations throughout the conditioned space of the building. will provide even greater economicadvantages for utilities Recent research (Gadgil et al. 1991; Miller 1991) has and their customers. shownthat currently available diffusers can provide accept- Almost every high-rise CoADbuilding surveyed listed able roomair diffusion with cold supply air temperatures space savings and the resulting cost savings as a major under the right operating conditions. A few engineers consideration for the inclusion of the CoADsystem. expressed a willingness to use diffuser outlet temperatures Smaller-sized ducts (typical ducts for a CoADsystem are as low as 45°F (7°C) when providing maximumcooling. 25 % to 40 % smaller than for a conventional 55°F system) They were confident that the higher supply-air volume allowed designers to offer a reduction in floor-to-floor under peak-load conditions was sufficient to ensure satisfac- height to the building owners. In at least two cases, this tory diffuser performance. On the other hand, some allowed one additional floor to be added to the building engineers avoided anything but conventional supply air while staying below the maximumbuilding height limitation temperatures by specifying continuously operating fan- imposedby the local building code. In one of these build- powered mixing boxes in all of their designs. At the ings, by combining smaller ductwork with structural reduced supply volumes characteristic of part-load condi- modifications to allow extra space for ducts to pass major tions, however, almost all engineers were reluctant to support beams, the floor-to-floor height was reduced by continue to use 45°F diffuser outlet temperatures. nearly eight inches. For this building owner, the rental Fan Energy Energy use of fan-powered mixing boxes revenue alone from the one additional floor far outweighed remains a concern for manyconsulting engineers because the extra construction costs and was the major economic of the long hours of annual use of these devices. Despite incentive for installing a TES/CoADsystem. these considerations, if faced with designing a cold air Room Air Distribution In a CoADsystem, after distribution system, nearly every design engineer inter- distributing the low-temperatureprimary air throughout the viewed would choose to specify parallel or series fan-pow- building, a critical consideration reported by all survey ered mixing boxes to ensure that conventional (55°F to

ASHRAETransactions: Research 121 65°F [13°C to 18°C]) diffuser outlet temperatures and approaches, (4) building codes and standards, and (5) utility adequate roo~n air circulation are maintained. In their incentives. Refer to Baumanet al. (1992) for a full discus- opinion, the energy penalty and increased installation, sion of the survey findings. operation, and maintenancecosts associated with the use of mixing boxes is muchmore preferable than the perceived alternative scenario: the potential for problemsand occupant CONCLUSIONS AND FUTURE complaints about air quality (e.g., poor air circulation) and TECHNOLOGY NEEDS comfort (e.g., diffusers dumpingcold air into the occupied zone of the building). A study was completed to provide a current assessment Economizers High-rise HVACdesign practice has of cold air distribution technologyin California. A series of changed over the past five to ten years in response to energy simulations of a three-story prototypical office changing tenant needs. As multi-tenant office buildings are building were completed using the DOE-2.1Ecornputer nowthe norru, most design engineers are installing floor- program. These simulations were used to exa~nine the by-floor air-handling units (AHUs)on all of their high-rise energy use and operating costs for six system configura- jobs. In a floor-by-floor air distribution system, the inclu- tions: (1)packaged system using conventional 55°F (13°C) sion of an air-side economizercan create big headachesfor supply air with no energy conservation strategies, (2) the owner and design team. If a centralized economizer is component-assembled conventional 55°F (13°C) air used, the same large space-consumingvertical shafts that distribution system without storage, (3) 42°F (5.6°C) can be eliminated with the floor-by-floor design become supply air with conventional chiller without storage, (4) necessary. On the other hand, if fresh air intakes and 42°F (5.6°C) supply air with partial (half) ice storage exhaust vents are provided for each AHU,the appearance system, (5) 42°F (5.6°C) supply air" with full ice storage of the building’s facade can be severely impacted