LA-10062-H

History - . 1.

. . ,...

- - --

Experience Gained from the Space Nuclear Rocket Program (Rorer) '

<./ �,·_ For Reference >..

Not to be taken from this room

n IA\� !{}n� rnr;l(R\o .� LosAlamos National L:at>oratory �� �U@u u 'W:V� LosAlamos,Newtv1ex1CO 87545 An Affirmative'Action/Equal Opportunity Employer

This work was supported by the Air Force Weapons Laboratory, Kirtland Air Force Base, Kirtland, New Mexico.

�-

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or processdisclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. LA-10062-H History

UC-33 Issued: May 1986

Experience Gained from the Space Nuclear Rocket Program (Rover)

Daniel R. Koenig

L

r.

n � f&n 0lf"n'hlt(5\� LosAlamos National Laboratory ���'Q' �Li@U Li Li�� LosAlamos,NewMexico87545 CONTENTS

ABSTRACT...... 1

I. OVERV IEW••••••••••• •••••••••••••••••••••••••••••••••••••••••••••••••• 1

I I. HISTORICAL PERSPECTIVES...... 4

I I I. REACTOR DEVELOPMENT. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 7 A. Kiwi-A...... 7

B. Ki wi-B and NRX...... •• • • •• 9

C. Phoebus•••.•• ••••••••••••••••••••••••••••••••••••••••••••••.••.•• 10

D. Pewee •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 11 E. Huclear Furnace, NF-1. ••••••••••••••••••••••. •••••••••••••••••••• 12

F. Fuel Devel opment •••••••••••••.••••••••••••••••••••••••••••••••••• 13

IV. ENGINE DEVELOPMENT•••.• ••••••••••••••.••.•••••.•••••••••••••••••••••• 16

A. Engine Tests •••••••••••••••••.••••••••••••••••••••••••••••••••••• 16

B. Engine Design Improvements. • • • • • • • • • . • • • . • • • • • • • • • • • • . • • • • • • • • • • • 17

C. NERVA and Smal l Engine Designs•• ••••••••••••••••••••••••••••••••• 18

D. Component Devel opment •••••••••••••••••.••.•••••••••••••.••••••••• 19

E. Testi ng Facilities ••••••••••••••••••••••••••••••••••••••••••••••• 21

V. FUTURE DEVELOPMENTS •••.••.••••••••••••.•••••••••••••••••••••••••••••• 21

A. Fl ight Engine ••••••••••.••••••••••••••••••••••••••••••••••••••••• 21

B. Space Power Generation•..••.•••• ••••••••••••••••••••••••••••••••• 21

C. Dual-Mode Reactors .•••••••••••••••••••••••••••••••••••••••••••••• 22

VI • SUMMARY. • • • . • • • • • • • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • 23

VI I. ACKNOWLEDGMENTS •••••••••••••••••••••••••••••••••••••••••••••••••••••• 24

VII I. SUPPLEMENTAL BIBLIOGRAPHY••••••••••••••••••••••• •••. ••••••••••••••••• 24

REFERENCES ••••••••••••••••••••••••••••••••••••••••••••••••.••••••••••••••••• 26

TAB LES •...••..•..•....•.•...... ••..•...•••....•.•••...... •••..•.....•. 30

FIGURES ...... •.•....•..•....•.....•...... •.•.••.••.•••...... •.•• 32

v EXPERIENCE GAINED FROM THE SPACE NUCLEAR ROCK ET PROGRAM ( ROVER)

by

Daniel R. Koenig

ABSTRACT

In 1955 the United States ini tiated Project Rover to develop a nuclear rocket engine for use in defense systems and space expl oration. As part of that project, Lo s Al amos devel oped a series of reactor designs and hi gh-temperature fuel s. Three high-power reactor series culmi nated in Phoebus, the most powerful reactor ever built, wi th a peak power level of 4080 MW . Two low-power reactors served as test beds for eval uation of high-temperature fuel s and other components for ful l-size nuclear rocket reactors. Los Al amos devel oped and tested several fuel s, incl udi ng a fuel consisting of highly enriched uc2 particles, coated with pyrolytic graphite, and imbedded in a graphite matrix and a composite fuel that formed a conti nuous web of uranium zirconium carbide th roughout the graphi te matrix. The program produced the design of the Smal l Engi ne, wi th a possibl e lifetime of several hours in space.

The Astronuclear Laboratory of the Westi nghouse Electric Corporation, havi ng re­ sponsibil ity for devel oping a prototype reactor based on the Los Al amos design, con­ ducted an extensive and successful test series that culmi nated wi th the NRX-6 reactor test that ran continuously for 60 minutes at design power.

Aeroj et-General Corporati on , prime contractor for devel opment of a compl ete rocket engine, developed two engine test series, the NRX/EST and the XE ' , to eval uate startup, full-power, and shutdown conditions in a variety of altitude and space simu­ lations.

The United States termi nated Project Rover in January 1973 at the poi nt of fl ight engi ne development, but testing had indicated no technol ogical barriers to a success­ ful fl ight system. Conceptual studies al so indicated that nucl ear rocket engine tech­ nol ogy coul d be appl ied to the generation of el ectric power in space.

I. OVERVIEW final mass at earth-escape velocity, as exempli­ In 1955 the Un ited States embarked on a pro­ fied in Fig. 1. In January 1973, after a total gram to develop a nuclear rocket engine. The expenditure of approximately one and a hal f bil­ program , known as Project Rover, wa s initiated at l ion dol lars, the program (al though judged a Los Al amos Nati onal Laboratory, then cal led technical success) was termi nated because of Lo s Al amos Scienti fic Laboratory. The concept to changi ng national priori ties. be pursued was a sol id-core , hy drogen-cool ed The expected appl icati on for nuclear rocket reactor in which the exiti ng gas expanded through engines changed several times duri ng the course a rocket nozzl e and discharged in space. The of the program. At first, nuclear rockets were moti vation for the devel opment of such a rocket considered a potential back-up for intercontinen­ engine was that it could provide about twice the tal bal listi c missile (ICBM) propul sion. Later ( l) specific impul se of the best chemical they were mentioned as a second stage for lunar rockets and , correspondi ngly, a reduction by a fl ight. A more durable possibility was their use factor of 5 in the ratio of ta ke-off ma ss to in manned Mars fli ghts. After pl ans for manned Mars fl ights were abandoned as too ambitious, the Laboratory of the Westinghouse Electri c Corpora­ final possibility advocated for nuclear engines tion ( WA NL) , the pri ncipal subcontractor to was earth orbit-to-orbit transfer. devel op the NERVA nuclear reactor. When analysis showed chemical rockets to be A seri es of reactors and engines was tested more economical for orbi t-to-orbit mi ssions, the at the Nucl ear Rocket Development Station ( NRDS ) need for a nuclear engine for rocket vehicle ap­ in the test site at Jackass Flats in Nevada where plication ( NERVA) evaporated, and the program was major testing facilities were built for the Rover canceled before achievement of a fl ight demon­ program ( Fig. 4) . These incl uded an assembly and stration. The design and the objectives of the disassembly faci lity and two testing facilities NERVA are shown in Fig. 2. Most of the design for the research and engine reactors. The test­ objectives were met or exceeded during the course ; ng program for the nuc 1 ear rocket reactors is of the program. summarized in Fig. 5. It was initi ated wi th a The NERVA in Fig. 2 is attached to a fami ly of research reactors named Kiwi ( for the sli ghtly pressurized, liquid hydrogen tank. fl i ghtl ess bird of New Zeal and) . The program Duri ng operation, the hydrogen is fed to the objectives were fi rst to demonstrate the proof of engine by a turbopump . The hi gh-pressure fl uid principle, then to establ ish the basic reactor first regeneratively cool s the nozzl e and the technol ogy and develop sound design concepts. reactor reflector as shown in Fig. 3, then passes These reactors were the fi rst to demonstrate the through the reactor core. Not shown in Fi g. 3 is use of high-temperature fuel s and to operate wi th a paral lel cool ant circuit to cool the core-sup­ liquid hydrogen. The Kiwi testing seri es culmi­ port tie rods; in the circuit the cool ant is nated with the Kiwi-B4E reactor, which operated heated sufficiently to dri ve the turbopump before for 11.3 mi n at a cool ant exi t temperature above the cool ant rejoins the main fl ow at the reactor 1890 K and for 95 s at 2005 K and a power level inlet. The core contains sol id hexagonal fuel of 940 MW . These tests led to the Nuclear elements banded together by 1 ateral support Reactor Experiment ( NRX) series of devel opmental spri ngs. Longitudinal holes in the fuel elements reactors. Their goal was to demonstrate a spe­ provide cool ant channel s for the hydrogen propel ­ cific impul se of 760 s ( 7450 m/s) for 60 mi n at a lant, which is heated to 2400-2700 K and final ly thrust level of 245 kN (55 000 lb) in a 1100-MW expanded through a thrust nozzle. Rotating drums reactor. These objecti ves were exceeded in the in the reflector containing neutron absorber last test of that series, the NRX-6 reactor, material provide reactivity control of the whi ch operated for 62 min at 1100 MW and a tem­ reactor, which has an epithermal neutron energy perature of 2200 K, with only an $0 .11 reacti vity spectrum. 1 oss. The aim of the Rover program , besides Another series of research reactors, cal led designing and demonstrati ng a practical rocket Phoebus, was devel oped wi th objectives to in­ engine, was to achieve the hi ghest-possibl e pro­ crease the specific impul se to 825 s, increase pel l ant temperature ( specific impul se is propor­ the power density by 50%, and increase the power tional to the square root of the temperature) for level to the range of 4000-5000 MW . These capa­ the duration of potenti a 1 mi ssions ( severa 1 bilities were demonstrated in the Phoebus-lB and hours). Thi s goal impl ied a strong technol ogy Phoebus-2A reactors. The latter, the most power­ devel opment program in reactor fuel s. ful reactor ever built, ran for 12 min at 4000 MW Los Al amos Nati onal Laboratory was given the and reached a peak power of 4080 MW . The 1 as t role of establ ishing a basic reactor design and two families of research reactors, Pewee and the of leadi ng the fuel s devel opment effort. Other Nuclear Furnace ( NF ), were tested only once each. key pl ayers were the Aerojet-General Corporation, They were lower-power reactors, 500 and 44 MW the prime contractor to devel op the complete respectively, designed primarily as test beds to rocket engine system, and the Astronucl ear demonstrate the capabilities of higher-temperature

2 fuel el ements. Pewee-1 ran for 40 mi n at 2555 K, For compari son, the mass of several Rover and NF-1 operated for 109 min at an average reactors is pl otted versus power in level in cool ant exi t temperature of 2450 K. Fig. 10 . An engine devel opment test program was part It was al so recognized that the design of a of the technol ogy demonstration. Its objecti ves nuclear rocket engine could be al tered so as to were to test nonnuclear system components, deter­ provide continuous station-keeping power for the mine system characteri stics duri ng startup, full­ missions. Design studies for such dual -mode power, and shutdown condi tions, eval uate control rocket systems were initiated in 1971-72 where concepts, and qual ify the engine test- stand oper­ one mode was the normal propul sion and the ations in a downwa rd-firi ng configuration wi th second, a closed-l oop, low-power electrical (4 5 simulated al ti tu de and space conditions. These mode. • l objectives were met or exceeded in the Nuclear The Rover program was termi nated in January Reactor Experiment/Engine System Test (N RX/EST) 1973 at the poi nt of fl ight engine development. and Experimental Engine (XE) programs. A proto­ For a fl ight system , it would be necessary to ty pe fl i ght engine system, XE, consisti ng of a veri fy the fl i ght reactor and engine design and fli ght-type reactor wi th nonnuclear fl ight compo­ to perform life and reproducibility testi ng . But nents, was tested in a space-simulated environ­ there are no apparent barriers to a successful ment, performing some 28 starts and restarts. nuclear rocket. A chronology of the major tests conducted The technol ogy devel oped during the Rover during the Rover program is shown in Fig. 6. program is directly appl icable to the generation The major emphasis of the reactor develop­ of el ectrical power in space, especial ly large ment program was to increase the reactor cool ant (mul timegawatt) bursts of electrical power. For exi t temperature because the specific impul se is an open-loop converter system, one would simply proportional to the square root of that tempera­ replace the rocket nozzl e wi th a power conversion ture and to increase the operati ng time of the sy stem . Some redesign of the core parameters reactor. The success of this part of the program would be i nvo1 ved because the power converter, is illustrated in Fig. 7. Cool ant exit tempera­ unless it were a magnetohydrodynamic (MHD ) sy s­ tures above 2500 K and operating time over 2 h tem , could not operate at the high temperature of were demonstrated . The cumul ati ve time-at-power the Rover reactors. The startup time for such a for the enti re Rover program is shown in Fi g. 8. power pl ant woul d be limited in part by the The major performances achieved duri ng the pro­ al lowabl e rate of reactor temperature change , gram are summarized in Fig. 9. about 83 K/s. However, a more severe 1 imitation The Rover program was terminated before al 1 is in the propel lant feed system, which requires of the NERVA objectives could be demonstrated, in approximately 60 s to overcome pump cavitati on parti cular, before showi ng that an engine coul d before chi 11-down and to chill various parts of be operated for 10 h with up to 60 starting the engine. In addition, there would be time cycles wi th a rel iability of 0.995. 1 imitations imposed by the power conversion sy s­ Toward the end of the program, emphasis was tem. bei ng pl aced on smal ler engines for the orbi ta l A cl osed-1 oop system would re qui re further transfer mi ssion. A comprehensive design study redesign to incorporate the gas circulators, and wa s done on a 367-MW , 72-kN (16 000-1 b) engi ne, the core design woul d have to be adj usted for the (2 3) the so-cal led Smal l Engine. • The total higher inlet temperatures. mass of this engine was 2550 kg, and its overal l As concerns dual electrical -power modes (a length was 3.1 m with the nozzl e skirt in a conti nuous, low-power mode and a short-tenn, fol ded position. The engine , together with a high-power mode) much of the technol ogy and many hy drogen tank containing nearly 13 000 kg of pro­ studies devel oped under the Rover program are pel lant, coul d be carri ed on the space shuttl e. appl icabl e if the high-power converter is to be a gas system.

3 II. HISTORICAL PERSPECTIVES 1959 Thi s chapter summarizes the major events in The fi rst reactor test, Kiwi-A, is �- the histor y of the Rover program. Information conducted successfully at the Nevada Test (7 ,8) for the test summaries was obtained primarily S'ti e. The reactor operated for 5 min at from Refs. 6-8. 70 MW and provided important design and material s 1945-1954 information. The fuel was hot enough (2683 K) to In 1945, at the suggesti on of Theodore von mel t carbide fuel particles. Vi brations in the Karman, the USAF Scientific Advisory Board core produced structural damage in the graphite studied the use of nuclear propul sion for rocket elements. The reactor employed uncoated, uo - 2 systems. However, because of a lack of a clear loaded, plate-type fuel elements and was cooled need for such systems, the shortage of fission­ with gaseous hy drogen. The reactor core con­ abl e material s, and the technical difficulties of tained a central isl and of D o to reduce the 2 developing such a propul sion system, no action amount of fissionabl e material required for was recommended. Nevertheless, paper studies of cri tical ity. Control rods were located in this nucl ear rocket systems were performed duri ng this island. (9,10 l period . 1960 1954 Kiwi-A' is tested for nearly 6 min �- Von Karman again suggests that, in view of at 85 MW to demonstrate an improved fuel-el ement the need for ICBMs and the good supply of fi s­ desi gn. The reactor used short, cyl i ndri cal , s ionabl e material , the Scienti fic Advi sory Board uo -l oaded fuel elements contained in graphite 2 reconsider nuclear propul sion. modul es. The fuel el ement had four axial cool ant 1955 channel s coated with NbC by a chemi cal vapor October 18. In a final report, an ad hoc deposition (CVD) process. convnittee of the Scienti fic Advisory Board recom­ August 29. A Memorandum of Understanding mends that because of the potential ly high spe­ defi ni ng NASA and AEC responsibilities and estab- cific impulses wi thin the realm of immediate 1 ishing a joint nuclear program office, the Space achievement from the nuclear rocket, substantial Nuclear Propul sion Office, is signed. development work should be started on the nuclear October 10. Kiwi-A3 reactor is operated in rocket system. excess of 5 min at 100 MW . The fuel was similar November 2. The nuclear rocket propul sion to that used in the previous test. As wi th the program is established as Project Rover at earli er tests, core structural damage occurred, Los Al amos and Lawrence Livermore Scienti fic indicati ng that tensile loads on graphite Laboratories. Several conceptual nuclear rocket structures should be avoided . This experiment desi gns already had been under study . (ll) But was the thi rd and last in the Kiwi-A seri es of the concept chosen to be pursued was a sol id­ proof-of-pri nciple tests conducted by Los Al amos. core, hydrogen-cooled reactor that would expand The test seri es demonstrated that thi s type of gas through a rocket nozzle. hi gh-power-density reactor could be control led 1957 and could heat hy drogen gas to high temperatures. March 18. The Atomic Energy Commi ssion 1961 ( AEC) decides to phase Li vermore out of the pro­ June-July. Industri al contractors, Aerojet­ gram as a resul t of budget restrictions and a General for the rocket engine and Westi nghouse Department of Defense reconvnendation for a more Electri c Corporation for the reactor, are se­ moderate level of support. The latter stenvned lected to perform the nuclear rocket devel op­ from the earl ier-than-anticipated availability of ment phase. The reactor in-flight tests (RIFT) chemical ICBMs, which reduced the urgency for program was initi ated at the Lockheed Corporation. devel opment of nucl ear propul sion.

