STEMS LIBRARIES T 3 4456 D3L1LaY & AEII RESEARCH A CENTRAL RESE A H. Co J i} d""*vLJ-LL.Ln,‘. - DOCUMENT COLLECTION MENT REPORT c-e. “Reacrors-steinf 7 Features of Aircraft Reactors ART FUEL PUMP AND XENON REMOVAL SYSTEM DEVELOPMENT, TEST EVALUATION, AND AIRCRAFT APPLICATION G. Samuels M. E. Lackey 5 3 ‘ fl A BECLASSIFIED CLASSIFICATION CHANGED To: * x % (,S i SR, E./ e o e e i dem - % - By Ayt “iou[ TY OF:_ _Q'Q.\__C- ____: — e .._._._(L"...,B Lot o e By: .. fil@d..(_x.?b_m_é- el OAK RIDGE NATIONAL I.ABORATORY OPERATED BY UNION CARBIDE NUCLEAR COMPANY Division of Union Carbide Corporation vee Y POST OFFICE BOX X * OAK RIDGE, TENNESSEE WL';EGAL‘_‘NOTI_CE Th!s reporr was prepared as -an account of Government sponsored work Ne lrher nor the Commlsslon, nor: any’ person octihg on. behalf of the Commlsslon. . a5 A, Makes ony warrunly or représentation,- express or lmplled wufh respec, completeness, or usefulness of the’ informotion contained ‘in’ H-ns ‘repo 't, oF rhut fhe Use .uny mformohon -upporotus, rnefhod .oF process dlsclosed in ‘this rep)orf moy: . not |nfr| . prwcnely owned rights; or A B. Assumes any hol:nlmes wn'h respect fo the use of, or for dumages resulf any mformohon, opperufus method or process disclosed in this reporf. ' As used |n the obove, person oclmg on beholf of the Commlsslon or' distributes, or prowdes access to, any ‘information pursuant f6. hls employmenf or confruci wuth the Commlss:on. ) : Coe s E3 LT 1R 1 8 ORNL-2376 e C~84%, AIRCRAFT REACTORS. This documlent consists of 186 pages. ) _‘Copy':7éof 205 copies. - Series A. Contract No. W-7405-eng-26 ;-_A:'Lfcre.ft Reactor Engineering Division ART FUEL PUMP AND XENON REMOVAL SYSTEM DEVELOPMENT | TEST EVALUATION, AND AIRCRAFT APPLICATION e G. Samuels . M. E. Lackey | DATE ISSUED —DEC 194957 OAK RIDGE NATIONAL LABORATORY Opera‘ted. by UNION CARBIDE NUCLEAR COMPANY A Division of Union Carbide and Carbon Corporatlon o Post Office Box X - C ‘ Oak Ridge, Tennessee ‘ /I//WIT//I/fl//////i//fii/?/?]/fl/fll/?”ifl/?il??filfifi 3 wuse | ! | e 1. S. 2. H. 3. D. L. F. 5. E. 6. A. 7. C. 8. G. 9. E. 10. R. 11. D. 12. C. 13. R. 14, W. 15. B. 16. W. 17. F. 18. C. 19. L. 20. A. 21. J. 22. D. 23. W. 2L. B. 25. W. 26. A, 27. E. 28. C. A. H. A G. W M. 78-80. 81-82. 83. 8k, 85. -1i- INTERNAL DISTRIBUTION E. Beall 35. W. Bertini’ 36. S. Billington 37. F. Blankenship 38. P. Blizard 39. L. Boch 4o. J. Borkowski h1. E. Boyd: 2. J. Breeding b3, B. Briggs Wi, W. Cardwell L5, E. Center (K-25) L6. A. Charpie L. G. Cobb L48. Y. Cotton k9.~ ‘B. Cottrell 50. L. Culler 51. W. Cunningham 52, “B. Emlet (K-25) 53. P. Fraas 5k, H. Frye, Jr. 55. E. Ferguson - 56. T. Furgerson 5T. ‘L. Greenstreet 58. R. Grimes - 59. G. Grindell. 60. Guth 61. S. Harrill 62. S. Householder 63 -6k, W. Hoffman ' . Hollaender - 65-Th. W. Keilholtz B T5. H. :Jordan T6-7T. T. Kelley : EKTERNALIDISTRIBUTION Alr Force Ballistic Missile Div151on AFPR, AFPR, AFPR, AFPR, Boeing, Seattle Boeing, Wichita . _ Lurtiss-Wright, Clifton Douglas, Long Beach 4 e = (')I.?:IC.'):fl'UI‘.l:ILq'J> EHPPHARGPEORTANEE DGR ORNL-2376 C-8k - Reactors-Special. Features of Aircraft Reactors M-3679 (20th ed. Rev.) E‘:I"UUZU:JN‘?N.L!UUSSUID—J:U{Zt‘“'dt\'l<fi:>fl?clm_‘h>l?d Lackey Lane Livingston Manly Mann Mann MacPherson Megrehblian Morgan Murray (Y-12) Nelson . Perry "McNally ‘Robinson amuels Savage . Savolainen Schultheiss . Shipley Skinner Snell Swartout . Taylor Trauger Weinberg . Whitman . ' . Wigner (consultant) Winters ORNL - Y-12 ‘Technical Library, Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. Central Research Library a 86-88: 90- 91 92, - 93. - 9k, 95. 