and most system, and (6) 42°F (5.6°C) supply air with weekly probably will be rejected by the architect. The implications storage system. All of the above simulations assumedthe of these considerations are that the majority of high-rise direct supply of conditioned air (42°F or 55°F) to the space buildings today do not have air-side economizers. In terms without the use of fan-powered mixing boxes. Simulations of energyconsumption, this situation flies in the face of the were repeated for three California climates, representing well-recognizedsignificant cooling energy benefits associat- areas of potentially rapid growthin newoffice construction: ed with economizeruse in California (see Figure 6). San Jose, Fresno, and San Bernardino. Additional simula- Controls and Operation Most of the engineers tions were performed to explore energy use and operating cost implications of restricted econornizer use and different stressed the critical importance of controls to a successful TES/CoADsystem. "Controls cause 90% of the problems control strategies for the use of fan-poweredmixing boxes. with TES/CoADjobs" was a commonphrase offered. As A survey was completed of consulting engineers, control strategies tend to be morecomplex for effectively equipment manufacturers, researchers, utility representa- operated systems, direct digital control (DDC)systems were tives, and other users of cold air distribution to assess the current state of practice related to CoADtechnology in considered essential, and somecontrols experts reco~nmend- California. Factors influencing the future developmentof ed the use of a moreexpensive industrial-grade controller cold air distribution were identified and discussed. on all TESprojects. The cost of one major breakdowncan The major conclusions from the whole-building energy easily be more than the difference in cost between a simulations were as follows: standard and higher quality control system. Most of the consulting engineers with whornwe spoke emphasized the significance of building conunissioning. 1. In all three climates, annual cooling energyuse for the Commissioningis important to the successful operation of four cases involving cold air distribution was always any air-conditioning system but, since many building greater than the base case, a conventional component- designers, contractors, and operators are relatively unfamil- assembled55°F air distribution system without storage. iar with TES/CoADtechnology, the co~nmissioning process The most energy-intensive of the six cases studied was becomeseven more critical for these systems. If a system the packaged system (configuration 1 above) followed is commissioned,its operation will more likely be optimized by the 42°F without storage case (configuration in terms of reduced costs, reduced energy use, and fewer above). Annualcooling energy use for these two cases occupant complaints. The selection of an experienced nearly doubled in comparisonto the base case. controls contractor plays an important role in the success of 2. Fan energy use for the four cases involving cold air building commissioning. After construction is completed, distribution always decreased in comparisonto the base the contractor will have primary responsibility for working case. These savings helped but did not completely closely with the building operator during building start-up offset the cooling energy increases. and the initial charge of the storage system. 3. Comparedto the system configuration using cold air In addition to the topics discussed above, the industry distribution without storage, all three combinationsof survey addressed (1) equipmentfor cold air distribution, (2) ice storage with 42°F supply air always reduced moisture and condensation, (3) system design engineering cooling and total building energy use.

122 ASHRAE lransactions: Research The base-case configuration always producedthe lowest indoor air quality, without the use of fan-poweredmixing total building energy use. However,with the fairly boxes. efficient cold-air systemdesigns used in this study, the Issues for further research are listed below. With largest increase in predicted total building annual advancements in these areas, a reasonable goal for a energy use was only 6.3 % over the base case for 42°F well-engineered ice storage/cold air distribution system without storage in San Jose, and the largest increase in would be to use the same or less total energy comparedto total building energy use for an ice storage/CoAD a conventional system design. system was only 4.8% over the base case for 42°F with