4 December 7. Kiwi-BlA reactor, fi rst of a and find sol utions for the severe structural dam­ new series, is tested by Los Al amo s. Kiwi-B age that was observed in the previous reactor reactors were designed for 1100 MW and used tests. The col d-fl ow designati on referred to refl ector control and a regeneratively cool ed reactor tests that contained fuel elements iden­ nozzle. Thi s test was the last to be run wi th tical to the power reactors except that they had gaseous hy drogen cool ant. After 30 s of opera­ no fi ssionable material and, therefore, produced tion, a hy drogen leak in the nozzle and the pres­ no power. These tests were performed wi th gas­ sure vessel interface forced termi nation of the eous nitrogen, hel ium, and hy drogen, and they run. The planned maximum power of 300 MW was demonstrated that the structural core damage was achieved, as limited by the capability of the due to fl ow-i nduced vibrations. Based on results nozzl e with gaseous hy drogen cool ant. of these tests and analyses, design changes that The core consisted of cyl i ndrical uo - were compl etely successful in el iminati ng core 2 loaded fuel elements, about 66 cm long, havi ng vi brations were made. seven axial cool ant channel s NbC coated by a �· Kiwi-B4D, the fi rst test at full tube-claddi ng process. The fuel elements were design power, is carried out wi th no indication contained in graphi te modul es. of core vibration. Thi s was al so the fi rst time 1962 a completely automati c start was accomplrshed for September l. Kiwi -BlB reactor test is the a nuclear rocket reactor. The test was termi­ first to operate wi th liquid hy drogen. The test nated after 60 s at ful l power when several met its primary objective of demonstrati ng the nozzl e tubes ruptured. The core consisted of ability of the sy stem to start up and run using ful l-length, uo -l oaded, 19-hole, hexagonal 2 liquid hy drogen. Fol lowing a smooth, stabl e fuel el ements with bores NbC coated by the tube­ start, the run was termi nated after a few seconds cladding process. at 900 MW when porti ons of several fuel el ements August 28. Kiwi-B4E , the eighth and final were ejected from the reactor. The core empl oyed Kiwi reactor, is tested by Los Al amos. The the same ty pe of fuel as Kiwi-BlA. reactor was operated for more than 12 mi n, of November 30 . Kiwi-B4A, the first design which 8 min were at nearly ful l power. The intended as a prototype fl ight reactor, is reactor operation was smooth and stabl e. Its tested. The power run was terminated at about duration was limi ted by the available liquid the 50% level when bright fl ashes in the exhaust hydrogen storage capacity. On September 10 , the ( caused by ejection of core materi al ) occurred reactor was restarted and ran at nearly ful l wi th increasing frequency . Subsequently, i nten­ power for 2.5 mi n. Thi s was the first demonstra­ sive analyses and component testi ng were con­ tion of the reactor' s ability to restart . ducted to determi ne the cause of the core dam­ The core consisted of ful l-l ength, 19-hol e, age. The core consisted for the fi rst time of hexagonal fuel el ements, loaded for the first ful l-l ength, extruded, 19-hol e, hexagonal fuel time wi th uc particles. The bores were NbC 2 elements, loaded still wi th uo • The cool ant coated by the tube-cladding process. 2 channel s were NbC coated by the tube-claddi ng September. Measurements, at zero power, of process. the neutroni c interacti on of two Kiwi reactors 1963 positioned adj acent to each other veri fy that December. The RIFT program is cancelled. there is littl e interaction and that, from a It was decided to revise the nuclear rocket pro­ nuclear standpoint, nuclear rocket engines may be gram to place emphasis on the devel opment of operated in clusters similar to chemical engines. ground- based systems and defer the devel opment of September 24. NRX-A2 is the fi rst NERVA fl ight systems. reactor tested at ful l power by Westinghouse 7) 1963-1964 Electric. ( The reactor operated in the range Several col d-fl ow tests of Kiwi-B-type of half to ful l power ( 1100 MW) for about 5 min, reactors are carried out to determi ne the cause a time limi ted by the available hydrogen gas

5 supply. The test was successful and demonstrated operated agai n at ful l power (-1100 MW) on an equival ent vacuum specific impul se of 760 s. June 23 for 14.5 min to bri ng the total operating The reactor was successful ly restarted on time at ful l power to hal f an hour. The liquid October 15 to investigate the margin of control hydrogen capacity of the test facility was not in the low-fl ow, low-power regime. sufficient to permi t 30 min of conti nuous opera­ 1965 tion at design power. January 12. Kiwi-TNT (Transient Nuclear 1967 Test) is successfully completed by Los Al amos. February 23. Phoebus-lB is operated for In this fl ight safety test, a Kiwi -B-type reactor 45 min of which 30 min, the maximum time planned, was del iberately destroyed by pl aci ng it on a were at design power of 1500 MW . The primary fast excursion to confirm the analytical model s purpose of the test was to detennine how the of the reactor behavior during a power excur­ higher-power operation affected the reactor. The (12,13) sion. fuel was the same as that used in Phoebus-lA . Apri l 23. NRX-A3 reactor is operated for December 15. NRX-A6 test exceeds the NERVA about 8 min with about 3.5 min at ful l power. design goal of 60 mi n at 1100 MW in a single run. The test was terminated by a spurious tri p from 1968 the turbine overspeed circuit. The reactor was June 26. Phoebus-2A, the most powerful resta rted on May 20 and operated at full power nuclear rocket reactor ever built, runs for for over 13 min. It was restarted again on 12.5 min above 4000 MW. The duration of the test May 28 and operated for 45 mi n in the low- to was determined by the available cool ant supply. medium-power range to expl ore the limits of the Designed for 5000 MW, the te st was limited to 80% reactor operati ng map . The total operati ng time of ful l power because the al umi num segments of of the reactor was 66 min wi th over 16.5 min at the pressure vessel cl amp band overheated pre­ ful l power. maturely. The reactor was restarted on July 18 June 25. The aims of Phoebus-lA, the first and operated at intermediate power level s. test of a new class of reactors, were to increase December 4. Pewee reactor testi ng is suc­ the specific impul se, the power density in the cessful ly completed. Pewee, designed to be a core, and the power level . The test is run suc­ small test-bed reactor, set records in power den­ cessfully at ful l power ( 1090 MW ) and core exit sity and temperature by operating at 503 MW for temperature ( 2370 K) for 10 .5 mi n. The reactor 40 min at a coolant exit temperature of 2550 K 3 was subsequently damaged when the facility's and a core average power density of 2340 MW/m • liquid hy drogen supply was exhausted . Thi s Thi s power density was 50 % greater than that re­ course of events was in no way rel ated to any quired for the 1500-MW NERVA reactor. The peak 3 defect in the reactor. The core consisted of power density in the fuel was 5200 MW/m . The ful l-length , 19-hole, hexagonal fuel elements cool ant exit temperature corresponds to a vacuum loaded with coated uc particles. The bores specific impul se of 845 s. The core contained 2 were NbC clad by the chemical vapor deposition the same ty pe of fuel elements as Phoebus-lA. (CVD) process. Ma rch. XE' , the first down-firi ng prototype 1966 nuclear rocket engine , is successfully operated February 3 to March 25. The NRX/EST , the at 1100 MW . The reactor was operated at various

first NERVA breadboard power pl ant, is operated power 1 evel s on different days for a total of duri ng 5 di fferent days for a total of 1 h and 115 mi n of power operation that included 28 re­

50 min, of which 28 min were at ful l power starts. Individual test times were 1 imited by (1100-1200 MW ). These times were by far the the faci lity's water storage system, which coul d greatest achieved by a single nuclear rocket not support operations longer than about 10 min reactor as of that date. at full reactor power. This test series was a June 8. NRX-A5 is operated successfully at significant mi 1 es tone in the nuclear rocket ful l power for 15.5 mi n. It was restarted and

6 program and demonstrated the feasibility of the engi ne longevity (initial ly 1 h, then 10 h) , it NERVA concept. was necessary to minimize hydrogen corrosion of In this year, the production of the chemi cal the fuel and breakage of the core from vi bra­ rocket Saturn V was suspended. It would have t i ona l and thermal stress. Only a few materi al s, (ll) been the prime launch vehicle for NERVA. including the refractory metal s and graph­ 1972 ite, are suitable for use in reactors designed to June 1. NF -1 test is successful ly accom­ run at high temperatures (up to 2700-2800 K). pl i shed. The reactor was operated for 109 mi n at Graphite was sel ected because in contrast to the the ful l design power of 44 MW, demonstrating metal s, it is not a strong neutron absorber, and fuel performance at a cool ant exit temperature to it does moderate neutrons leadi ng to a reactor 2500 K and a near-record peak power density in with a smal ler cri ti cal mass of enriched ura­ 3 the foe l of 4500-5000 MW/m . NF-1 was designed nium. Graphite has excel lent high-temperature wi th a remotely repl aceabl e core in a reusabl e strength, but its great disadvantage is that it test bed, intended as an inexpensive approach to reacts wi th hot hydrogen to form gaseous hydro­ mul tiple testing of advanced fuel material s and carbons and, unless it is protected, it rapidly structures. Another special feature of th is test erodes. Consequently, one of the greatest chal- series was eval uation of a reactor effl uent 1 enges of the nuclear rocket program was to cleanup system. The sy stem performed as expected devel op fuel elements of adequate lifetime in in removing radioacti ve contaminants from the hi gh-pressure hot hy drogen. effl uent reactor gas. The designers of the nuclear rocket engine Two types of fuel elements, (UC-ZrC)C "com­ had to consider many factors such as neutronic posite" fuel and the pure (U,Zr)C carbide fuel , and heat-removal requirements; high mechanical were tested in NF-1. loadi ngs; and the complex probl ems of startup, 1973 control , shutdown, and safety. To permit prel im­ January. The Rover nuclear rocket program inary eval uation of the neutronic calculations, a is terminated . It was judged a technical mockup or cri tical assembly of each reactor ty pe, success, but changi ng national priori ti es known as Honeycomb, was built as shown in (l4,l5 resul ted in the deci sion to cancel the program. Fig. 12. l It consisted of graphite slabs, enriched uranium foils, pl astic to simu­ III. REACTOR DEVELOPMENT late the propel lant, and beryllium-refl ector The concept of a nuclear rocket engine is blocks. Later, during construction of each new simpl e. As shown schematical ly in Fig. 11, it ty pe of reactor, a more exact mockup of the final consists of a cryogenic pro pel l ant tank, a reactor, known as Zepo (Zero Power) , was built turbopump to feed the propell ant through the (Fig. 13) usi ng actual fuel el ements to determi ne system, a nuclear reactor to heat the pro pel l ant the sy stem's neutroni cs. Such testing facilities to the highest temperature possible, and a thrust were built at Los Al amos and WANL. The actual nozzl e through which the hot gas is expanded. reactor and engine tests were carried out at the The propel lant is hydrogen because a gas wi th the NRDS . lowest-possibl e mol ecul ar wei ght is most desirabl e. A. Kiwi-A The reactor design goal s presented a real The fi rst reactor tested under the Rover chall enge in reactor design and materi al s program was named Kiwi-A. It was designed and devel opment. The core exi t temperature of the built by Los Alamos as were al l of the Kiwi (l6 cool ant had to be maximized to achieve the seri es of reactors. The reactor design, l as highest-possibl e specific impul se. The core shown in Fig. 14, was intended to produce about power density al so had to be maximi zed to 100 MW of power. It wa s, in fact, tested for minimize reactor mass. To achi eve a practical 5 min at 70 MW . The Kiwi-A core consisted of an

7 annular stack of four axial layers of fl at-pl ate , shattered and was ejected out of the nozzle along graphite fuel el ements loaded wi th highly en­ with the graphite wool between the center isl and riched uo particles. The fuel el ements were and the core. The functions of this plate were 2 retained and supported in graphite structures to contain the carbon wool insulation and to called whims. The whims, shown in Fig. 15, were serve as a gas seal that prevented gas from by­ wheell i ke structures wi th 12 wedge-shaped boxes passing the annular core into the central of fuel plates fitted between their spokes, each region. Failure of the closure pl ate all owed a box containing 20 fuel plates. A fifth whim con­ lot of gas to fl ow radial ly inward through sl ots tained unloaded fuel pl ates and served as an end in the inside wal l of the whims (Fig. 15) and refl ector for the outlet end of the core. The into the central part of the core , thereby by­ inlet and radial reflectors consi sted of several passing the power-producing region of the core . conti nuous graphite cyli nders. Power flattening This by passed gas was not heated to ful l tempera­ was achi eved by varying the fuel loading. The ture. Because the test conditions demanded a core was separated from the radi al refl ector by a prescri bed average gas outl et temperature , it carbon wool region. The hole in the center of fol l ows that the gas that did pass through the the core contained a "o o isl and ," the functi on acti ve core had to be heated to a higher tempera­ 2 of which was to moderate neutrons, thereby ture. The high fuel temperatures that resulted 235 reducing the critical mass of u, and al so to led to mel ti ng of the uc fuel and high erosion 2 provide a low-temperature, low-pressure container of the graphite fuel pl ates. for the reactor control rods that were cooled by For the next two reactors, the Kiwi-A core ci rcul ating o o. The enti re reactor was en­ design was modi fied to replace the whims and fuel 2 cased in an aluminum pressure shell to which a pl ates wi th graphi te modules containing cylindri­ 118) li ght-water-cooled nickel nozzle was attached . cal fuel elements as shown in Figs. 16 and The nozzl e was designed for choked-fl ow outlet 17. Thi s modification entai led a compl ete change conditions for the core cool ant (that is, sonic in the fuel fabrication process from pressi ng and flow at the th roat of the nozzle). mol di ng to a new graphi te extrusion process. The The hydrogen cool ant fl ow through the fuel cyl i nders were segmented in short lengths reactor is as follows. Coolant is deli vered to and six of them were stacked on top of each other the plenum near the top of the pressure vessel . in each hole of the graphite modules to make up a The gas then flows axial ly downward through holes compl ete fuel modul e. The fuel cyl i nders con­ in the refl ector segments and into the plenum at tai ned four axial cool ant channel hol es that were the bottom of the pressure vessel where the fl ow coated by a CVO process with NbC to reduce hydro­ reverses, passing upward through hol es in the gen corrosion of the graphite. This modi fied inlet refl ector. The gas now conti nues upward core configuration was tested twice for 5 to between the fuel plates of each whim, through the 6 min in the power range 85-100 MW in the lg) 20 unloaded plates of the top whim, and out through Kiwi-A'( and Kiwi-A3 ( l tests. Fracture the nozzle . of fuel modules was experienced in both of these The Kiwi-A experiment was a first step tests, but the general appearance of the fuel toward demonstrating the feasibility of a high­ elements after each test was quite good even temperature, gas-cooled reactor for nuclear pro­ though several elements showed bl i steri ng and pul sion, and as such it provided important severe corrosion. 1 8 17) 1 21 reactor design and materials information. • The Kiwi-A series of tests l demon­ Much higher fuel temperatures ( up to 2900 K) strated that hydrogen gas could be heated in a than anticipated were reached during the test nuclear reactor to the temperatures requi red for because early in the run the graphite closure space propul sion and that such a reactor coul d plate , located just above the o o isl and, indeed be control led. 2

8 B. Kiwi-B and NRX cool ant channel , sized to provide approximately Built on the experience gained wi th the the same exit gas temperature for al l channel s. Kiwi-A reactors, a new reactor design evol ved The core , which contained 182 kg of uranium that more nearly resembled what would be needed ( enri chment 0.9315) , was surrounded by a graphite for a fl i ght engine. The Kiwi-B te st series was cyl inder about 46 mm thick and a beryll ium 22} initi ated wi th the Kiwi-BlA ( test in refl ector 114 mm thick. Twel ve rotating drums

December 1961 and culmi nated 2 years and 8 months 1 ocated in the reflector contained segments of later with the successful Kiwi-B4E test accom­ boron carbide neutron absorber that could be plished in August 1964. During this test series, swung toward or away from the core to provide improvements were made with the extruded fuel reactivity control of the reactor. The entire desi gn and the protecti ve NbC-coating technol ogy . reactor was encased in an aluminum pressure Severe structural damage to the core was experi­ vessel to which the exhaust nozzle was attached. enced wi th the second test in the series The pressure vessel was approximately 21 nm ( 23> ( Kiwi-B1B} when the hot ends of seven fuel thick, 1.9 m in length, and 1.3 m in outer diam­ modules were ejected from the core duri ng the eter. transient ri se to ful 1 power. It took several The fl ow of hy drogen cool ant through the subsequent ful l -power tests, in parti cular reactor was as fol lows (Fig. 3) : liquid hy drogen ( 24 25} Kiwi-B4A, Kiwi-B4D , • and several col dflow entered the aft end of the nozzl e to cool the tests to di scover and confi rm that core damage nozzle wal l before entering the reflector plenum. was caused by flow- i nduced vibrations and to From this plenum the hy drogen traveled forward demonstrate, after design modi fications were through the refl ector and control drums, al so appl ied, that a stabl e design had been achieved. cool ing the pressure vessel . It entered a pl enum This successful reactor configuration again before fl owi ng forward through the outer ( 26-30 l (6} ( Kiwi-B4E ) led to the NRX series of region of the simul ated shiel d. The fl ow dis­ NERVA devel opmental reactors from which emerged charged from the shiel d and entered the plenum 31 32} the final NRX-6 design ( • shown earl ier in region between the shiel d and the dome of the Fig . 3. The reactor was designed for a nomi nal pressure vessel . Here the fl ow reversed, and the power of 1100 MW . It was al l graphite moderated, gas fl owed aft through the inner region of the and it had an epithermal neutron spectrum. The shiel d, then through a fi ne mesh screen and the extruded graphite fuel elements were hexagonal core support pl ate. Most of the cool ant then and contained 19 cool ing channel s. The channel flowed through the channel s in the fuel el ements wal ls and the exterior surfaces of the fuel el e­ where it was heated to a high temperature. A ment were coated wi th NbC to reduce hy drogen cor­ smal l part of the flow cool ed the peri phery rosion. The fuel was assembl ed in clusters of region between the core and the beryllium six el ements supported by a tie rod in the cen­ reflector, and some cool ant al so fl owed past the tral location as shown in Fi g. 18. The tie rod tie rods in the core. These coolant flows were was attached to an aluminum support pl ate at the mixed in the nozzl e chani>er at the reactor exit col d end of the reactor. Irregularly shaped before expul sion through the nozzle. clusters were fitted on the core peri phery to One aim of the devel opmental series of tests obtain a cylindrical core configuration. The conducted by Westi nghouse Electri c was to reduce core dimensions were 1.32 m in length and approx­ the fraction of coolant fl ow that did not pass imately 0.89 m in diameter. Lateral support for through the fuel in order to obtain the highest­ the core wa s obtained wi th a spri ng and a ri ng­ possible gas temperature in the nozzle chamber. seal arrangement as described in Fi g. 19. Power Thi s aim was achieved by applying design modifi- fl attening was achieved by varying the fuel cati ons described below for the Phoebus loadi ng , and the cool ant fl ow di stribution was reactors. The duration of ful l -power runs was colltrol led by orifices in the inlet end of each gradual ly increased with each NRX reactor unti 1

9 4o-42 the test in December 1967 in wh ich the NRX-A6 ran The Phoebus-2A reactor( l incorporated conti nuously for 60 min at 1125 MW with an exit al l of the features mentioned above. The core, cool ant temperature at or above 2280 K, corre­ which conta i ned about 300 kg of uranium, con­ spondi n9 to a vacuum specific impulse of sisted of 4068 fuel ed el ements pl us 721 regenera­ ( 33-35l 730 s. The test duration and power tively cool ed support elements. The active core level exceeded the NERVA design goal s at that dimensions were 1.39 m in diameter and 1.32 m in time. length. The 19-hole fuel elements were similar in geometry and had the same external dimensions C. Phoebus as those of earl ier reactors ( Kiwi-B4E to Fo11 owi ng the successful performance of the NRX-A6) , but the coolant channel diameter was

Kiwi-B4E reactor, the Los Al amos Scientific Labo­ increased from 2.54 mm to 2.79 nm. The channel s ratory devoted its attention to a new class of were coated wi th NbC of tapered th ickness and reactors similar in design to Kiwi-B but havi ng were overcoated wi th a 1ay er of Mo to reduce greater cool ant exit temperatures, power densi­ hy drogen corrosion of the graphite. As before, ties, and power level s. Power density was to be the fuel was assembl ed in clusters of seven el e­ increased mainly by enlarging the di ameter of the ments where the central element was unloaded coo1 ant fl ow channe 1 s in the fue1 e1 ements from graphite containi ng the ti e-tube axial support 2.54 mm to 2.79 mm to reduce thermal stress and assembly. Details of construction and cool ant core pressure drop . The temperature increase was fl ow paths for the regeneratively cool ed tie to be obtained by some mi nor design modi fications tubes are shown in Fig. 21. The core- refl ector in the fuel elements but mostly by reducing the interface had an alumi num interface cylinder amount of cool ant flow that by passed the core. assembly that separated the high-pressure

The cool ant fl ow along the core peri phery wa s refl ector system region from the 1 ower-pressure reduced, and the single-pass cool ing of the metal core periphery, transmitted the axial pressure­ tie rods in the core was reduced and eventua11 y drop 1 oad to the nozzl e, and contained the sea1 changed to two-pass regenerative cool ing by rings. This assembly was cool ed by refl ector replacing the tie rods with tie tubes. These bypass cool ant. The 203-mm-thi ck beryllium- tubes were cooled by diverting 10% of the fl ow to refl ector assembly contained 18 control drums the core support and returning thi s fl ow to the rather than 12 as used for the earl ier smal l er main core cool ant flow at the inlet of the fuel reactors. The reactor was contai ned in an alumi­ elements. These cool ant fl ow modifications num pressure vessel 2.54 nm thick with an out­ greatly reduced the mixing of col d cool ant wi th side diameter of 2.07 m and an approximate length the core exi t gas in the nozzl e chamber. The ( excludi ng the nozzl e) of 2.5 m. Reactor mass power level was increased simply by increasing incl udi ng the pressure vessel was 9300 kg . A the number of fuel elements in the core . two-dimensional model of the reactor that was 36 (37-39 Three tests, Phoebus-lA, ( l -lB , l used in neutronic calcul ations is shown in and -2A, were carried out in this series. The Fig. 22. Reactor control was obtained by first two tests were essential ly vehicl es for adjusting two basic parameters, namely, the cool­ experiments leadi ng to the Phoebus-2A design. ant fl ow rate and the control -drum position. Phoebus-2A (Fig. 20 ) designed for 5000 MW was the The successful ful l-power test of Phoebus-2A most powerful nuclear rocket reactor ever built. took pl ace in July 1968 and lasted for 12.5 mi n,

It was intended original ly to be a prototype a time 1 imited by the available hy drogen cool ant optimum-thrust nucl ear propul sion engine for supply ( coolant, dri ven by two Rocketdyne Mark-25 ambitious planetary mi ssions. The reactor had a turbopumps operating in paral lel, flowed through nominal thrust of 1110 kN (250 000 lbf) and a the reactor at a rate of 120 kg/sl . The maximum specific impul se of 840 s, corresponding to a power level reached duri ng the test was 4080 MW . nozzle chamber temperature of 2500 K. The reactor coul d not be operated up to the design power level of 5000 MW because part of the

10 al umi num pressure vessel assembly was overheati ng toward providi ng a real istic nuclear, thermal , prematurely as a result of unexpected poor ther­ and structural environment for the fuel elements mal contact wi th an LH -cool ed clamp ri ng . The in a core contai ning one-fourth the number of 2 maximum fuel-element exi t-gas temperature at­ el ements in these reactors, and one-tenth the tained was 2310 K, and the maximum nozzle chamber number of el ements in Phoebus-2A. temperature, nearly as high, was 2260 K. Thi s Mo st of the basic design features of Pewee smal l temperature difference is an indication of were similar to those of the precedi ng reactors. the effectiveness of the measures taken to reduce The fuel el ements were simi lar, and they were mixing of col d cool ant with the core exit gas. held in pl ace by support elements; the control At design power, the core power density woul d drums were incorporated in the beryllium radial have been nearly twice that of the Ki wi-B refl ector; and liquid hy drogen was used as the reactors. working fl uid. There were , however, significant The Phoebus-2A test revealed some neutronic differences that distinguished Pewee from earl ier discrepancies when compared wi th pretest calcul a- reactors. The core diameter was reduced from tions and zero-power cri tical ity experi- 1400 mm in Phoebus-2A to 533 mm to reduce the 43,44> ments. ( Specifical ly, the ful l -scale number of fuel el ements. Sufficient reactivity reactor test resul ted in larger col d-to-hot wi th the smal ler core was obtained by inserti ng changes in reactivity than had been predicted. sleeves of zirconium hy dri de around the tie rods The anomal ies were eventual ly resol ved and at­ in the support el ements as shown in Fig. 23. The tributed mainly to the combined effect of low hy drogenous materi al moderated the core neutrons beryllium-refl ector temperatures and the presence and reduced the critical mass of uranium in the of cold high-density hy drogen in the al uminum core to 36.4 kg . The ratio of support el ements interface cylinder and in the refl ector. The to fuel el ements was increased from 1:6 to 1:3, re sult was to produce a large negative change in as illustrated in Fig. 24, to increase the amount reactivity and a substanti al reduction in con­ of ZrH moderator to the desi red level . Thi s x trol-drum wo rth. Nei ther of these effects had el iminated the traditional cl usters-of- seven con­ been correctly accounted for in pretest analysis. cept; each fuel element was supported redundantly The successful concl usion of the Phoebus-2A by two pedestal s. The core contained 402 fuel tests was a mil es tone in nuclear rocket tech­ elements and 132 support el ements. Because Pewee nology because of the high-power capability that was designed as a test bed for fuel el ements, no the test demonstrated. Some probl ems remained, attempt was made to maximize the specific impul se particularly in the area of fuel longevity and by maintaini ng a high temperature in the nozzl e temperature capability, but the feasibility of chamber; the support-el ement cool ant was di s­ practical nuclear space propulsion had been con­ charged directly into the chamber. This di s­ vincingly demonstrated by this stage of the Rover charge reduced the nozzl e temperature signif­ program. Phoebus-2A was the last reactor design icantly because the hydri de moderator required a in direct support of the NERVA devel opment that larg er amount of cool ant than a graphite support was tested by Los Al amos. Two smaller reactor element wi thout moderator and because a con­ designs were subsequently tested by Lo s Al amos, servatively low coolant di scharge temperature was but they were primari ly te st beds for improvi ng chosen. the fuel technol ogy. The smal l size of Pewee required a thick (205-mm) beryl lium refl ector that was built in D. Pewee two concentric parts as shown in Fig. 25. The ( 45l Pewee was a small reactor designed to inner part consisted of beryl lium ri ngs that re­ serve as a test bed for the eval uation of ful l­ pl aced the interface cylinder of previous size Phoebus and NRX fuel el ements and other designs. The outer part was made from Phoebus-1- components. The general design was directed type sectors and contained nine control drums.

11 The mass of the Pewee reactor, includi ng the having a low fuel inventory. It was never meant aluminum pressure vessel, was 2570 kg. to be a candidate concept for a rocket engine. The Pewee test series conducted in The reactor, described in Fi gs. 26 and 27, con­ November-December 1968 was successful, and it set sisted of two parts: a permanent, reusabl e por­ several records for nuclear rocket reactors. The tion that incl uded the refl ector and external primary objecti ve was to demonstrate the capabi l­ structure ; and a temporary, removable portion ity of this new reactor as a fuel -el ement test that consisted of the core assembly and asso­ bed. Pewee ran for a total of 192 min at power ciated components. level s above 1 MW on two separate days. The A major objective of th is design was to have ful l -power test consisted of two 20-min hol ds at a reusable test device that woul d reduce both the design power (503 MW) and an average fuel -element time between reactor tests and the cost of test­ exi t-gas temperature of 2550 K. This temperature ing. After completion of a te st series, the core was the highest achi eved in the Rover program. assembly would be removed and disassembled for It corresponds to a vacuum specific impul se of exami nation, whereas the permanent structure 845 s, a level in excess of the design goal set would be retai ned for use wi th a new core. for the NERVA. The peak fuel temperature al so Actual ly, the NF-1 was tested only once before reached a record level of 2750 K. The average termination of the program, but the removabl e 3 power density in the core was 2340 MW/m , al so feature of the design was demonstrated. a record high and greater than that requi red for The NF -1 core was a 34-cm-di ameter by the NERVA . The peak power density in the fuel 146-cm-long al umi num can that contai ned 49 fuel 3 was 5200 MW/m . The fuel elements were similar elements as compared to 402 in Pewee. Th is core to those of Phoebus-lA except for a few elements was surrounded by a 27-cm-thi ck beryl lium radial CVD-coated with ZrC instead of NbC . The ZrC­ reflector that accommodated six rotati ng control coated fuel elements performed signi ficantl y drums. The fuel inventory was about 5 kg of better. uranium (93% enriched) . Sufficient reactivity The reactor performed close to design condi­ for critical configuration wi th such a small fuel tions except for an unexpected, larg e, radial inventory was obtai ned by designing the core as a variati on of 220-310 K in the fuel -element exi t­ heterogeneous water-moderated thermal reactor. gas temperature and a heat pickup in the Each fuel cel l contained a standard 19-hole, refl ector 14% greater than predicted at ful l hexagonal fuel element encased in an al uminum power. But the successful performance of the tube as descri bed in Fig. 28. The cell tubes Pewee reactor design was important because it were inserted inside al umi num sleeves, and water demonstrated that small reactors, with fl owed through the core in two passes, fi rst be­ low-temperature moderati ng material s inside the tween the sleeves and the eel l tubes to the aft core , coul d be operated in the configurati on and end of the core, where the flow turned around and in the extreme temperature environment of a went back between the elements. The hydrogen rocket engine. A second test of the Pewee cool ant, after making several passes in the reactor had been planned, but Pewee-2 was never reflector assembly, made a single pass through built. the core wi thi n the fuel cool ant channels. The hy drogen exhaust gas was handl ed di f­ E. Nuclear Furnace, NF-1 ferently than in previous reactors. Instead of The last reactor test of the entire Rover being exhausted through a convergent-divergent 46 47) program was that of the NF-1 , ( ' a reactor nozzl e directly to the atmosphere , the hot hy dro­ ten times less in design power than Pewee. The gen was first cool ed by injecti ng water directly NF-1 was devised to provide an inexpensive means into the exhaust gas stream as shown in Fig. 29. of testi ng ful l-size nucl ear rocket reactor fuel The resulti ng mixture of steam and hydrogen gas elements and other core components in a reactor was then ducted to an effluent cl eanup sys tern to

12 remove fi ssion products before release of the fuel mel ti ng temperature was 2683 K, the mel ti ng cleaned gas to the atmosphere. temperature of the uc -c eutectic. 2 The primary obj ecti ves of the NF -1 test The fuel pl ates for the ori gi nal Kiwi-A series were to verify the operati ng characteri s­ reactor were molded and pressed at room tempera­ ti cs of the NF-1 and associated facilities and to ture, then cured to 2723 K. The plates had no operate at ful l power with a fuel -el ement exi t­ coati ng to protect the carbon agai nst hydrogen gas temperature of 2440 K for at least 90 min. corrosion. Al l subsequent reactors used fuel Al l primary objectives were attained duri ng the el ements that were extruded and coated, initially test series. A weal th of data was obtained on with NbC , to reduce hy drogen corrosion. The fuel the dy namic and static characteri stics of the element for the early reactors through Kiwi-BlB NF-1 and the faci lity, and no major NF-1 design were extruded cylinders with fi rst four, then deficiencies were found . seven cool ant channel s. The cylinders were con­ The reactor was operated at the design power tained in graphite modules. Kiwi-B4A was the of 44 MW and a fue1 -element exi t-gas temperature first Kiwi-B design intended as a prototype of approximately 2440 K for a record time of fl ight reactor; and it used 19-hole, one-piece,

109 mi n and at or above 2220 K for 121 mi n. The hexagonal fuel elements, 19 mm across the flats. maximum exit temperature reached was about Thi s fuel element shape became the adopted stan­ 2550 K. Two new ty pes of fuel elements were dard for al l the remaini ng reactor designs. tested in NF-1. They were the ( U,Zr)C graphite The Kiwi-B4E test was the first use of composite) elements that compri sed 47 of the 49 coated uc particles in place of uo ( 2 2 fuel cel ls in the core and two cells containi ng particles in the fuel . The major problem with ( U,Zr)C ( carbide) el ements. The carbide elements oxide-l oaded fuel elements was the so-cal led 3 wi thstood peak power densities of 4500 MW/m back-reaction. Micrometer-size uc particles 2 but experienced severe cracki ng. These elements are extremely reacti ve and revert to oxide in the were smal l ( 5.5 mm across the fl ats) , hexagonal presence of air, particularly humid air. Thus, elements wi th a single 3-mm-di ameter cool ant oxi de-carbide-oxide reactions occurred duri ng hol e. Redesign, by reducing the web thickness by each heating and storage cycl e, including graphi­ 25%, woul d substantial ly decrease the temperature tizi ng , coati ng, and reactor operation; and each gradients and reduce the cracking. The composite cycl e caused loss of carbon by CO gas evolution elements withstood peak power densities in the and degraded the el ement. Dimensional changes 3 fuel of 4500-5000 MW/m and achieved better al so were noted in stored elements. Oxidation of corrosion performance than was observed pre­ the uc loadi ng materi al caused the el ement to 2 viously in the standard , graphi te-matrix, swel l as much as 4% so that the final dimensions Phoebus-type fuel el ement. could not be control led. The sol ution to this probl em was the intro­ ( 8) F. Fuel Development duction of uc particles that were considerably 2 The major technol ogy effort of the Rover larger, 50-150 µm diameter, and coated with program was expended on devel oping fuel s. Al l of -25 µm of pyrolyti c graphite. The fi rst coated the Kiwi reactors except the last one, Kiwi-B4E , pa rti cl es had a low-density pyrocarbon coat that used highly enriched uo fuel in a graphite 2 could not wi thstand high temperatures. At matrix. The uo particle size was 4 µm and the 2 2273 K, the uc core would mi grate through the 3 2 parti cl e density was about 10 .9 g/cm . At high coati ng, thus destroying the protection against temperatures ( 1873-2273 K) duri ng processing, the the back-reaction. Consequently, the graphi­ uo reacted wi th the carbon surrounding it and 2 ti zing temperature had to be hel d lower at was converted to uc wi th evol ution of CO and 2 2173 K. Thi s temperature gradual ly increased consequent loss of carbon from the el ement. The with improved coated particles to 2573 K. Coated

13 parti cles were eventual ly developed that could the temperature ( 2623 K) at which the niobium in wi thstand 2873 K for 0.5 h. Thi s work subse­ contact with carbon was converted to NbC . quently led to the devel opment of TRISO fuel Meanwhile, during the Kiwi-B testing seri es, beads used in co11111ercial high-temperature gas­ the CVD technology was improving and becoming a cooled reactors. The coated particles in the sophisticated process in which 19 ful l-l ength, nuclear rocket engine were not intended as a con­ 1321-mm-long, 2 .4-mm-di ameter bores coul d be tainment for fission products, the pri ncipal re­ coated with NbC tai l ored in thickness over the quirement in conunercial reactors, but to provide ful l length of the el ements. And so al l the fuel stability during fuel -element processing and elements for the reactors from Phoebus-lA through storage and to el imi nate reaction wi th humid air the last one , and includi ng the NRX series of and coating gases. reactors, were CVD coated. The early CVD

Coati ng technol ogy evol ved greatly during coati ngs had a useful 1 ife of about 10 mi n, but the Rover program. As mentioned earl ier, the by the end of the program, NbC and ZrC coati ngs fuel elements of al l the reactors tested in the had been tested for as long as 5 h. Pewee was program, except for Kiwi-A, were coated with NbC the fi rst nuclear test to empl oy some fuel ele­ ( or ZrC late in the program) to reduce hy drogen ments coated with ZrC . They performed s i gni f­ corrosion. It had been real ized early that i cantl y better than e 1 ements with NbC . The pro­ hy drogen and graphite, at the anticipated high gressive improvements achieved in fuel perform­ temperatures of a rocket engine, would react to ance during the NRX and Pewee seri es of tests are form methane, acetylene, and other hydrocarbons. shown in Fig. 30 , where corrosion measured in Further, the graphite loss from hy drogen corro­ terms of mass loss has been normal ized to one for sion duri ng reactor operation would seri ously the NRX-A2 and -A3 te sts. No maj or change in affect the reactor neutronics. So a fuel -element fuel-element design, fabrication method, or coati ng effort was undertaken in 1959 for the characteristics occurred in the NRX series. The Kiwi-A' reactor to develop thin ( 0.025- to elements were al l made from coated uc beads 2 0.05-mm-thick) NbC or ZrC coati ngs to act as a di spersed in a graphi te matri x, extruded wi th 19 barrier to hydrogen attack for the length of time cool ant channel s in a hexagonal pri sm, and coated the reactors were to operate. Niobi um carbide with NbC. The mai n contri buti ng factors for the was selected initial ly because it has a higher improvements were the use in NRX-A6 of a Mo metal eutectic temperature (3523 K) with carbon than overcoat over the NbC bore coat in the fi rst 1-m does ZrC (3123 K). Much later, attention was length of the el ements (this overcoat reduced the shifted to ZrC because it adhered better to midband corrosion, which wi ll be di scussed below, graphite and was more desirabl e neutronical ly. by a factor of 10); the use of thinner NbC The coati ngs were appl ied initially wi th CVD coati ngs, which reduced their tendency to crack; techniques for the fuel el ements in Kiwi -A' and tighter control of processing and tighter control -A3. These fuel el ements were short: 216-mm of the fuel-element external dimensions to reduce cyl inders containing four axial cool ant chan­ interstitial gaps between el ements ; adjustments nel s. The cyl inders were designed to nest into in fl ow orifici ng and fuel loadi ng to improve the one another end to end to build up the total ele­ radial power and temperature profile across the ment length. The Kiwi-B elements were much core. The corrosion at the end of the NRX series longer, and CVD deposition of NbC on fuel -element was reduced to 30% of that at the beginning of bores had not devel oped to the poi nt where they the NRX series, and improvements planned for coul d be coated successfully and reproducibly. Pewee-2, which was never bui 1 t, would have Consequently, a di fferent claddi ng technique wa s reduced this to 10% . used for these reactors. Simply stated, this Much of the fuel testing was done in a hot technique was to insert niobium tubes into the gas te st furnace, which simulated the operati ng fuel -element bores and heat the lined el ements to condi tions, without radiation, of the nuclear