weekly storage, also for San Jose. 1. Develop improved methods for supplying low-tempera- The reduction in peak electrical demandfor the three ture air to occupiedspaces. ice storage/CoAD systems (approximately -15 % for 2. Newproducts and operating strategies are needed for half storage and -30% for full and weekly storage) the following three categories: contributed to lower annual operating costs in compari- (a) fan-powered mixing boxes, including improved son to the base case when a favorable utility rate control strategies and more efficient motors and structure was applied (-8 %to - 11%for half storage fans; and - 13 % to - 17 %for full and weekly storage). The (b) system-powered induction boxes; and highest annual operating costs were consistently ob- (c) direct supply of low-temperature air through high tained for the packaged system in all climates and for induction diffusers. all utility rates. Since cold air distribution without 3. Research required to support the above developments storage provided only minimal, if any, peak demand include (1) independent laboratory testing of cold air reductions, operating costs were always higher than the delivery products, (2) field monitoring of operational other cold air systems and the 55°F base case. CoADsystems, and (3) detailed numerical modeling Economizer use played an i~nportant role in energy improve our understanding of the fundamentals of room savings, particularly for mild marine-influenced Cali- air motion and air quality resulting from the use of fornia climates. In San Jose, failure to use an econo- CoADsystems. mizer with 55°F supply air increased the total building 4. Systemenergy use and operating costs are highly sensi- annual electrical use by nearly 7 %and operating costs tive to chiller configuration and operating conditions. by 5 % to 6 %. The economizer penalty was so severe Chiller optimization studies could provide significant that, if it was not included in the base-case 55°F supply energy and cost benefits. air syste~n (a surprisingly commonpractice in high-rise 5. Implementationof effective control strategies can make construction in California, as discoveredin the survey), the difference between a successful TES/CoADinstal- the comparative energy picture for TES/CoADsystems lation and an unsuccessful one. Development of new was significantly improved. A sample simulation found and advanced control strategies can help to further that operating without an economizerin San Bernardino optimize the overall system performance. using full storage and 42°F supply air used essentially the same amount of energy annually (0.2% decrease) ACKNOWLEDGMENTS comparedto the 55°F supply air case without storage and also without an economizer. Use of fan-powered mixing boxes (FPMBs)increased This workwas supported by the California Institute for distribution energy consumption and peak demandover EnergyEfficiency (CIEE), a research unit of the University a wide range (from 0%to 150%), depending primarily of California. Publication of research results does not imply on the type of FPMBand modeof operation. CIEEendorsement of or agreement with these findings nor that of any CIEE sponsor. We would like to thank Fred The industry survey demonstratedthat cold air distribu- Buhl, Ender Erdem, Fred Winkelmann,and Bruce Birdsall tion (and ice storage) systemsare still not being applied of the Simulation Research Group in the Energy and a widespread basis in California. The numberof ongoing or EnvironmentDivision at LawrenceBerkeley Laboratory for completed projects was rather limited. As more designers their advice and assistance in getting the new version of and building owners becomefamiliar with using ice storage DOE2.1E up and running for this study. The industry systems for load management,largely in response to utility survey could only have been completed with the remarkable incentive programs, cold air distribution will also be willingness of our contacts to share their substantial exper- consideredas an increasingly attractive option. In its current tise and experience, along with their valuable time, with us. state of development, however, significant energy-saving Amongthis group, we would like to especially recognize features of CoADtechnology are not being effectively Scot Duncan of Retrofit Originality, Inc., Huntington utilized in installed systems. This situation stems from a Beach, CA; Dave Peters of Southland Industries, Long lack of confidenceon the part of consulting engineers in the Beach, CA; and Ben Sun of Flack + Kurtz, San Francisco, ability of a cold air distribution system to provide accept- CA, for their generous contributions to this study. A able room air distribution, both in terms of comfort and complete list of contacts whoprovided input to our survey