14 4a,49l reactors. The high-pressure furnace, whi ch is an attempt to reduce mi dband corrosion. ( shown in Fig. 31, provided a reasonabl e simul a­ The structure of the composite fuel is compared tion of reactor power density, temperature and to that of the standard gra phi te-matri x fuel in thermal stress, and the effects of fl owi ng hy dro­ Fig. 33. The composite fuel is made from un­ gen. These tests provided, of course, no infor­ coated {U,Zr)C particles in such a way as to form mation about radiation damage, but it was fel t a continuous phase of carbide , as a web through­ that the high temperatures and the small burnup out the graphi te matri x. The structure of the in actual reactor operati ons woul d minimi ze standard fuel shows coated uc particles 2 radiation effects. The fuel el ement under test embedded in a conti nuous graphi te matrix. When was resistively heated wi th de current. The the carbide coati ng lining the cool ant channel s vol ume heat generation produced by ohmic heati ng cracks in this fuel , carbon is lost indefinitely was not an accurate simulation of nuclear through the cracks because the graphite matrix is heating, and changes in fuel-element composition continuous. With the composite fuel , carbon is during the test affected the el ectri cal conduc­ lost through coati ng cracks unti l the carbide tivity of the element potential ly causing prob­ di spersion phase is exposed to the cracks, and lems. But in general , furnace testi ng was then carbon stops escaping except for a smal l valuabl e in the devel opment of new fuel-el ement amount di ffusing through the carbide. As is

technol ogies and a 1 so in qua 1 i ty-contro l sampl ing evident in Fig. 32, the composite fuel did indeed during manufacture of fuel el ements for a spe­ perform better than the graphite fuel . However, cific reactor. for reasons that have not been ful ly determi ned, A maj or probl em al l uded to earl ier through­ the midrange corrosion was still greater than out the fuel devel opment program was the midrange expected . This unexpected corrosion was attri b­ corrosion, as exempl ified in Fig. 32. It was the uted in part to cracking from excessive thermal region about one-third the length from the col d stress that resul ted from a decrease in thermal end of the core where corrosion was greatest. conductivity duri ng the power run. This The inlet end of the core had low corrosion rates decrease, which was measured, is thought to have because the temperatures were low. The fuel been caused by fission fragments. Presumably operated at much higher temperatur�s toward the such an effect woul d not occur in the standard , nozzl e chamber end of the core , but the fuel wa s coated-particl e, matrix fuel because the fi ssion processed during fabricati on to accept the high­ fragments do not penetrate through the particle end temperatures. Al so the neutron fl ux, and coati ngs to damage the matri x. hence the power density, was low, resul ti ng in Pure (U,Zr)C carbide fuel s were al so tested low thermal stresses and consequently mi nimal in NF -1 as another approach to reducing corro­ cracking. There, mass loss was mostly due to sion. The fuel elements were fabricated as smal l carbon di ffusion through the carbide coati ng. hexagonal rods wi th one cool ant channel at their However, in the midrange , the power density was center. The fuel experienced mi nimal corrosion , high and the temperature was now appreciabl e, yet but it cracked extensively as a resul t of its low

still much lower than that at which the fuel was fracture resistance and 1 ow thermal conduc­ processed; the carbide coati ngs woul d crack tivity. However, by increasing the strai n-to­ because of mi smatched expansion coefficients, and fracture characteri stics of the el ements and high mass losses woul d occur through the crack s. redesigni ng their shape to reduce their cross­ The improved performance of the ZrC coati ng is secti on and web thickness, the performance of the clearly shown in Fig. 32, as is that of a new carbide el ements coul d be substantially improved. type of fuel cal led composi te fuel . Yet another advanced fuel was being devel oped The composite fuel was devel oped near the near the end of the program. This fuel wa s end of the Rover program and tested in the similar to the standard uc -coated particles in 2 Nucl ear Furnace al ong wi th pure carbide fuel as graphite-matrix fuel , but the graphite matri x wa s

15 made wi th POCO carbon-filler fl our to yiel d a fuel s for electri cal-power production appl ica­ matri x having a higher coefficient of thermal tions. expansion (CTE ) that cl osely matched that of the Fuel structures were al so a maj or probl em in NbC or ZrC channel coati ngs. This fuel , referred the reactor devel opment. Reactor cores in the to as the high-CTE graphi te-matrix fuel , was early Kiwi seri es essential ly fel l apart from fabricated into fuel el ements, and it exhibited vibrations; these were induced thermal-hydraulic better strain-to-fail ure characteri stics than the interactions. By the end of the Kiwi series, an standard fuel . It was intended for the NF-2 test adequate structural support system had been that, unfortunately, did not take pl ace. But the demonstrated. Improvements in the structural promi sing resul ts obtained before cancel lati on of system conti nued to be made , wi th the results the program should be seri ously considered in any su111nari zed in Tabl e I. It should be emphasized future graphite fuel -el ement devel opment. that what structural anomal ies existed were The demonstrated operating performance of determi ned after the tests and did not cause a the standard graphite-matrix fuel was 1 h at a termi nation of power. At the end of the NRX-A6 cool ant exit temperature in the range of test, some cracking in the beryl lium-refl ector 2400-2600 K. This performance was obtained ri ng, support blocks, peri pheral composite cups, mainly from the NRX-A6 and Pewee tests. The and one tungsten cup was found. These cracks demonstrated performances of the advanced com­ were bel i eved to be the resul t of excessive posite and pure carbide fuel s were nearly 2 h thermal gradients. ( 109 mi n) at 2450 K and at a peak power density 3 in the fuel of 4500-5000 MW/m , as obtained in IV. ENGINE DEVELOPMENT the NF-1 test. Based on the extensive fuel s work achieved during the Rover program, projections of A. Engi ne Tests endurance limi ts were estimated as shown in An engine devel opment program wa s carried Fig. 34. These projections indicate that the out as part of the nuclear reactor research and composite fuel shoul d be good for 2-6 h in the devel opment te st series of Phoebus and NRX. temperature range of 2500-2800 K. Similar per­ Prime responsibility for this effort rested wi th formance can be expected at 3000-3200 K for the the Aerojet-General Corporation. The primary carbide fuel s, assumi ng that the cracking probl em objecti ve of thi s test series was to further can be reduced through improved desi gn. For 10 h extend nuclear rocket technology in preparation of operation, the graphi te-matrix fuel would be for a fl ight sy stem. Thi s invol ved incorporati ng limi ted to a cool ant exit temperature of the advances made in reactor devel opment into an 2200-2300 K, the composite fuel could go to engine that compri sed the nonnuclear components nearly 2400 K, and the pure carbide to about of a compl ete fl ight system. Of parti cular 3000 K. interest were the investigation of engine And so the program was terminated wi th three startup , shutdown, and restart characteristics promi sing fuel forms at hand, the carbide-carbon for different initial condi ti ons; the eval uating composite, the pure carbide, and the high-CTE of vari ous control concepts; and testi ng the graphite matrix. As di scussed above, much performance of nonnuclear engine components in testi ng was performed on these fuel s, but their the nuclear environment. Two ful 1-power nuclear corrosion behavior was not completely under­ test series can be categori zed in this program. (50l stood. Most of the work was done in the tempera­ These are NRX/EST , which was carried out in 51 52) ture range of 2000-2800 K. This range would have February-March 1966, and XE' , ( • which took to be extended to lower temperatures ( below pl ace in March through August 1969. Both of 1500 K) and testi ng done wi th gases other than these tests empl oyed 1100-MW NRX-type reactors. hy drogen to eval uate the performance of these In addition, a col d-fl ow test seri es Experimental

16 Engine Col d Fl ow (XECF ) was conducted in shutoff val ve is first opened, the pump tends to February-Apri l 1968. vapori ze the fl uid until sufficient fluid passes The NRX/EST di spl ayed in Fig. 35 was the through it to chill down the pump to cryogenic fi rst operation of a NERVA breadboard power plant conditions. The nozzl e al so tends to be a choke wi th engine components connected in a fl i ght­ poi nt, as are the core and refl ector inl ets. functi onal rel ati onship. The te st demonstrated Therefore, a certain amount of fl uid must be the stability of the power plant under a number passed through the sy stem to remove the stored of different control modes whi le the engi ne oper­ heat in the lines, val ves, and refl ector. Once ated over a broad area of its performance map . this is accompl ished, the pump can be started and The endurance capability of the reactor and other wi ll operate normal ly. Approximately 1 mi n of engine components was demonstrated by operating fl uid fl ow is necessary to accomplish this func­ the power plant at significant power duri ng 5 tion. Duri ng thi s time period, the reactor can different days for a total of 1 h and 50 min, of be brought up to a low-power level . Reactor which 28 mi n were at ful l power. These tests drums are prograR111ed out rapidly, almost to the served to demonstrate the mul tiple restart capa­ col d cri tical point, and then put on a sl ow tran- bilities of the engine, includi ng the feasibility sient. Once appreciable temperature ri se is of restarting the engine without an external sensed in the chamber, the reactor can be power source. swi tched to closed-l oop temperature contro 1. Operation of the XE ' Engine ( Figs. 36 and Thi s scheme does not require any neutroni c 37) was the first te st of a down-fi ri ng nucl ear instrumentati on. When appreciable power has been rocket engine with components in a fl i ght- type, achieved and the turbopump is running, the engine cl ose-coupled arrangement. The te st stand pro­ can be accel erated at the rate of 83 K/s. vided a reduced atmospheric pressure ( about Experience on the NRX/EST and XE ' engine programs 1 psia, or 60 000 ft al titude) around the engine showed that the engine system can be control led to partially simulate space conditions. The in a predictabl e and safe manner. engine was successful ly operated at ful l power. ( SS) It ran at various power level s on di fferent days B. Engi ne Design Improvements for a total of 115 min of power operation that Other goal s of the engine devel opment pro­ incl uded 28 restarts. The bootstrap startups gram, besides demonstrati ng engine feasibility (without external power) were accompl i shed over a and control , were fi rst of al l to maximize spe­ range of pump inlet sucti on pressures and with cific impul se, which is proportional to the reactor conditions spanning the range that woul d square root of the nozzl e chamber temperature; to be encountered in fl ight operations. Compl etely meet various design thrust level s that are pro­ automatic startup was demonstrated. The capa­ portional to fl ow rate and that demonstrate the bil ity of the engi ne to fol l ow demanded tempera­ capabi lity to throttle the engi ne down and ture ramp rates up to 56 K/ s was demonstrated, operate at a reduced thrust; to minimize engine and based on thi s information, assurance wa s size and weight; and to increase longevity from gained that rates up to 83 K/s could be achieved an initial 1 h to 10 h. In fact, system oper­ wi thout exposi ng any of the engine components to ati ng life is real ly determined by the amount of a transient condition that would exceed its propel lant that can be transported to space in a design limi tations. reasonable payload to perform the mi ssion. Figure 38 shows some of the characteri stics Increases in the chamber temperature were ( 52-54l of sta rti ng an engine of th is sort. The made first of a 11 by improvi ng the reactor fuel engine components must be conditioned before high performance to permit raising the operating tem­ power can be reached. The turbopump , nozzl e, perature, as discussed in the preceding chapter. refl ector, and core inlet are al l designed to A number of design changes al so were made to ope.rate at low temperatures. When the pump improve reactor performance. The cores of the

17 56 57) early reactors were supported axially wi th ti e engine( • and Phoebus-2A were tested. Ma ss rods attached to the col d-end support plate. The estimates are detailed in Tabl e I II. A compari ­ rods were cooled by hy drogen that di scharged into son of just the reactor masses was shown earl ier the nozzl e chamber, and this cool ing 1 owe red the in Fi g. 10. Both the NERVA and Small Engine rocket speci fic impul se because the ti e-rod cool ­ designs took advantage of the ful l-flow cycle 1 58-61) ant exi t temperature was much 1 ower than that of idea. The NERVA was designed for the the cool ant exiting the main fuel elements. In 337-kN ( 75 000-lb) thrust level using a 1570-MW later reactors and in engine sy stems desi gned for reactor, and the Smal l Engi ne was designed for a fl ight, tie tubes were substi tuted for tie rods. 72-kN thrust level with a 367-MW reactor. The The tubes were regeneratively cool ed wi th the equival ent vacuum specific impul se of the XE' coolant di scharging into the core inlet rather engi ne was 710 s. This was improved to 825 s for than the core exit. Al so, in the early reactors, the NERVA fl i ght engine and 875 s for the Small a large core peripheral fl ow rate was used to Engine. The increasing specific impul se level s protect the core-refl ector interface, and thi s are reflections of the chamber temperatures that cooler flow was discharged into the nozzl e went from 2270 K as demonstrated in the XE' test chamber. Thi s fl ow was steadily decreased to to a design val ue of 2695 K for the Small Engine. 3l almost nothing to increase the nozzl e chamber The Smal l Engine(2• depicted sche- temperature . A final optimi zation would be to matically in Fi g. 39 real ly re presents an accumu- employ a regeneratively cooled heat exchanger for 1 ation of al l of the knowl edge gained in the the peri phery assembly. nuclear rocket program. It used hy drogen as the The engine fl ow cycle was al so changed to propel lant, the ful l -fl ow engine topping cycle, increase specific impul se. The XE engine em­ and a singl e-stage centri fugal pump with a ployed the "hot-bleed" cycl e to dri ve the cool ant singl e-stage turbine. It had a regenerati vely turbopump. In th is cycl e, some cool ant is ex­ cool ed nozzl e and tie-tube support elements . A tracted from the chamber and mixed wi th cool ant radi ation shield of borated zirconium hy dri de was from the refl ector outl et, and tne combi ned cool ­ incorporated above the reactor. Thi s was mainly ant is used to dri ve the turbine. Because the to reduce heati ng to the propel lant tank above turbine exhaust pressure is low, this cool ant the engine, al though it al so provided shiel di ng coul d not be reintroduced into the mai n fl ow for the payload and crew. Reactor control wa s stream; it was di scharged into space at a 1 ow done wi th six actuators for the 12 control drums temperature rel ati ve to the nozzle chamber condi­ in the beryllium refl ector. The engine empl oyed tions, thereby reducing the overal l specific only five valves and their actuators, incl uding a impul se of the engine. propel lant tank shutoff val ve (PSOV ) 1ocated at Fi nal engine designs evol ved to empl oy a the bottom of the propel lant tank to provide a ful l -flow or topping cycl e in which the turbine ti ght seal against propel lant leakage when the recei ves fl uid from the tie- tube outl et and di s­ engine is not in use; a nozzl e control val ve charges it into the core inl et. Thi s cycle, ( NCV) to adjust the fl ow spl it between the nozzl e never tested experimental ly, significantly raises cool ant tubes and the tie tubes; a turbine seri es the specific impul se of the engi ne. The turbine control val ve ( TSCV) that coul d be used to inlet temperature in the ful l-flow cycle is much isolate the turbine duri ng preconditioning and lower than in the hot-bleed cycl e. Consequently cool down to remove after-heat and to extend the much greater turbine flow rates and di scharge control range; a turbine by pass control valve pressures are requi red than in the former cycl e. ( TBCV) to regul ate the amount of fl ow to the turbi ne and thus the turbopump speed and flow C. NERVA and Small Engi ne Designs rate ; and a cool down control val ve (CCV) to Design characteristics for several engi nes regul ate hy drogen fl ow for decay heat removal are listed in Tabl e II. Only the experimental XE ' fol lowi ng engine operation and together wi th a

18 smal l pump , to provide prepressuri zati on for the fl at. The effective core diameter was 570 mm and tank. the overall reactor diameter was 950 mm. A regeneratively cool ed nozzle was used out As shown in Table III, the overall Small to the area ratio of 25:1. It was fol lowed by an Engine weight was 2550 kg , with the reactor uncooled skirt section. Thi s section extended (minus the shield) almost 1600 kg of thi s. the nozzl e out to an area ratio of 100: 1. The Figure 42 su11111arizes the Smal l Engine state uncooled nozzl e skirt was hinged to facilitate points at design conditions. The chamber tem­ packaging in the 1 aunch vehicl e. This arrange­ pera tu re was nearly 2700 K. ment. provided room for a 1 arger propel 1 ant tank A study was made to see how the Smal l Engine in the 1 aunch vehicle. The overall engi ne 1 ength would mate wi th the propel lant tank so as to fit was 3.1 m with the skirt folded, or 4.4 m with within the space shuttl e. The resul ts are sum­ the skirt in place. The total mass of the system marized in Fig. 43. The nuclear stage with the was 2550 kg . Smal l Engine would weigh close to 18 000 kg , of The reactor core, pictured in Fig. 40 , was which almost 13 000 kg woul d be propel lant. If designed to produce about 370 MW . There were additional propel lant modules were sent up sepa­ t �64 hexagonally shaped, (UC-ZrC )C composite fuel rately, the stage woul d then weigh 23 000 kg , el ements contai ni ng a total of 52.4 kg of uranium wi th over 21 000 kg of hy drogen availabl e for (0.9315 enri chment) . Each fuel element had 19 propul sion. At a fl ow rate of 8.5 kg/s, a singl e coolant channel s. There were 241 support el e­ nuclear stage would operate for approximately ments, containing zirconium hy dri de , ZrH , as a 1500 s in space. An additional propel lant modul e 2 neutron moderator. The core peri phery incl uded woul d add 2500 s to the operati ng time . Thus, a an outer insulator layer, a cooled inboard slat possible lifetime of several hours woul d allow section, a metal wrapper, a cool ed outboard sl at performing many significant mi ssions. section, and an expansion gap. The core was sup­ (60 ported on the col d end by an al uminum alloy plate D. Compo nent Devel opment l with the support plate resting on the reflector A vigorous program for the devel opment of system. The reactor was contained in an aluminum nonnucl ear engine components accompanied the pressure vessel. A beryllium bar.rel wi th 12 reactor and engine test programs. The pri ncipal reactivity control drums surrounded the core. components were the mai n cool ant turbopump, the The reactor was designed for 83 K/s temperature valves and actuators, the nozzl e assembly, the transients. reactor pressure vessel , radiati on shiel ding, and Figure 41 provi des more detai ls on the fuel the control s and instrumentation. modul es. The fuel provided the heat transfer Figure 44 is a picture of the turbopump surface and the energy for heating the hy drogen. devel oped for the XE' engine, including a listing It consisted of a composite matrix of UC -ZrC of some key parameters related to the turbopump. sol id solution and carbon. The channel s were The pump performs the function of pressuri zing coated with zirconium carbide to protect against the propel lant for the engine feed system. The ti;'drogen reactions. The ti e tubes transmi tted 1 ow flow rates required by the Small Engine made the core axial pressure load from the hot end of it possible to run the shaft speed at a much the fuel elements to the core support pl ate. higher rate than for the XE' engine or NERVA . They al so provided an energy source for the turbo­ The XE' turbopump was a singl e-stage, radial exit pump and contained and cool ed the zirconium fl ow, centri fugal pump wi th an al uminum pro­ carbide moderator sl eeves. They consisted of a pel ler, a power transmi ssion that coupled the counterfl ow heat exchanger of Inconel 718 and a pump to the turbine, and a two- stage turbine wi th zirconium hy dri de moderator wi th zirconium stai nl ess steel rotors. On NRX/EST , this pump carbide insulation sleeves. The el ements were performed eight starts and operated 54.4 min at 0.89 m long and measured 19.1 mm from fl at to high power. In the XE' engine, it performed

19 28 starts and restarts, incl udi ng runs to rated stainl ess steel for the cool ant channel s. The power. Potential problem areas were wi th the graphi te nozzl e extension was uncool ed out to an shaft system bindi ng at the beari ng cool ant, a area ratio of 100:1. The cool ed section used di fficulty that was experienced in the XE tests. U-tube constructi ons and a divergent section. The solution was to increase clearance and to The major unresolved probl ems in achieving a 10-h improve al i gnment. Beari ngs are probably one of life were associ ated wi th some remai ni ng stress the few life-limi ting components in the non- probl ems in the al umi num al loy. Fabrication nuclear subsystem. Resul ts of life tests are problems appear to have been resol ved. listed in Table IV. The solution to the beari ng Figure 48 depicts the pressure vessel and (64) probl em seemed to depend on maintai ni ng adequate enclosure. Its functions were to support cool ing to reduce wear. the components of the reactor assembly, to form a Various valves were requi red in the system pressure shel l for the hy drogen propellant, and (52 60-62) to control the hy drogen fl ow. • These to transmit thrust to the thrust structure. The val ves were binary valves, except for the in-out design conditions, as speci fied for the NERVA , control val ves and check valves, such as shown in were a maximum fl ow rate of 37. 6 kg/ s, maximum Fig. 45. Val ve operati ng experience wi th a pressure of 8.66 MPa, temperature range from 20 reactor was obtai ned in both the NRX/EST and XE' to 180 K, rel iability of fewer than three 6 engine tests. Tabl e V lists the val ve and failures in 10 fl ights, and a service life of actuator characteri stics. The major potential 10 h. Similar designs were demonstrated in the problems appeared to be from seal damage by con­ fi ve NRX tests and the XE' engine. The pressure tami nants, erroneous position indicators, and vessel was constructed of a cyl i nder that had a leakage from poor lip-seal tol erances. top cl osure wi th bol ts and seal s. A one-piece The number of val ves in the Small Engine was extruded forging of al uminum alloy 7075-773 was fi ve. However, there was some desire in the used, wi th a surface coati ng of Al o • The 2 3 NERVA fl i ght engine to increase system rel iabil­ maj or i terns still bei ng designed were the best ity by having two turbopumps, either of which ways of assuri ng bulk prel oad and of final izing could provide ful l flow and pressure to the the surface coati ngs. engine systems, and redundant valve configura­ A radiation shield was located between the tions. In order to provide swi tching between the reactor co re and the propellant tank. The shiel d turbopumps and to provide high reliability by was intended to prevent neutron heati ng of the backing up each val ve in case of a failure , some propel lant, and it al so provided biological 26 val ves woul d be needed , as seen in shielding for the crew as the tank emptied of its (6l) Fig. 46. Indeed it becomes questionabl e propel lant. It did not present any difficul t (65) whether the redundancy gained is worth the added design problems. system complexi ty. Control s and instruments were another major The nozzl e assembly is used to expand the area of devel opment duri ng the Rover pro­ !35 heated gas from the reactor in order to pro vi de gram. l For control -drum actuators, pneu­ maximum thrust. A nozzl e is pictured in matic-type actuators were devel oped. These were (63) Fig. 47. The design conditions for the demonstrated on the XE' engine without apparent NERVA fli ght engine were a thrust level of degradation or anomal ies. 337 kN , an area ratio of 100:1, a service life of Instrumentation was al so a maj or devel op­

10 h, rel iability of fewer than four fai l ures in ment. Thermocoup1 es demonstrated performance at 4 10 fl ights, chamber pressure of 3.1 MPa, 2667 K for 1 h wi thout degradation. Thermocouple chamber temperature of 2360 K, flow rate of di spl acement, pressure, and vibration sensors 41.6 kg/s, and cool ant channel temperature were devel oped for several hours of operation. A between 28 and 33 K. The regeneratively coo 1 ed 1% measurement accuracy wi ll requi re some further nozzle part used an al umi num alloy jacket and devel opment.

20 Control logic reached a high degree of auto­ experimental engine , but the highest equivalent mati on wi th demonstration of automatic control specific impul se achieved was 845 s in Pewee, systems in XE' for al l operational phases. Feed­ which operated at a peak cool ant exit temperature back control loops and drum positi on, power, tem­ of 2550 K and a peak fuel temperature of 2750 K. perature, turbine control val ve position, and A core average power density as high as 3 pressure were developed. 2340 MW/m and peak maximum fuel power density 3 of 5200 MW/m were obtai ned as well in Pewee. E. Testi ng Faci lities The NF-1, which experienced nearly as hf gh peak Testing facilities were another maj or devel ­ power densities, ran at ful l power and an average opment item . During the nuclear rocket program, cool ant exit temperature of 2445 K for an accumu- major test facilities were devel oped at the NRDS 1 ated time of 109 mi n. The experience gained at Jackass Flats in Nevada (Ref. 2, Vol . III). from these tests indicates that the composite

These included reactor test facilities, engine fuel would 1 ast for 6 h under these condf ti ons t49 test facilities, and assembly and di sassembly before appreciabl e fuel loss occurs. l facilities (shown earl ier in Fi g. 4) . The The nonnuclear components of the engine reactor test facilities were designed to te st the proved to be capable of achievi ng the endurance reactor in an upward-firi ng position. These used required for testi ng . Ful ly automatic control facility-type feed systems for providing the and bootstrap startup were demonstrated for a hy drogen to cool the reactor and to support the wide range of operating conditions in NRX/EST and reactor tests. The first downward-firi ng faci l­ XE ', the latter experienci ng 28 engine starts and ity, which al so included some atmosphere simul a­ restarts wi th temperature ramp rates up to ti on, was the engine te st faci lity. Thi s was 56 K/s, which could be rai sed with confidence to used in the XE col d-flow te sts and al so the XE ' 83 Kl s. ful l -power test. Thi s secti on describes the Rover technol ogy For mai ntenance, assembl ing, and disassem­ base. The program was termi nated at the poi nt of bl ing of the reactor and engine systems, there fli ght engine devel opment. For a fl i ght system , were buildings cal led Maintenance Assembly and it would be necessary to verify the fl ight Di sassembly (Ml\D) , which provided the necessary reactor and engi ne design, perform duration facilities, hot cel ls, and other equipment for testi ng , and veri fy reproducibility. There putti ng together and taking apart the reactors. appear to be no technological barriers to the The MAD buildings were linked (Fi g. 49) to their construction of a successful nucl ear rocket. respecti ve test facilities by railroads that were used to transport the reactors by remote control B. Space Power Generation if necessary. The nuclear rocket engine technol ogy base is directly appl icable to the generation of electric V. FUTURE DEVELOPMENTS power in space, particul arly for hi gh-power (so­ c all ed mul timegawatt, 10- 100 MW ) , open-cycl e (66) A. Fl ight Engi ne systems. A schematic drawi�g of such a The basic research and technol ogy devel op­ power pl ant is shown in Fig. 50 . The pl ant is ment required for a nucl ear rocket fl i ght engi ne similar to a rocket engine in which the engine were essentially compl eted during the Rover pro­ nozzl e has been repl aced by a turbine to generate gram. Power level s in the range of 500-4100 MW electricity. The cool ant gas is then exhausted were demonstrated in the NRX , Phoebus, and Pewee in such a way as not to produce any thrust. The test series. A thru st level of 930 kN (200 000 core exi t temperature of the hy drogen cool ant 1 bl was reached in Phoebus-2A with a hydrogen would have to be much lower than that of the fl ow rate of 120 kg/s. A speci fic impul se of rocket engine because of materi al limitations on 710 s (6060 m/s) was obtai ned in the XE the turboal ternator. Thi s means that for a given

21 cool ant fl ow rate the power would be reduced, or carbide embedded in a graphi te matri x, except for for a given power level , the flow rate would have NF-1 where pure UC fuel was tested al so. The to be increased as compared to that in the rocket pure UC fuel was promi sing but not ful ly devel­ engine. A mul timegawatt, open-cycle power plant oped, as extensive cracking of the fuel occurred woul d necessarily have a durati on at power meas­ in NF-1 . For a variety of materi al s considera­ ured in hours, bei ng limited by the amount of tions, the density of UC in the composite fuel s cool ant that can be re asonably transported to is limited to 700-800 mg/cc so that smal l reactor space. Because of the low operati ng temperatures size can be achieved only by incorporati ng of such a power pl ant, hy drogen corrosion in the moderati ng materi al such as zirconium hy dri de in reactor would be greatly reduced and would not be the core as was done in Pewee and the Smal l a life-l imiting factor. Engine, or water as was done in NF-1 . The cri ti­ The Rover reactor designs are al so applic­ cal mass for the Pewee and Smal l Engine reactors 235 able to cl osed-l oop multimegawatt space power was in the range of 35-60 kg of U. The systems. The reactor inlet and outlet conditions fractional burnup that can be tolerated in these woul d be different from those of the nuclear reactors is not limited by the capabilities of rocket engines, and the cool ant would be hel ium the fuel but rather by the reactivity margin or a mixture of hel ium and xenon instead of available from the refl ector control sy stem. A hydrogen. The three reactor fuel s devel oped practical burnup limi t based on reactivity con­ prior to tenni nation of the Rover program woul d trol margi n is probably in the nei ghborhood of certainly be promi sing candidates for such power 10%. Thus, assumi ng an energy equivalent of 1 g 235 pl ants al though much work woul d need to be done of u per megawatt day or O. 36 kg/MW yr, the to map out their corrosion and erosion behavi or integrated power capabi lity of this class of as a functi on of time and temperature wi th these reactors is in the range of 10-20 MW yr. Pos­ gases. The lower operati ng temperature woul d sibly this capabi lity coul d be increased through penni t use of the TRISO-design fuel beads in fuel the use of burnabl e poi sons or breedi ng of less­ fabrication. The TRISO beads, devel oped and used enriched fuel . in commercial gas-cool ed reactors, woul d provide the advantage of reducing possible fi ssion prod­ C. Dual-Mode Reactors uct contami nation of the working fl uid to low Near the end of the Rover program, it was level s. re al i zed that it woul d be convenient to use the The simplest power plant woul d be a di rect nuclear rocket engi ne to provide long-durati on, (67-69) . Brayton eye l e power convers i on sy stem 1 n auxiliary el ectri cal power for station keeping. which hi gh-pressure gas enters the reactor where Studies were carried out to modify the Smal l it is heated, expands through a turbine gener­ Engine design to perfonn two modes of operati on ating el ectri c power, and passes through a re­ as shown in Fig. 51. !4•5•7o.7ll cuperator and heat-sink heat exchanger where it The nonnal , rocket propulsion mode of the is cool ed. The gas is then repressurized in the Smal l Engi ne was unchanged. But a separate mode compressor and partial ly heated in the recuper­ was incorporated into the design, where the ator before re-enteri ng the reactor. hy drogen cool ant, which nonnal ly fl owed through Cl osed-cycle power pl ants are attracti ve the tie- tube core support structure and subse­ because their mission life at ful l power is not quently through the turbopump , coul d be diverted limi ted by the coolant supply. They are limited, and isol ated to fl ow conti nuously in a closed­ however, by burnup of fuel in the core. Several loop system as shown in Fig. 52. In this mode, smal l reactor designs were studied duri ng the 10 to 25 kW of el ectri c power could be generated Rover program includi ng Pewee, Nuclear Furnace, conti nuously . the Smal l Engine, and others. The fuel used in Thi s dual-mode design employs an organic these studies was always some fonn of uranium Rankine cy cl e usi ng thiophene as the worki ng

22 fluid. For a 10-kW system, the parti cular open-cycle el ectri c-power sy stem coul d be incor­ e design case shown in Fig. 52, the radiator would porated wi thout el iminating the rocket engine be incorporated into the surface of the propul­ th rusting potenti al by using a poppet val ve, as ( ?2) sion modul e coolant tank. For higher powers, shown in Fi g. 54. The valve, normal ly addi tional radiator area would be needed. The retracted into the convergent section of the maj or engine modi ficati ons required for the aux­ nozzl e, is moved to cl ose off the rocket nozzle iliary power plant were the addition of isolation and open a bl eed line to di vert the heated val ves for the fl ow through the ti e-tubes support hydrogen to the turboal ternator in the power­ structure; a change of materi al from al umi num to generati ng mode. stainless steel for the reactor dome , core sup­ In sunvnary, the Rover nuclear rocket tech­ port plate, and tie-tube supply lines; and a nology is relevant to the generation of electric materi.al change in the control -actuator wt ndi ngs power in space, parti cularly for mul timegawatt, from polyimi de to ceramic. The material changes open-cycle, si ngl e- or dual -mode systems. Con­ were needed to accommodate the wi der temperature ceptual studies performed during the program show range of dual-mode operation. The reactor inlet that a station-keeping auxi liary power pl ant in

and the reflector are uncool ed duri ng el ec tri c­ the range of 10-25 kWe can be incorporated into power generation, and the ti e tubes run hotter the nucl ear rocket engine, wi th only mi nor modi­ than duri ng the propul sion engine operation mode. fications to the engine al ready devel oped, to The incremental mass penal ty of a 10-kW produce a dual -mode sy stem. But extensive re­ e power plant was calculated to be about design would be required to provi de greater, con­ 50 kg/kW , incl uding the main engine modifica­ ti nuous power. Likewi se, rel atively minor modi­ e tions. Thi s penal ty could be reduced to fications would be re qui red to convert the 35 kg/kW for an optimi zed 25-kW sy stem. nuclear rocket engi ne to the power source for an e e The lifetime of th is dual-mode power pl ant is open-cycle Brayton power pl ant. Other than con­ limited by the radiati on degradation limi t of the ceptual studies, however, littl e work was done in thiophene working fl uid. The study indicated that the program to devel op the el ectri c conversion in 2 yr of operation at 10 kW , the radiati on systems themsel ve s. e dose to the thi ophene woul d be approximately a factor of 10 bel ow that limi t. In the same time, VI. SUMMARY the burnup of fuel in the core woul d be insigni f­ In th is report we presented fi rst a summary icant. A feature of a dual-mode engine not yet of mil estone achievements obtained duri ng the mentioned is the possibility for reduci ng the Rover nuclear rocket propul sion program. This amount of propel lant used to remove the reactor summary was fol l owed by a bri ef chronol ogical after-heat fol lowi ng an engine shutdown by run­ hi story of the program, a general ly chronol ogical ning the auxiliary power pl ant to cool the descri ption of the reactor technol ogy development reactor. For the 10-kW system, a maximum e wi th emphasis on the fuel s program, and a des­ therma 1 power of 140 kW coul d be removed in this cri pti on of the engine testi ng program, engi ne manner, 1 imi ted simply by the radiator configura­ design, evol ution, and component technology tion. status as of the end of the Rover program. We The dua 1-mode rocket engine desert bed here ended by discussing a few of the ideas and can be altered to become a bimodal power pl ant, studies that were generated near the end of the as shown in Fig. 53, in which the nozzl e has been program in recognition of the need to apply the replaced by an open-cycl e turboal ternator. Thi s demonstrated nucl ear rocket technol ogy to the configuration provides a high electri c-power generati on of el ectri c power in space. A val u­ generation mode for a duration limited by the able part of this report is the significant but amount of stored cool ant, pl us a continuous, low unavoidably incomplete (over 100 000 vol umes of electri c-power mode for station keeping. The data were generated duri ng the program)

23 bibliography that we have compiled. We have made VII. ACKNOWLEDGl>ENTS an attempt to list only the most perti nent refer­ I am greatly indebted to David Buden for his ences that we could find. many contri buti ons to this work. He suggested Today , a dozen years after termination of the scope of the report , did some of the research the Rover program, as we ponder future requi re­ for it, and graciously reviewed the final prod­ ments for nuclear power in space, we can pl ace uct. am al so grateful to William Kirk and the rel evancy of the Rover technology to our cur­ Michael Seaton for their careful and cri tical rent needs in the fol lowing perspecti ve. The reviews, and to Betty Murphy , Sal ly Sul livan, and NERVA nucl ear rocket reactor, as exemplified by Miquel a Sanchez for preparing the manuscri pt. I the NRX-A6 reactor, was not the fruit of a paper am al so grateful to Rosina Martinez for making study. It was the resul t of a long series of corrections and completing the final report . experimental tests and was demonstrated to work and to operate under conditi ons of temperature VIII. SUPPLEMENTAL BIBLIOGRAPHY and power densities more severe than are likely to be encountered in space reactors designed Rover Program Reviews and Status Reports specifical ly for power production. The NERVA "Nuclear Rocket Propulsion," National Aeronautics reactor design was demonstrated to be sound and Space Admi ni stration publ i cation NASA SP-20 , structural ly, thermohydraul ically, and neu­ December 1962. tronical ly. Al though it was optimi zed for thrust production rather than power generation, it can R. Spence, "The Rover Nucl ear Rocket Program," serve as a real reference point against whi ch Science, 160, No. 3831, May 1968. large, future space reactor designs of any sort can be judged and eval uated, if the di fferi ng R. W. Schroeder, "NERVA--Enteri ng a New Phase," operati ng conditions are properly extrapol ated. Astronautics and Aeronautics, p. 42, May 1968. The class of Rover reactors was tested over an enormous power range from 0 to 4000 MW , there­ R. E. Schreiber, "Nucl ear Propul sion for Space," by pro vi di ng val uabl e seal ing rel ati onships for Los Al amos Sc ientific Laboratory report size , mass, and other parameters, based on real LA-DC-9931, October 1968. data. The hardware technology acquired through the W. R. Corl iss and F. C. Schwenk, "Nuclear Propul­

Rover program and the basic reactor desi gn itsel f sion for Space ," US Atomi c Energy Commi ssion a re directly appl icabl e to future space power­ Technical Information, Library of Congress cata­ p l ant designs based on either closed- or open- log card No. 79-171030 , 1968 (rev. 1971) .

1 oop gas cycles. Probably the outstanding ques­ tion here is how to extrapolate the erosion and K. Boyer, "Status of the Nucl ear Rocket Propul ­ corrosion behavior of the best Rover reactor sion Program ," Proc . XXth International Astro­ fuel s, obtained at hydrogen gas exit temperatures nautical Congress (Selected Papers) Ma r del Plata of 2500 K, to gas temperatures likely to be below 1969, p. 287, Pergamon Press, 1972. 1500 K, and for gases other than hy drogen such as He or He-Xe mi xtures. W. H. Esselman and M. R. Kefl er, "The NERVA Final ly, many of the design and testi ng Nuclear Rocket-A Status Report ," Atompraxi s � methodol ogies, as wel l as tool s, computer codes, Heft 4, 1970 . and facilities, developed during the Rover pro­ gram can be adapted for use in the devel opment of M. T. Johnson, "NERVA Reactors," Proc. Ameri can future nuclear power pl ants in space. Nucl ear Society 1970 Topical Meeti ng, Huntsville, Al abama, Apri l 28-30 , 1970 .

24 M. Klein, US 91st Congress, Second Session, W. L. Kirk, J. R. Hopkins, and L. J. Sapir, Senate Commi ttee on Aeronautical and Space "Study of He terogeneous Reactor for Rocket Pro­ Sciences, "National Aeronautics and Space pulsion," Los Al amos Sci enti fi c Laboratory report Admi nistration Authori zati on for Fi scal Year LAMS-2930 , Apri l 1963. 1971. Hearings, Part 2," Washington, D.C. , Pri nting Office , March 5, 1970 . W. J. Houghton and W. L. Kirk, "Phoebus II Reactor Analysi s," Los Al amos Scienti fic Labora­ D. S. Gabriel , US 93rd Congress, Senate Committee tory report LAMS-2840 , Apri l 1963. on Aeronautical and Space Sciences, "National Aeronautics and Space Admi nistration Authori za­ J. M. McCl ary et al ., "Thermal Analysis of the tion for Fiscal Year 1973. Heari ngs," Kiwi-B-4D Reactor Parts I-V," Los Al amos Scien­ Washington, D.C., Pri nti ng Office, 1972. ti fic Laboratory report LA-3170-MS, December 1964 .

Aerojet-General Corporati on Yearly Progress W. L. Kirk et al ., "A Design Study of Low-Power,

Reports (for NRX and XE tests) , 1965-1969. Light-Weight Rover Reactors," Los A 1 amos Scien­ ti fic Laboratory report LA-3642-MS , June 1968. Los Al amos Sc ienti fic Laboratory, "Quarterly Status Reports of LASL Rover Program," 1965-1972 . Small Engine Design Study - See Ref. 2, and .!.!_. .!.