&SHRAETransactions: Research 123 is contained in Baumanet al. (1992). Wewould also like EPRI. 1990a. Proceedings: Ventilation workshop. EPRI acknowledge David Lovberg, undergraduate student of Report CU-6972, Vols. 1 and 2. Electric Power Re- Environmental Resources Engineering, Humboldt State search Institute, Inc., September. University, Arcata, CA, for his assistance in graphics and EPRI. 1990b. Conm~ercial cool storage. EPRI EU.3024. analysis. Electric PowerResearch Institute, Inc. EPRI. 1991a. Proceedings: EPRI workshop on funda~nen- REFERENCES tals and applications of cold air distribution. May2-3, Fort Collins, CO. Anderson, R., V. Hassani, A. Kirkpatrick, K. Knappmil- EPRI. 1991b. Cold air distribution with ice storage. EPRI ler, and D. Hittle. 1991. Visualizing the air flows from CU.2038.7.91. Electric PowerResearch Institute, Inc. cold air ceiling jets. ASHRAEJournal 33 (5): 30-35. Gadgil, A.J., F.S. Bauman, and E.A. Arens. 1991. Cold ASHRAE.1990a. Cool storage and low temperature air air distribution systems for office buildings. CIEE system design and applications. ASHRAETechnical Final Report--Phase I. California Institute for Energy Seminar, May2, San Francisco, CA. Efficiency. ASHRAE. 1990b. ANSI/ASHRAE Standard 113-1990, Hittle, D., and A. Bhansali. 1990. Expected energy use of Method of test#~g for room air diffusion. Atlanta: ice storage and cold air distribution systems in large American Society of Heating, Refrigerating and commercial buildings. EPRIReport CU-6643.Electric PowerResearch Institute, Inc., February. Air-Conditioning Engineers, Inc. Bauman,F.S., T. Borgers, P. LaBerge, and A.J. Gadgil. Huang, J., H. Akbari, L. Rainer, and R. Ritschard. 1991. 1992. Cold air distribution in office buildings: Tech- 481 prototypicai commercial buildings for 20 urban nology assessment for California. Center for Environ- market areas. Lawrence Berkeley Laboratory Report mental Design Research, University of California, No. 29798, April. Berkeley, June. Kirkpatrick, A., T. Malmstrom,K. Knappmiller, D. Hittle, Berglund, L.G. 1991. Comfort benefits for sunm~er air P. Miller, V. Hassani, and R. Anderson. 1991. Use of conditioning with ice storage. ASHRAETransactions low temperature air for cooling of buildings. Proceed- 97(1). ings: Building Simulation Conference, Nice, France. Knebel, D.E. 1990. Off-peak cooling with thermal storage. California Energy Cmnmission. 1992. Nonresidential ASHRAEJournal 32(4): 40-44. manual: For compliance with the 1992 energy e.ffi- ciency standards. Report P400-92-005. California Landry, C.M., and C.D. Noble. 1991. Case study of cost- Energy Conurtission, July. effective low-temperatureair distribution, ice thermal Catanese, D.L. 1991. An energy analysis of low-tempera- storage. ASHRAETransaction 97(1). ture air distribution systems and reduced economizer- MacCracken, C.D. 1990. OPAC--All new for the 1990s. cycle cooling. ASHRAETransactions 97(1). ASHRAEJournal 32(4): 44-46. Dorgan, C.E., and J.S. Elleson. 1987. Field evaluation of Merten, G.P., S.L. Shum, R.H. Sterrett, and W.C. Ra- cold air distribution systems. EPRI Report EM-5447. cine. 1989. Operation and performance of commercial Electric PowerResearch Institute, Inc., October. cool storage systems. EPRI Report CU-6561, Vols. 1 Dorgan, C.E., and J.S. Elleson. 1988. Cold air distribution and 2. Electric PowerResearch Institute, Inc., Septem- design guide. EPRI Report EM-5730. Electric Power ber. Research Institute, Inc., March. Miller, P.L. 1991. Diffuser selection for cold air distribu- Dorgan, C.E., J.S. Elleson, and M.S. Downey. 1990. tion. ASHRAEJournal 33(9): 32-36; Detailed fieM evaluation of a coM air distribution Peters, D., W. Chadwick, and J. Esformes. 1986. Equip- system. EPRI Report CU-6690,Vols. 1 and 2. Electric ment sizing concepts for ice storage systems. Proceed- PowerResearch Institute, Inc., February. ings: International Load ManagementConfetence. Elleson, J.S. 1991. High-quality air conditioning with cold EPRI Report EM-4643, Electric Power Research air distribution. ASHRAETransactions 97(1). Institute, Inc,, June, pp. 46-1-46-29. EPRI. 1987. Seminar proceedings: Commercial cool Piepsch, J. 1991. Water-loop heat pumpsystems: Assess- storage. EPRI Report EM-5454-SR. Electric Power ment study update. EPRI Report CU-7535. Electric Research Institute, Inc., October. PowerResearch Institute, Inc., October.

124 ASHRAETransactions: Research