Fuel Development Sy stem Studies J. C. Rowl ey , W. R. Pri nce, and R. G. Gi do, "A "Mi ssion Ori ented Advanced Nuclear System Param­ Study of Power Density and Thermal Stress Limita­ eters Study ," TRW Space Technology Laboratori es tions of Rover Reactor Fuel El ements," Los Al amos report 8423-6009-RLOOO , March 1965. Prepared for Scienti fic Laboratory report LA-3323-MS , July the George C. Marshal l Space Fl i ght Center, 1965. National Aeronautics and Space Admi nistration, under contract NAS 8-5371; J. M. Napier, A. J. Caputo, and F. E. Clark, I "Summary Technical Report ," "NERVA Fuel Element Devel opment Program Summary II "Detai led Technical Report ; Mission and Report - July 1966 through June 1972, Parts 1-5," Vehicl e Analysi s," Oak Ri dge Y-12 Pl ant Report Y-1852, Parts 1-5 , III "Parametri c Mission Performance Data," September 1973. IV "Detai led Technical Report; Nucl ear Rocket Engine Analysis," "Technical Summa ry Report of NERVA Program - v "Nucl ear Rocket Engine Analysis Resul ts," Phase I NRX and XE - Addendum to II - NERVA Fuel VI "Re search and Technol ogy Impl ications Devel opment ," Westi nghouse Astronucl ear Labora­ Report," tory report TNR-230 , Addendum to Vo 1. .!.!_, NERVA VII "Computer Program Documentation; Mission Fuel Devel opment, July 1972. Optimization Program; Pl anetary Stopover and Swi ngby Mi ssions," Reactor Design Analysis VII I "Computer Program Documentation; Mi ssion J. A. McCl ary , "Kiwi Reactor Cores Heat Transfer Optimi zation Program; Planetary Flyby and Propel lant Fl ow Studies," Los Al amos Scien­ Mi ssion," tific Laboratory report LAMS-2555, July 1962. IX "Computer Program Documentati on; Nucl ear Rocket Engine Optimi zati on Program."

25 REFERENCES 10 . R. E. Schreiber, "The LASL Nuclear Rocket 1. "Electri c Propul sion Mission Analysis Propul sion Program," Los Al amos Sci entific Terminology and Nomenclature ," National Laboratory report LAMS-2036, Apri l 1956. Aeronautics and Space Admi nistrati on, NASA-SP-210 , 1969. 11. R. W. Bussard , "Nuclear Rocket Reactors A Si x-Month Study Review," Los Al amos Scien­ 2. F. P. Durham, "Nuclear Engi ne Defi nition tific Laboratory report LAMS- 1983, December Study Prelimi nary Report," I-III, · Los 1955. Al amos Scientific Laboratory report

LA-5044-MS, September 1972. 12. c. A. Fenstermacher, L. D. P. King, and W. R. Stratton, "Kiwi Transient Nuclear 3. J. D. Balcomb, "Nuclear Rocket Reference Test," Los Al amos Sc ientific Laboratory Data Summary," Los Al amos Scientific Labo­ report LA-3325, July 1965. ratory report, LA-5057-MS, October 1972. 13. L. D. P. King et al ., "Descri ption of the 4. J. H. Beveridge, "Feasibility of Using a Kiwi-TNT Excursion and Rel ated Experi­ Nuclear Rocket Engine for Electri cal Power ments," Los Al amos Scientific Laboratory Generation," AIAA paper No. 71-639, report LA-3350-MS , August 1966. AIAA/SAE 7th Propul sion Joint Special ist Conference, Sal t Lake City, Utah, June 1971. 14. H. C. Paxton, "Thi rty-Five Years at Pajari to Canyon Si te," Los Al amos Nati onal 5. J. A. Al tseimer and L. A. Booth, "The Laboratory report LA-7121-H, May 1981. Nuclear Rocket Energy Center Concept," Lo s Alamos Scienti fic Laboratory report 15. H. C. Paxton, "A Hi story of Cri tical LA-DC-72-1262, or AIAA paper No. 72-1091, Experiments at Pajari to Site," Los Al amo s Novent>er 29 - December 1, 1972. Nati onal Laboratory report LA-9685-H, March 1983. 6. J. Destefano and R. J. Bahori ch, "Rover Program Reactor Tests Performance Summary 16. J. C. Rowl ey , "Kiwi-A Operati ng Manual NRX-Al through N RX -A6 ," Westi nghouse Astro­ Functional Description of the Reactor," nucl ear Laboratory report WANL-TME -1788, Los Alamos Scientific Laboratory document July 1968. N-3-479, August 1958.

7. R. E. Smi th, "Tabulation of LASL Test Pl ans 17. D. P. MacMillan and A. R. Dri esner, for Test Cel l 'A' and Test Cel l 'C' ," "Post-Mortem Exami nation of Kiwi-A ," Los Al amos Scientific Laboratory internal Los Al amos Scienti fic Laboratory report memorandum J-17-126-70 , June 30 , 1970 . LA-2430 , July 1960 .

8. J. M. Taub, "A Review of Fuel Element 18. V. L. Zeigner, "Kiwi-A ' Operating Manual Devel opment for Nuclear Rocket Engines," Functional Description and Operati ng Los Al amos Scienti fi c Laboratory report Characteri stics of the Reactor," Los Al amos LA-5931, June 1975. Scienti fic Laboratory document N-3-792 , February 1960 . 9. E. P. Carter, "The Appl ication of Nuclear Energy to Rocket Propul sion, A Literature 19. D. W. Brown, "Kiwi-A Pri me Test Seri es - Search," Oak Ri dge National Laboratory Part I: Final Report on the Kiwi-A Prime report Y-931, December 29, 1952. Full Power Run," Los Al amos Scienti fic Laboratory report LAMS-2492, December 1960 .

26 20. D. A. York, "Sull'l11ary Report of Kiwi-A3 30 . E. A. Pl assman, H. H. Helmick, J. D. Di sassembly and Post-Mortem ," Los Al amos Orndorff, "Some Neutronic Resul ts of the Scientific Laboratory report LA-2592, July Kiwi-B-4E Nevada Test," Los Al amos Scien­ 1961. ti fic Laboratory report LA-3327-MS , May 1965. 21. R. W. Spence, "A Prel imi nary Report of the Kiwi-A Tests," Los Al amos Scienti fic Labo­ 31. "NRX-A6 Design Report Vol ume 1 Configura­ ratory report LAMS-2483, February 1961. tion Design and Mechanical Analysi s of the NRX-A6 Reactor," Westinghouse Astronucl ear 22. D. W. Brown and S. Cerni , "Final Report on Laboratory report WANL-TME-1290, October the Kiwi-B-lA Full-Power Run," Los Al amos 1965. Scientific Laboratory report LAMS-2708, Apri l 1962. 32. "NRX-A6 Design Review, " Westi nghouse Astro­ nucl ear Laboratory report WANL-TME-1629, 23. D. W. Brown, "Final Test Report Kiwi-B-lB June 1967. Reactor Experiment," Los Al amos Scientific Laboratory report LA-3131-MS , November 1963. 33. J. C. Buker et al ., "NRX-A6 Test Prediction Report ," Westi nghouse Astronucl ear Labo­ 24. H. J. Newman, K. C. Cooper, A. J. Giger, ratory report WANL-TME-1613, Supplement "Kiwi-B-4D Design Summary ," Los Al amos No. 2, November 1967. Scientific Laboratory document N-3-1536, September 1963. 34. J. Destefano, "NRX-A6 Final Report," Westi nghouse Astronuclear Laboratory report 25. M. Elder, "Prel imi nary Report Kiwi-B-4D-202 WANL-TNR-224, January 1969. Full Power Run," Los Al amos Sc ientific Laboratory report LA-3120-MS, August 1964. 35. C. M. Rice and W. H. Arnol d, "Recent NERVA Technology Devel opment," J. Spacecraft .§_, 26. "General Description of the Kiwi-B-4E -301 No. 5, p. 565, May 1969. Reactor," Los Al amos Sc i enti fic Laboratory internal memorandum N-3-1591, December 19, 36. M. El der, "Prelimi nary Report Phoebus-lA," 1963. Los Al amos Sci enti fic Laboratory report LA-3375-MS, September 1965. 27. M. Elder, "Prel iminary Report Kiwi-B-4E- 301 , "Los Al amos Scientific Laboratory 37. "Quarterly Status Report of LASL Rover Pro­ report LA-3185-MS, October 1964. gram for Period Endi ng August 31, 1966," Los Alamos Scientific Laboratory report 28. A. R. Driesner, "Summa ry of Di sassembly and LA-3598-MS , September 1966. Post-Mortem Visual Observations of the Kiwi-B4E-301 Reactor," Los Al amos Sc ien­ 38. V. L. Zeigner, "Phoebus-lB General Descri p­ tific Laboratory Report LA -3299-MS , Apri l tion," Los Al amos Scienti fi c Laboratory 1965. internal report N-3-1786, March 1968.

29. V. L. Zeigner, "Survey Description of the 39. "Phoebus-lB Disassembly and Post-Mortem Design and Testi ng of Kiwi-B-4E-301 Propul­ Resul ts," Los Al amos Scienti fic Laboratory sion Reactor," Los Al amos Scienti fic Labo­ report LA-3829, September 1968. ratory report LA-3311-MS , May 1965.

27 40 . J. C. Hedstrom, S. W. Moore, and L. J. 49. L. L. Lyon, "Performance of ( U,Zr)C-Graph­ • Sapir, "Current Phoebus-2 Reactor Model ," ite (Composite) and of ( U,Zr)C (Carbide) Los Al amos Scientific Laboratory internal Fuel Elements in the Nuclear Furnace 1 Test memorandum N-3-1740 , May 24, 1965. Reactor," Los Al amos Scienti fic Laboratory report LA-5398-MS, September 1973. 41. "Quarterly Status Report of Los Al amos Scientific Laboratory Rover Program for 50 . "Summary Report CY '66 for Peri od Ending 30 Peri od Endi ng May 1967 ," Los Al amos Scien­ September 1966," Aerojet-General Corpora­ tific Laboratory report LA-3723-MS, June tion report AGC-RN-A-0008 , November 1966. 1967. 51. "Summary Report CY-1969 for Peri od Endi ng 42. "Phoebus 2A Prel iminary Report ," Los Al amos 30 September 1969," Aerojet-General Corpo­ Scientific Laboratory report LA-4159-MS, ration report AGC-RN-A-0012, November 1969. January 1969. 52. "XE-Prime Engine Final Report ," _!_, .!.!_, .!.!.!_, 43. J. C. Hedstrom , W. L. Kirk, and L. J. Aerojet-General Corporation report AGC -RN­ Sapir, "Phoebus 2A Reactivity Analysis," S-0510 , December 1969. Los Al amos Scienti fic Laboratory report LA-4180-MS, May 1969. 53. D. Buden, "Operati onal Characteri sties of Nuclear Rockets," AIAA paper No. 69-515, 44. J. L. Sapir and J. D. Orndoff, "Neutronics AIAA 5th Propul sion Joint Special ist Con­ of the Phoebus-2 Reactor," Los Al amos ference, US Air Force Academy , Colorado, Scientific Laboratory report LA-4455, Ma rch June 9-13, 1969. 1970 . 54. K. R. Conn, "NERVA Engine Operati ng De­ 45. "Peewee 1 Reactor Test Report," Los Al amos scription," Aerojet Nuclear Systems Scienti fic Laboratory report LA-4217-MS, Company , a Di vi sion of Aerojet-General , June 1969. report AGC-110-f-4b-1, May 1972.

46. H. H. Helmick and H. J. Newman, "Design of 55. "Technical Summary Re port of NERVA Program,

a Nuclear Furnace Reactor," AIAA paper No. Phase I: NRX & XE ," Westi nghouse Electric 69-513, AAIA 5th Propul sion Joint Spe­ Corpora ti on, Astronucl ear Laboratory report cial ist Conference, US Air Force Academy , TNR-230 , July 1972. Colorado, June 9-13, 1969. "Engineeri ng Design and Analysis, Techniques and Development," 47. W. L. Kirk, "Nuclear Furnace-1 Test I I "NERVA Component Devel opment and Test­ Report ," Lo s Al amos Scientific Laboratory ing," report LA-5189-MS , March 1973. Addendum to II: "NERVA Fuel Develop­ ment," 48. K. V. Davidson et al ., "Development of III "Full Scal e Test Program," Carbide-Carbon Composite Fuel El ements for IV "Technol ogy Utilization Survey," Rover Reactors," Los Al amos Scienti fic V "Abstracts of Significant NERVA Docu­ Laboratory report LA-5005, October 1972. mentation. "

28 56. "XE-1 Engine Estimated Wei ghts and Bal ance 66. L. A. Booth and J. c. Hedstrom, "A Prel im- Report Revi sed ," Aerojet-General Corporati on inary Study of Open-Cycle Space Power Sys- report AGC-RN-S-0231A, December 1965. terns," Los Al amos Scientific Laboratory report LA-5295-MS , June 1973. 57. J. H. Al tseimer et al ., "XE-Engine Sy stems Analysis Design Data Book," Aerojet-General 67. J. G. Gallagher and L. H. Bowman, "The NERVA

Corporation report AGC-RN-S-0289A, Apri 1 Technol ogy Reactor for a Brayton Cycle Space 1967. Power Unit," American Nuclear Society Trans­ action of 1969 Winter Meeti ng , g, No. 2, 58. "NERVA Engine Reference Data," Aerojet- December 1969. General Corporation report AGC-Sl30-CP 090290-AFl, September 1970 . 68. J. G. Gallagher et al ., "NERVA Technol ogy Reactor Integrated wi th NASA Lewi s Brayton 59. "Dynamic Analysis Report ," Westinghouse Cycle Space Power Sy stems," Westi nghouse Astronuclear Laboratory report WANL-TME - As tronucl ear Laboratory report , WANL-TNR- 2755, January 1971. 225, May 1970 .

60 . "Engi ne/Component Design Status Review, 69. R. E. Thompson and B. L. Pierce, "Gas Cooled Phase II, Engine, Methods and Ongoing Compo­ Reactors for Space Power Pl ants," AIAA paper nents, Presentation to SNSO May 11-14, No. 77-490 , March 1-3 , 1977. 1971," Aerojet-General Corporati on report AGC-N8000R:71-002 , Book 1 and 2, May 1971. 70 . L. A. Booth and J. H. Al tseimer, "Summary of Nucl ear Engine Dual -Mode Electrical Power 61. J. H. Al tseimer, G. F. Mader, and J. J. System Prel iminary Study," Los Al amos Sci en­ Stewart , "Operating Characteri sties and tifi c Laboratory report LA-DC-72-1 111, Requirements for the NERVA Fl ight Engi ne," October 1972. J. Spacecraft, · No. 7, July 1971. ! 71. J. P. Layton, J. Grey , and W. Smith, "Pre­ 62. "NERVA Turbi ne Bl ock Val ve Design Surnnary limi nary Analysis of a Dual -Mode Nuclear Report ," Aerojet-General Corporation report Space Power and Propul sion Sy stem," AIAA AGC -122-f-03-j-l, May 1972. paper No. 77-512, March 1-3, 1977.

63. J. L. Watkins, "NERVA Nozzl e Forging Devel ­ 72. "Fi nal Report Propul sion/El ectri cal Power opment (ARMCO 22-13-5) Report ," Aerojet­ Generation Engineeri ng Operations Report," General Corpora ti on report AGC-N8500-72-141- Aerojet-General Corporation report, AGC-395- f-198, Apri l 1972. 72-1, June 1972.

64. D. W. Tracy , "NERVA Pressure Vessel and Closure Status Report ," Aerojet-General Corpora ti on report AGC-N8500: 72-620-F-096, May 1972.

65. E. A. Warman, J. C. Co urtney, and K. 0. Koebberl ing, "Final Report of Shiel d Sy stem Trade Study ," Aerojet-General Corporation report AGC-S054-023, S054-CP090290-Fl, .!_, July 1970 . .!.!_,

29 SUr+1ARY OF NRX - REACTOR ST RUCTURAL ANOMALIES

Phenomenon Causing/ Resul ti ng Technique , Cri teri a of Source Significant Structural Anomal ies Al lowi ng Occurrence Geometry Change (Corrective Ac tion)

NRX -Al 1) Control-drum binding. 1) Defl ection of control drum 1) Compl ete defl ection analysis of all and outer reflector due to interacti ng components for all reactor thermal gradients. conditi ons indicated. Increase bore diameter in outer refl ector.

2) Pressure vessel -outer reflector 2) Defl ecti on due to thermal 2) Deflecti on analysi s for startup, support ri ng interference. gradients. steady state, and shutdown re quired.

3) Re fl ector crack s. 3) Contraction due to thermal 3) Finite eleme nt analysi s technique was gradients on nozzl e end ri ng. final technique before A6 .

NRX-A2 1) Al uminum barrel tab bent. 1) Combi nati on of barrel pretest 1) None indicated, not considered serious. axial movement, shri nkage, and inner reflector axial contraction.

NRX -A3 1) Support bl ock cracks, aft c' 1) Stresses due to thermal 1) Initiated fi rst finite-el ement computer bore fillet to inner flow gradients and mechanical loads analysis study . (Thi s technique was hol es mi ssed on A2 , same cond. at termi nation of startup ramp . developed extensively .)

2) Control-drum rubbing. 2) Defl ecti on of sector and drum 1) Transient analysi s required , drum due to thermal gradients duri ng materi al removed. cool down after fl ow initiati on.

NRX-A4 1) Cracking of support-bl ock 1) Stresses due to thermal 1) Reduction of start-ramp on A5 , al so fi llets, al so peri pheral and gradients . pl anned test hol ds, al so pa rti al lobe s axial cracks. removed, c' bore radius was increased.

2) Fl ari ng and spl itti ng of 2) Excessi ve growth due to over­ 2) Reduce liner tube leng th . liner tube ends. heati ng .

3) Li ght rubbing of control drum. 3) Defl ection due to thermal 3) Analysi s techniques improved, drum gradi ents. longitudinal sl ots added .

NRX-A5 1) Peri pheral and 1) Stresses due to thermal gradient 1) Seri ous probl em at end of report ing axial support-block cracks. and mechanical loadi ng . peri od.

2) Li ner tube fl ari ng . 2) Cl uster components' overheati ng 2) Reduced line tube length . caused liner tube expansion that coupled wi th tapered sl eeve restraint to cause fl ari ng .

3) 1/3 of fuel el ements broken. 3) Reduced strength due to 3) Increase emphasis on coati ng excessi ve corrosion pl us inflow technol ogy . gradients and end of life.

4) Hot buffer filler stri p 4) Stresses due to thermal 4) El iminate hot buffer. extensi ve breakage. gradients .

NRX -A6 1) Beryllium-refl ector ri ng 1) Stresses due to thermal 1) Components tests veri fy ing 3-D finite cracked. gradients. el ement analysis technique.

2) Support-bl ock cracks. 2) Stresses due to thermal 2) A predictable probl em re quiri ng gradients. extended efforts for similar components in the fl ight engine era of NERVA .

3) Cracking of peri pheral 3) Stresses due to hi gh non­ 3) Nonexpected, protecti on cup composite cups and one symmetri c transverse temperature performance not impaired. tungsten cup . gradient.

30 TABLE I I

ENGINE DESI GN CHARACTERISTICS

Smal l Dimensions Engi ne XE' NERVA Phoebus-2A

Thrust ( kN ) 72 245 337 1123 Specific Impul se ( s) 87 5 710 82 5 840 - -- Thermal Power ( MW) 367 1140 1570 5320 Turbopump Power (MW) 0.93 5.1 6.9 19.4 Turbopump Speed ( rpm) 46 950 22 270 23 920 34 000 Pump Di scharge Pressure ( MP a) 6.03 6.80 9.36 8.54 Engine Fl ow Rate ( kg/s) 8.5 35.9 41.9 129 Chamber Temperature ( K) 2695 2270 2360 2500 Chamber Pressure (MPa) 3.10 3.86 3.10 4.3 Core Diameter (m) 0.89 0.84 0.57 1.39 Core Length (m) 1.32 1.32 0.89 1.32 Be Refl ector, o.d. (m) 1.25 1.12 0.95 1.99 Be Re fl ector Thickness (mm) 117 114 134 203 Pressure Vessel o.d. (m) 1.3 1.3 0.98 2.07 Pressure Vessel Length Im) 1.9 1.9 1. 7 2.8 Pressure Vessel Thickness (mm) 21 25.4 25.4 25.4

TABLE III

ENGINE MASS EST IWl.TES ( kg)

Smal l Dimensions Engi ne XE' NERVA Phoebus-2A

Reactor Core and Hardware 868 1270 3 367 5511 Refl ector and Hardware 569 1467 1 660 2676 Internal Shi el d 239 1316 1 583 Pressure Vessel 150 422 863 1075 Turbo pump 41 182 243 Nozzl� and Sk irt Assembly 224 558 1 051 Propell ant Li nes 15 450 500 2008 Thrust Structure and Gimbal 28 480 663 Val ves and Actuators 207 930 1 262 Instrumentati on and Electronics 159 400 508 Contingency 50 225 600 TOTAL 25"50 7700 IT"J'O"'O"

TABL E IV

BEAR INGS

• One of the few life-limi ti ng components in nonnucl ear subsystem

• Special bearing tests showed wi th Polyhenzemedazol -graphite fabric composite (PBI/graphite) :

- 23 h and 13.8 h for two tests in nonirradiated environment

- 6 h of svccessful operation in three tests irradiated to 10 10 4.8 x 10 and 6 x 10 ergs/gm

• More radiation tolerant than armalon ( tefl on-gl ass fabric)

• Adequate cool ing essenti al to reduce wear

31 TABLE V

VALVE AND ACTUATOR CHARACTERISTICS

TURBINE TURBINE COOLDOlfl NOZZLE PROPELLANT BYPASS SHUTOFF TURBINE CONTROL CONTROL CONTROL SHUTOFF CONTROL CONTROL VALVE DRUM VALVE VALVE VALVE VALVE VALVE ACTUATOR ACTUATOR (CCV) (PSOV ) (TBCV) (TSCV ) (TVA ) (CDA) � Actuator size smal l smal l large large 1 arge 1 arge smal l

Seat diameter, cm 2.54 2.54 15.24 3. 175 11.43

Seat area, cm2 5.06 5.06 182 7_g3 103

Stroke 1. 52 cm 1.52 cm 5.08 cm 1.52 cm 3.18 cm •o .44 rad 3.14

Load max. vel ocity O .15 cm/s 0.15 cm/s 0.5 cm/s 2.5 cm/s 0.2 cm/s 0.007 rad/ s 0.2 rad/s

Seati ng force, N 3550 3558 13 340 2800 10 000

Pl us max . % at 760 275 20 .7 103 103 rated speed , N cm-2

Lead screw diame ter, cm 1.27 1.27 1. 9 1. 9 1.9

Lead sc w advance , 0.0809 0.0809 0.0809 0.0809 0.0809 cm rad-re

Max . ste pi ng rate , 62 62 200 210 75 143 122 steps s-r

Motor gear ratio 50 50 50 10 50 380 917

Torque at speed N-m 3250 34

CCV and NCV are identical .

PSOV and TBCV have same actuator.

THEORETICAL MASS RATIO* ENERGY PHYSICAL SPECIFIC EARTH SO URCE SYSTEM IMPULSE (s) ESCAPE LIQUID HYDROGEN

CHEMICAL 440 15 REACTION ----i� LIQUID OXYGEN > OR FLUORINE H2 SOLID COR E 980 3.2 2750 K HYDROGEN LIQUID H MOL TEN CORE HEATED BY 2600 1.5 FISSION 5250 K REACTOR H+ +e- GASEOUS CORE 6700 1.2 �21000 K DIRECT FISSION 7 THERMONUCLEAR FUSION - ENERGY -- 1 .3 - 3 x 1 0 �1 AN NIHILATION OF MATTER

• RATIO OF TAKE-OFF TO FINAL MASS

Fi g. 1. Compari son of various energy sources for rocket propulsion. Hydrogen heatea oy a solid-core fission reactor offers the potential for more than doubl ing the specific impulse and for reducing by nearly a factor of 5 the ratio of take-off to final mass for earth escape , relative to the performance of chemical rocket en�ines . Specific impulse, I, is also expressed in meters per second. [I(m/s) = 9.8 I(s)].

32 INTERNAL SPECI FICATIONS SHIELD

- - THRUST (N) 334,000 REACTOR SPECI FIC IMPULSE (s) 825

POWER (MW) � 1500 CHAMBER PRESS (MPa) 3.10 CHAMBER TEMP (K) 2360 EXPANSION RATIO 100:1 WEIGHT 11,250 (kg) (W/O EXT SHIELD) NOZZLE ASSY TOTA L OPERATING TIME 600 (min) NO. OF CYCLES 60 (TO 2360 K) RELIABILITY 0.995

Fig. 2. The NERVA and its design specifications.

ACTUATOR DRIVE

SHAFT PRESSURE VESSEL

REFLECTOR CORE INLET OUTLET PLENUM PLENUM

LATERAL SUPPORT SYSTEM CORE SHIELD REFLECTOR SUPPORT PLATE FORWARD END SUPPORT RING

CONTROL DRUM SEAL SEGMENT REFLECTOR BERYLLIUM RING AFT END SUPPORT -r-""""'�L a b RING �---'

Fi g. 3. Cut-away and schematic fl ow descri ption of the NERVA reactor.

33 ENGINE TEST FACILITY REACTOR TESTING FACILITY

Fig. 4. Photographs of some of the maj or testing faci lities built at NRDS at Jackass Fl ats, Nevada, for the Rover program.

TECHNOLOGY DEMONSTRATION

RESEARCH

KIWI B

1955 1960 1965 1970 1975

Fi g. 5. SuITTTiary of the testing program for the nuclear rocket reactors . The two top arrows point to what would have been the development of a fli ght engine had the program proceeded through a logical continuation .

34 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972

I NRX I I NRX-A3 NRX-I A6 � NRX-A1 . 41 <( REACTOR a: TEST NRX-I A2 . (!) 0 ·1NRX-A. 5 a: a.. <( ENGINE XE' > • a: TESTS w NRX./EST • I XECF z I KIWI B1 B KIWI B4D KIWI A3 I . KIWI KII •WI A 1 KIWI B4A• l K1w1 TNT • t .K�IWI B4E'. KIWI B1A 41 KIWII • A' I I PHOEBUS 1A PHOEBUS 18 PHOEBUS :i:: • • u .PHOEBUS 2A a: <( I w I Cl.I w a: PEWEE PEWEE • I

NUCLEAR NF-1 FURNACE • I Fig. 6. Ch ronol ogy of major nucl ear rocket reactor tests.

NUCLEAR 160 T >2500 K FURNACE-1 m FE T 2200 K TO 2500 K FE 140 � <2200 K � TFE Ql 120 ....:i D c: .E 2450 K NRX-A4/EST w 100 :E PHOEBUS- i== 2A

XE PRIME

_____._._ ___ ...... �_._....._ o.._�1964...... __..__...__.__._ 1965 _.__._ ...... 1 966 1967 _.__._1968 _____ 196---.9 , NF-1 _T 1972 CA LENDAR TIME (years)

Fi g. 7. Operating time versus coolant exi t temperature for the full-power reactor tests conducted after develollllent of a successful core design (Kiwi-B4D). Note the general trend toward an increase in both running time and temperatu re .

35 1020 MINUTES 1000

900

800

700

Vl UJ � CUMULATIVE TIME AT POWERS � 600 ABOVE ONE MEGAWATT � z :;:; 500 � � 400 376 MINUTES

300

200

100

0 1964 1965 1966 1967 1968 1970 1971 1972 CA LENDAR YEAR

Fig. 8. Cumulative testing time for all Rover reactor and engi ne system tests .

RECORD PERFORMANCES

POWER (PHOEBUS-2A) 4100 MW THRUST (PHOEBUS-2A) 930 kN HYDROGEN FLOW RATE (PHOEBUS-2A) 120 kg/s EQUIVALENT SPECIFIC IMPULSE (PEWEE) 845 s MINIMUM REACTOR SPECIFIC MASS 2.3 kg/MW (PHOEBUS-2A)

AVERAGE COOLANT EXIT TEMPERATURE 2550 K (PEWEE)

PEAK FUEL TEMPERATURE (PEWEE) 2750 K CORE AVERAGE POWER DENSITY (PEWEE) 2340 MW/m3 PEAK FUEL POWER DENSITY (PEWEE) 5200 MW/m3 ACCUMULATED TIME AT FULL POWER (NF-1) 109 MIN GREATEST NUMBER OF RESTARTS (XE) 28

Fi g. 9. Summary of major performances achi eved in actual tests duri ng the Rover program.

36 8 2 0 1- 5::1 a: Fig. 10 . Reactor mass versus design power for t;; 6 several Rover reactors. The Smal l ::;: Engine was not built but was the resul t of a design study done near the end of �- the program. � 4 � � w 2 a:

*INCLUDES PRESSURE VESSEL, EXCLUDES ANY SHIELDING

DESIGN POWER, GW

PROPELLANT TANK

-FUEL -REFLECTOR ALUMINUM

Fig. 12. Honeycomb assembly machine used to �odel the reactors for neutronic cri ticality experiments. NUCLEAR REACTOR HEAT EXCHANGER

HEATED PROPELLANT

Fig. 11. Schematic descri ption of a nuclear rocket propul sion engine.

37 Fig. 13. The co re being assembl ed in the zero-power mockup of the Phoebus-2A reactor. Figure 12 shows the much cruder simul ation of the same reactor in the Honey­ comb machine.

Fig. 14. Cut-away description of Kiwi-A , the fi rst Rover reactor des ign tested.

1. Nozzle 2. Core support liner 3. Center island 4. Graphite re flector 5. Fuel pl ates 6. Pressure shel l 7. Control rod 8. Lif t band 9. Test stand support ri ng 10. Lock ring 11. Main coolant inlet manifold

38 Slot for

Plate Support Shoulder

Outer Wall

Plate

Plates)

Fig. 15. Fi fth whim of Kiwi-A during assembly. The Kiwi-A design was the only one that empl oyed fuel elements in the form of pl ates.

I 0-= 'b-. D c ( (l . rb'.� I 11..::.

Fig. 16. Cy lindri cal fuel el ements in a graphi te modul e empl oy ed in Kiwi-A' and Kiwi-A3. Si x cyl i nders were stacked end to end in each hole of the graphi te modul e to form a complete fuel modul e.

39 REFLECTOR 1:;;=1��1/ CYLINDER INNER REFLECTOR SEGMENT ��\OR COREi�l

INLET REFLECTOR

Fig. 17. Details of Kiwi-A' core design. The central isl and has been removed .

CLUSTER SUPPORT PLATE BLOCK FUEL ELEMENT

Fig. 18. Fuel-element cluster employed in most of the later Rover reactor desi gns . l.t consists of six, full-length , hexagonal fuel elements supported by a centrally located tie rod. Each extruded graphi te fuel element had 19 cooling channels.

40 ltEFUCTOI llNG

LATERAL SUPPORT SPltlNG OUTER SEAL SEGMENT

Fig. 19. Detai l of core peri phery and lateral support for the NRX-A6 reactor.

Fig. 20. Phoebus-2A being towed on rai ls to its testing station at NRDS. Desi gned for 5000 MW and tested at 4000 MW , it was the most powerful nuclear rocket reactor ever bui lt.

41 Fig. 21. Detail of the regeneratively cooled tie tube empl oyed to support the Phoebus-2A fuel cl usters.

ORE \SUPPORT BLOCKS r rUNLOAOEO TIPS STAINLESS LINER TUBE :mr ·- 113.51 r ' 193 I ..._ . 111.Sl',_ . 197 ,: """ . PRESSURE ' - .•·, .. .. VESSEL PYROGRAPHITE . .. .i! I SLEEVE 9 18 27 36 45 54 72 165 Bt REFLECTOR · .. 63  �1�1� i,..-'- ,:• � ': � - -� ··.-:" � INTERFACE _..... CY LINDER ' - ZIRCONIUM 8 17 26 35 44 53 - 164 179 HYDRIDE . 62 71 � :: MODE RA TOR - :r:. SEAL REGION -'d-1.::.Ml&Ull II ''°"' .... .,,. � . t--...__ SLATS SUPPORT ::.; 7 16 25 34 43 52 61 70 ::::::163 I• 178 ' � ....._ PYROGRAPHITE ELEMENT •! "'- WRAPPER --'1-1-!..i.ri G II •00.04 "' � I'.._ UNLOADED FUEL 6 15 24 33 42 162 !ii 6( 69 . in FILLERS .::z: ELEMENT � :z: 11.12 PEDESTAL '--Ii!: SPIDER ... SHEATH 5 14 23 32 41 50 5l 68 161 t--1 76...... BORON t� !� 1; SUPPORT .. ,__ ROD 4 13 .22 31 49 67 175 40 58 160 4t.2".. ��b � 3 12 21 39 57 66  ,__ PYROGRAPHITE 30 48 ... II 20 29 38 47 65 159 174 """ 2 10 19 28 37 46154 55 64 . . F. ... 1 32.41 ;' Poison 193 ... 1 e2 �m MOLYBDENUM CONE ,,,. Plotts.I ..... m 190 ..... Inlet Plenum 184 ; 22.to I 185 191 Fig. 23. Detai l of the hot end of the Pewee fuel SYpport Plott support system. Sl eeves of zirconium

OD.. hYdri de a round the tie rods moderated � .. � the core neutrons and greatly reduced � � � ' ' s � Q"":Q9!"':�"!��:l!����t:; . � � � �J �g . . the critical mass of uranium in the RAOIUS, cm - core.

Fig. 22. Two-dimensional model of Phoebus-2A reactor.

42 INNER BERYLLIUM CYLINDER POISON PLATE /

UNCOOLED FILLER CONTROL [LEMENT ORUM CYLINDER

E CENTER

Fig. 24. A 60° sector of the Pewee core showing Fig. 25. Detail cross section of the Pewee the three-to-one ratio of fuel and beryl lium- reflector assembly that support el ements. contained nine control drums.

MODERATOR INLET MANIFOLD

MODERATOR OUTLET MANFOLD

fi;r--i!lr-1:1----- CORE TUBE ASSEMBLY BERYLLIUM INNER SECTOR BERYLLIUM OUTER SECTOR

PRESSURE VESSEL BARREL

Fi g. 27 . Cross-secti onal viewof NF-1. The WATER INJECTOR core assembly was water moderated

REACTOR STOOL to reduce the urani um critical mass to about 5 kg. The beryl lium-reflector DRIVE SHAFT assembly was desi gned to be reusable.

Fig. 26. Ax ial view of the NF-1 reactor. Desi gned stri ctly as a nuclear test bed for fuel el ements, it was the last reactor tested in the Rover program.

43 INSTRUMENT4TION TUBE

a

Fig. 28. Detai ls of the NF -1 fuel -cel l assembly.

b

Fig . 29. Detail of the NF-1 water injection sy s­ tem (al so visible at the bottom of Fig. 26) empl oy ed as part of the effl uent cleanup system used to remove fi ssion products from the exh aust hYdrogen ga s.

NRX-A2-A3 1.0 ------

0. 8 - UJ I- I- NRX-A6-XE

0 I I 1965 1966 1967 I 1968 I 1969

Fi g. 30. Progress achi eved in reducing hydrogen corrosion of fuel elements normalized to the corrosion rate observed in the NRX-A2 and -A3 tests . The corrosion rate shown for Pewee-2 is a projection because that reactor was never bui lt.

44 Fig. 31. The hot gas test furnace used for tiYdrogen corrosion tests of fuel el ements in a nonnucl ear environment.

2400 TEMPERATURE Q 10x1022 x- a: w 2000 1ox1023 :'i ------...... I- cc NEUTRON FLUENCE 8 .... <( ::::> 1600 '-. 8 "-N" I- ::!: OR 6 z E <( 3 6 1Cl) cc '2 z cc 1200 FISSIONS/m z I-� 0 <( w ...... 0 <( cc - W O.. , ;;; ::E 4 I- w ::!: ::!: 800 4 ::::> u w ...... !!? u.. w z I- 400 ...... 2 u.. 0 2 z w 0 100 0 0 w 80 I- GRAPHITE FUEL <( 60 COATED WITH NbC cc 50 ___. __. - - --- � -;; 40 GRAPHITE FUEL .o . 30 --- -'N .. - -- COATED WITH ZrC "' E ... - Cl) - 20 ,. .. <( ___ .. - HIGH <:TE COMPOSITE ::EE' - .. �------�FUEL COATED WITH 10 ...... ,_____ 0 ZrC 0 200 400 600 800 1000 1200 AXIAL POSITION (mm)

Fig. 32. Mass loss from Pewee and NF-1 fuel elements versus axial position and reactor environment. The peak in the mass loss curve is the so-called midrange corrosion. The best performance was obtained with the com­ posite fuel coated wi th ZrC.

45 NbC OR ZrC COAT

GRAPH ITE SUBSTRATE

-�@)�­ UC-ZrC DISPERSION

COATED-PARTICLE uc2 PARTICLE MATRIX COMPOS ITE MATRIX

Fig. 33 . Compari son of the fuel structure in the standard , coated-particle graphite matri x with the composite matrix fuel . The continuous , webbed UC-ZrC dispersion prevents hyd rogen , entering through cracks in the top coating, from eating deeply into the graphi te matrix.

10 \ \ 8 \ COMPOSITE FUEL = \ w u 2 c ' 2 \ w 4 GRAPHITE PRELIMINARY MATRIX � FUEL ' ' 2 '

2200 2400 2600 2800 3000 3200 TEMPERATURE (Kl

Fig. 34. Comparison of projected endurance of several fuels versus coolant exi t temperature .

46 --

Fig. 35. View of NRX/EST test assembly on the test pad. This was the fi rst test of a NERVA breadboard power pl ant with all the major fl ight engi ne components functi onally connected.

47 Fig. 36. XE' engine instal led in its test stand. The engine is positi oned in a downward-firi ng atti tude. The te st stand provided a parti al simulation of outer space conditions.

48 TEST STAND ADAPTER ---

UPPER THRUST STRUCTURE TURBOPUMP ASSEMBLY --

,.,___ TURBINE BLOCK VALV E COOLDOWN LINE TURBINE POWER CONTROL VALVE -----�--'dili IF'-----l----

PRESSURE VC:SSEL -----tt-iin_

TUR Bl NE INLET LINE-----,.--n;r--it,

TURBINE EXHAUST LINE----i

PUMP DISCHARGE LINE------1 . , _.

Fig. 37. XE engine design concept.

500 200 20,000

450 180 18,000 500 40

� 400 160 16,000 w er: 400 800 c;; 350 140 �., 14,000 Cl ::> z f- 30 z � w Uzf- u er: 20 -' "' w - 0 ..J N 80 z 8000 ... w 0 ..J :::;: LI.w er: z 0 ...... ::> er: 0 200 200 er: :::;: u f- ::> ... 60 z ... 6000 u 0

50 20 2000

•

0 0 0 0 0 0 10 20 30 40 50 70 100 60 80 90 TIME, ls)

Fig. 38. Ty pical characteri stics of the nucl ear rocket engine startup. Note that chill-down of the various engine components takes about 60 s. Then the engine can be turned on to ful l power at a rate limited by thermal stresses in the core resulting from the temperature tran­ sient (not to exceed about 83 k/s) .

49 e HYDROGEN PROPELLANT VALVE TANK e FULL FLOW TOPPI NG CYCLE

TAN K PRESSU RE VALVE e SINGLE-STAGE CENTRIFUGAL PUMP AND SINGLE-STAGE TURBINE

e REGENERATIVELY COOLED META L-CORE SUPPO RT

ELEMENTS ITIE TUBES)

eRADIATION SH IELD OF BORATED ZIRCONIUM HYDR IDE

e6 CONTROL-D RUM ACTUATORS

VALVE e5 VALVES AND VALVE ACTUATO RS

TIE TUBES e REGENERATIVELY COOLED NOZZLE, AREA RATIO = 25:1 CORE eUNCOOLEO NOZZLE SKIRT, EXIT AREA RATIO = 100:1 REFLECTOR

eUNCOOLED NOZZLE SKIRT HINGED ANO ROTATABLE

NOZZLE eOVERALL ENGINE LENGTH = 3. 1 m 1123 in.) WITH SKIRT FOLDED 4.4 m 1174 in.) WITH SKIRT IN PLACE

e TOTAL MASS = 2550 kg 15620 lb)

Fig. 39. Schematic fl ow descri ption of the Smal l Engine conceptual design. The engine, which required only five control val ves, was designed to produce 72 kN of thrust from a 370-MW reactor.

e PRODUCES 365 MW

e564 HEXAGONALLY SHAPED IUC-ZrCI C COMPOSITE FUEL ELEMENTS

e241 SUPPORT ELEMENTS CONTAINING ZrH NEUTRON MODERATOR

e19 COOLANT CHANNELS PER ELEMENT

ecORE PERIPHERY CONTAINS AN OUTE R INSULATION LAYER, A COOLED INBOARD SLAT SECTION, A METAL WRAPPER, A COOLED OUTBOARD SLAT SECTION, AND AN EXPANSION GAP

eREFLECTOR IS BERYLLIUM BARREL WITH 12 REACTI· VITY CONTROL DRUMS

eCORE SUPPORT ON COLD END BY AN ALUMINUM-ALLOY PLATE. SUPPORT PLATE RESTS ON REFLECTOR SYSTEM

e REACTOR ENCLOSED IN ALUMINUM PRESSURE VESSEL

eCAPABLE OF 83 K/s TEMPERATURE TRANSIENTS

Fi g. 40 . Cross-sectional descri ption of the Smal l Engine concept. The overal l

reactor diameter was 950 mm. The design used ZrH-moderated support elements , as was done in Pewee , to reduce the uranium cri tical mass.

50 FUEL FUEL ELEMENT FUNCTION e - PROVIDED ENERGY FOR HEATING HYDROGEN SUPPORT PROPELLANT ELEMENT - PROVIDED HEAT TRANSFER SURFACE INNER TIE TUBE ZrH DESCRIPTION e MODERATOR 235 - u IN A COMPOSITE MATRIX OF UC-ZrC SOLID OUTER SOLUTION AND C TIE TUBE - CHANNELS COATED WITH ZrG TO PROTECT AGAINST INSULATOR .- H REACT IONS 2

TIE TUBES

FUNCTION e - TRANSMIT CORE AXIAL PRESSURE LOAD FROM THE HOT END OF THE FUEL ELEMENTS TO THE CORE TI E TUBE SUPPORT PLATE SUPPORT - ENERGY SOURCE FOR TURBOPUMP COLLAR AND CAP - CONTAIN AND COOL ZrC MODERATOR SLEEVES

MINIARCH DESCRIPTION e - COUNTER FLOW HEAT EXCHANGER OF INCONEL 718 - ZrH MODERATOR - ZrC INSULATION SLEEVES

Fig. 41. Descri ption of the Smal l Engine fuel module design showi ng the ZrH sleeve in the regeneratively cool ed tie-tube support element.

"SMA LL ENGINE" STATE POINTS AT DESIGN CONDITION PSOV 2 CHAMBER PRESSURE 310 N/cm CHAMBER FLOW RATE 8.51 kg/s CHAMBER TEMPERATURE 2696 K REACTOR POWER 367 MW SPECI FIC IMPULSE 8580 m/s THRUST 7297 N NOZZLE FLOW FRACTION 44.9% TURBINE BYPASS FLOW FRACTION 11.8% TURBOPUMP SPEED 4917 rad/s TU RBOPUMP SHAFT WORK 0.93 MW PUMP EFFICIENCY 65% TURBINE EFFICIENCY 80% NOZZLE VALVE AREA 2.63 cm2 TURBINE CONTR OL VALVE AREA 3.02 cm2 CONTROL DRUM ACTUATOR FLOW RATE PRESSURE TEMPERATURE REFLECTOR STATE POINT DESCRIPTION (kg/sl (MPal (Kl

PUMP INLET 8.51 0.12 17.0 PUMP EXIT 8.51 6.03 19.8 TIE TUBE MANIFOLD INLET 4.05 5.72 20.3 PRESSURE TIE TUBE FIRST PASS EXIT 4.05 5.38 56.9 VESSEL TIE TUBE EXIT 4.05 5.02 428.9 COOLED SLAT MANIFOLD INLET 0.64 5.72 20.3 NOZZLE SLAT FIRST PASS EXIT 0.64 5.24 167.1 SLAT EXIT 0.64 5.02 431.5 TURBINE INLET 4.13 4.86 428.6 TURBINE EXIT MIXED 4.69 4.13 415.6 TURBINE BYPASS INLET 0.55 4.86 428.6 NOZZLE IN LET 3.83 4.63 21.4

NOZZLE EXIT 3.83 4.21 240.4 REFLECTOR EXIT 3.83 4.21 294.9 SHIELD INLET 8.51 4.06 361.0

CORE INLET 8.51 3.96 370.1 UNCOOLED SKIRT FUEL ELEMENT EXIT 8.33 3.10 2728.0 !LATCHED POSITION) CORE BYPASS EXIT 0.18 3.10 370.1

CHAMBER 8.51 3.10 2695.A

Fig. 42. De scription of Smal l Engine operati ng parameters at design conditions.

51 STAGE MASS, USABLE PROPELLANT kg (lb) MASS, kg (lb)

NUCLEAR STAGE 17783 (39205) 12814 (28250) (INCLUDING SMALL ENGINE)

P.ROPELLANT MODULE (18.3 m) 23181 (51105) 21265 (46880)

PROPELLANT TANK

-

4.5

·- -- : I

�''i___ :J, ,J -1�o.9 f-���� . 18.3 -- I ALL DIMENSIONS IN METERS

Fi g. 43. Description of a nucl ear rocket stage employi ng the Smal l Engine and compatibl e with the space and mass constraints of the space shuttl e.

FUNCTION • - PRESSURIZE TH E PROPELLANT FOR THE ENGINE FEED SYSTEM

DESIGN CONDITIONS • XE' NE RV A SMALL ENGINE PUMP DISCHARGE 6.69 9.36 PRESSURE (MPa) 6.03 PUMP FLOW RATE (kg/s) 35. 8 20.9 -41.7 8.5 TU RBINE TEMP (K) 648 154 429 TURBINE FLOW RATE (kg/s) 3.32 19 4.1 SHAFT SP EED (rpm) 21989 23920 46952 CONSTRUCTION (XE) • - SINGLE-STAGE, RADIAL-EXIT-FLOW-CENTRIFUGAL PUMP WITH AN ALUMINUM IMPELLER, A POWER TRANSM ISSION THAT COUPLES THE PUMP TO THE TURBINE, AND A TWO-STAGE TURBINE WITH STAINLESS STEEL ROTORS REACTOR TESTS EXPERIENCE • - NRX/EST 8 STARTS INCLUDING 54.4 MINUTES AT HIGH POWER - XE-28 STARTS/RESTARTS INCLUDING RUNS TO RATED POWER

POTENTIAL PROBLEMS • - SHAFT SYSTEM BINDING AT BEARING COOLANT LABYRINTH IN XE TESTS (INCREASE CLEARANCE AND IMPROVE ALIGNMENT) - BEARING LIFE

Fig. 44. Tu rbopump experi ence and design conditions for the nuclear rocket engine propel lant feed sy stem.

52 __.----ACTUATOR MOUNTING eFUNCTION FLANGE CONTROL HYDROGEN FLOW IN THE ENGINE SYSTEM SPLINE COUPLING

ROLLER BEARING (2 PLACES)

e CONSTRUCTION BINARY VALVES EXCEPT FOR ANALOG CONTROL VALVES AND CHECK VALVES

eREACTOR TEST EXPERIENCE NRX/EST AND XE-PRIME ENGINE SYSTEMS

BUTTERFLY SUPPORT SHAFT

e POTENTIAL PROBLEMS :JllliiP----- THRUST BEARING SEAL DAMAGE BY CONTAMINANTS ERRONEOUS POSITION INDICATIONS LEAKAGE FROM LIPSEAL TOLERANCES FOUR-INCH BUTTERFLY VALVE (TPCV/TBV)

Fig. 45. Ty pical cool ant val ve empl oyed in the nuclear rocket engine.

STAGE PRESSURIZATION LINE

TURBOPUMP

NOZZLE TANK

l>

N CHECK VALVE

TURBOPUMP

Fig. 46. Schematic engine cool ant fl ow di agram showi ng the consequence of avoiding single-point failures in the coolant circuit descri bed in Fig. 39 by addi ng redundant val ves and turbopumps. Some 26 val ves would be required .

53 FUNCTION • - EXPAND GAS TO PROVIDE MAXIMUM THRUST

DESIGN CONDITIONS (NERVA ENGINE) • - THR UST 334 kN - AREA RATIO 100:1 - SERVICE LIFE 10 h - RELIABI LITY 0.9998 - CHAMBER PR ESSURE 3.1 MPa - CHAMBER TEMPERATURE 2360 K - FLOW RATE 41.6 kg/s - COOLANT CHANNEL 28-33 K

CONSTRUCTION • - LH COOLED SECTION TO 24:1 OF ARMCO 2 22-13-5 JACKET AND CRES 347 COOLANT CHANNELS

- GRAPHITE NOZZLE EXTENSION UNCOOLED TO 100:1 - LI-TUBE CONSTRUCTION IN DIVERGENT SECTION

UN RESOLVED PROBLEMS • - USE OF ARMCO 22-13-S APPEARS TO HAVE RESOLVED STRESS PROBLEMS. FABR ICATION PROBLEMS ALSO APPEAR RESOLVED

INLET TORUS

POL INTERFACE NOZZLE EXTENSION

Fig. 47. Experi ence and NERVA design conditions for the nozzl e assembly.

FUNCTION • - SUPPORTS COMPONENTS OF REACTOR ASSEMBLY TO FORM A PRESSURE SHELL FOR THE HYDROGEN PROPELLANT - TRANSMITS THRUST TO THE THRUST STRUCTURE DESIGN CONDITIONS (NERVA ENGINE) • - MAXIMUM FLOW RATE 37.6 kg/s - MAXIMUM PRESSURE 8.66 MPa - TEMPERATURE RANGE 20 - 180 K - RELIABI LITY 0.999997 - SERVICE LIFE 10 h REACTOR TESTS • - NRX-5 TESTS, XE' CONSTR UCTION • - CYLINDER, TOP CLOSURE, BOLTS AND SEALS - ONE-PIECE EXTRUDED FORGING OF ALUMINUM ALLOY 7075-773 - SURFACE COATING M203 UNRESOLVED DESIGN ITEMS ON NERVA ENGINE • - METHOD TO ASSURE BULK PRELOAD - FINALIZE SURFACE COATINGS

Fig. 48. Experi ence and NERVA design conditions for the reactor pressure vessel .

54 TEST CELL •c•

.....__. f'\ CH•c• am·ION

�

\ MILE

Fig. 49. Arrangement of facilities at the Nuclear Rocket Devel opment Station.

EXTERNAL PROPELLANT TURBINES SHIELD VALVES CONTROL CONTROL VA LVES DRUM DRIVE REACTOR L TURBO-ALTERNATOR

�� CONTROL REFLECTOR PROPE LLANT DRUM TA NK

Fig. 50 . Schematic coolant fl ow diagram of an open-cycle, gas-cool ed, nuclear space power plant.

55 MODES TANK HIGH-POWER ROCKET MODE OF 365 10 LIFE, 2600-2660 K e MN, h PROPELLANT TEMPERATURE, NEGLIGIBLE FUEL BURNUP TO e LOW-POWER ELECTRICAL MODE OF 10-25 kW le), Ml'llt), I BOI LER ORGANIC RANK INE CYCLE DRIVEN BY THERMAL ENERGY

FROM STR UCTURA L SUPPORT SYSTEM, USES PROPULSION FROM MODULE SURFACE TO SUPPORT RADIATOR UP TO 10 kW(e) BOILER

ENGINE MO DIFICATIONS PUMP e STRUCTURAL SUPPORT SYSTEM ISOLATION VALVES FOR DOWN LOW-POWER LINE • CHANGE REACTOR DOME AND TIE TUBE CO RE SUPPORT PLATE LINES FOR WIDER TEMPERATURE RANGE OPERATION TO STAIN LESS ST EEL FROM Al, AND ACTUATOR WINDINGS TO VALVE TIE TUBES CERAMIC FROM PO LYI MIDE CORE

LIFETIME REFLECTOR e STUDY SHOW.ED TWO YEARS LIFE WILL NOT SIGN IFICANTLY AFFECT REACTIVITY CO NTROL MA RGIN NOZZLE

• STUD IED IN EARLY 1970's BY J. ALTSE IMER, L. A. BOOTH, "THE NUCLEAR ROCKET ENERGY CENTER CONCEPT" LA·DC·72·1 262, 1972, BASED ON IDEAS OF JOHN BEVERIDGE

Fi g. 51. Schematic cool ant flow diagram for a dual-mode nuclear rocket engine based on the Small Engine design pa rameters . The engine provides the normal high-power rocket propulsion mode, plus a low-power, continuous electric-power generation capabi lity using an organic Rankine cycle driven by a closed loop that cool s the support elements in the core.

7

4

REACTOR �-4<1 BOILER· TIE-TUBE CONDENSER RADIATOR SUPERHEATER I CIRCUIT �,...... -'----' 11 0•47.0 kW I 0•35.7 kW PRIMARY \ 5 RADIATOR LOOP LOOP BLOWER BLOWERS 0.8 q.0.8 q. • 0.255 kW • kW 0 REGENERATOR 0 0.359 0 8.69 kW .iP1= 1.08 N/cm2 .iP• 0.62 N/cm2 Q•

FLOW, TEMP., PRESS., COMPONENT INLET (kg/s) (K) (N/cm2 ) POWER BREAKOOWN (kW)

PRIMARY LOOP- GH2 REACTOR THERMAL POW ER 47.0 1 BOILER 0.0622 555.6 419.7 RAOIATOR THERMAL POWER 35.7 2 REACTOR 0.0622 504.0 420.0 ALTERNATOR OUTPUT 11.0 CONTROL SYSTEM 0.39 POWER CONVERSION SYSTEM - THIOPHENE PRIMARY BLOWER 0.26 3 TURBINE 0.0891 550.0 140.0 RAOIATOR BLOWERS 0.36 4 REGEN-VAPOR 0.0891 438.9 5.02 NET ELECTRIC OUTPUT 10.0 5 CONOENSER 0.0891 354.8 4.85 6 JET PUMP 0.0891 333. 1 4.52 PUMP 0. 178 334. 5 14.90 TOTAL SYSTEM SPECIFIC 7 8 AL TERNA TOR 0.0891 34 2. 1 155.6 MASS 48.6 kg/kW le) a 9 REGEN-LIOUIO 0.0891 346.2 154.8 • 10 BOILER 0.0891 411.6 150.0

HEAT REJECTION LOOP - GH 2

11 RAOIATOR 0.0798 355.2 419.5 12 CONOENSER 0.0798 304 .6 420.0

Fig. 52. Design operating parameters for a 10-kWe organic Rankine power-generati ng pl ant to be used wi th the Small Engine.

56 RADIATOR

TURBO-ALTERNATOR HEAT ; :::: WP ) EXCHANGER

• L- PROPELLANT VALVES __...... , / CONTROL J/ VALVES TURBO-ALTERNATOR (HIGH POWER) �/,

REFLECTOR PROPELLANT EXTERNAL TA NK SHIELD

Fi g. 53. Schematic diagram of a bimodal power-generating plant based on the Rover reactor design. It would have an open-cycle, high-power generati on mode plus a con­ ti nuous, closed-cycle, low-power generation capabi lity.

· �OZZL� PLUG·-· . . CALTERNATE MODE> ROCKETr-r MODE • ·- /�-�;s:=f- =t GAS FLOW .. 1 DIRECTION + ///d

.. I ,,...---·· ,l.f: �),;;1 . ��:,..._ 1 11r�OZZLEROCKET PLU(MODE . ALT. �ODEl GAS� r\\ 1 . FLOW DIRECTION �\· \

•

Fig. 54. Poppet valve design used to switch from a rocket propul sion mode to a power generation mode.

57 Printed in the United States of America Available from National Technical Information Service US Department of Commerce 52S5 Port Royal Road Springfield, VA 22161

Microfiche (AO I)

NTIS NTIS NTIS NTIS Page Range Price Code Page Range Price Code Page Range Price Code Pase Rans• Price Code

001-025 A02 151-175 A08 301-325 Al4 45 1-475 A20 026-050 A03 176-200 A09 326-350 Al5 476-500 A21 05 1-075 A04 201-225 AJO 35 1-375 Al6 501-525 A22 076- 100 A05 226-250 All 376-400 Al7 526-550 A23 JOl- 125 A06 251-275 Al2 401-425 AIS 551-575 A24 126-150 A07 276-300 Al3 426-450 Al9 576-600 A2S • 601-up A99

•contact NTIS for a price quote. >·- (/') {. UJ r ' .... _"' .... _ ;

. (, ) f' . (_