96. 97-110. 111. 112. 113-115. 116. 117. 118. 119. 120-125, 126. 127. 128. 129, 130. 131. 132, 133. 134, 135, 136. 137. 138. - 139-142. 143, 14k, 145, 146, 147, 148, 1h9 - 150. 151.. 152, 153. 15k, 155. 156. 157. 158, 159-162. 163. 7 - ‘ : e SADPISE -1ife - o R AFPR, Douglas, Santa Monica . AFPR, Lockheed, Burbank AFPR, Lockheed, Marietta ' AFFR, North American, Canogda Park AFPR, North American; Downey Air Force Special Weapons Center Air Research and Development - Command (RDGN) ‘Air Research and Development Command. (RDTAPS) Air Research and Developmeént :Command (RDZPSP) Alr Technical Intelligence Center Air Unlver31ty Library = .- ‘ -ANP Project Office, Convair, Fort Worth Argonne National Laboratory Armed Forces Special Weapons Project, Sandia _ Armed Forces Special Weapons Project, Washington. Assistant Secretary of the Air Force, R&D - -~ Atomic Energy Commission, Washlngton Bettis Plant (WAPD) Bureau of Aeronautics Bureau of Aeronautics General Representatlve BAR, Aerojet-General, Azusa = BAR, Convair, San Dlego BAR, Glenn L. Martin, Baltimore BAR, Goodyear Aircraft, Akron . BAR, Grumman Aircraft, Bethpage Bureau of Ships ' Bureau of Yards and Docks Chicago Operations Office Chicago Patent Group Curtiss-Wright Corporation General Electric Company (ANFD) Hartford Area Office Idaho Operations Office Knolls Atomic Power Laboratory Lockland Area Office Los Alamos Scientific Laboratory- Marguardt Aircraft Company ' National Advisory Committee for Aeronautics, Cleveland National Advisory Committee for Aeronautics, Washlngton Naval Air Development Center Naval Air Material Center Naval Air Turbine Test Station Naval Research Laboratory Nuclear Development Corporation of America ] Office of Naval Research 0ffice of the Chief of Naval Operations (OP 361) Patent Branch, Washington : Pratt and Whltney Aircraft Division San Francisco Operations Office et vt . ) L o Carae ot e T . - s | | - ~iv= 16k4. Sandia Corporation 165. Sandia Corporation, Livermore 166. School of Aviation Medicine . 167. USAF Headquarters 168. USAF Project RAND 169. U. S. Naval Radiological Defense Laboratory 170-1T71. University of California Radiation Laboratory, Livermore 172-189. Wright Air Development Center (WCOSI-3) 190-204. Technical Information Service Extension, Oak Ridge 205. Division of Research and Development, AEC, ORO ..v.. CONTENTS Abstract . e+ e * & & 8 e e s ._~-o o _..;.o' -o LI o e @ -'. ._. . . . . Section I - DESIGN CRITERTA. ,'..; e e eeve e e e e e e - Precepts and Requlrements . e s e e e s e . ,';fl.g{ . « v Tllustrations and Nomenclature. . . « « & s o o o o s o ow'e o 4 Circuit Description . . ® . * * .u * . - . » 7:0 » .. e . ‘n ‘e u o Primary Fuel Circuit + « o e o o ¢ o o o0 o o o o o o o o BypaSS Fuel Circuito . _» o s & o s+ 2 s e & . ‘® e e & & o @ _. Dynamic Seals 0 Ld 'C' . . e * . . . . * * . * . . * .. . e - . Section II‘ DESIGN'AND DEVELOPMENT OF ART FUEL PUMP AND XENON REMOVAL SYSTEM_ Xenon Removal System Design « « + + «-.+ . ..;“.‘..f‘. .-, . . - Xenon Removal System FLOW. . o o« & o o o o o 0 s o o o o o - Xenon Removal System Gas Removal e« o o o .. I ART Fuel Pump Development . . . . ctle e e e e e e Comblned.ART Fuel Pump and Xenon Rémoval System.Develqpment . SYStem Pressure Fluctuations’ . ;f.'.-. D ee e e e e e SyStem Pressure Level R ,ou-c- ¢ el e e . ¢« o o _-o) s o o o ¢ Bypass Flow Determination &nd Control. « o o o o & 4 ¢ v o | Fuel Leakage Past Upper Slinger Seal . . o o o s s o o e - Experiment Results of Combined System Tests.. . « « & « .o &« s & ~ ART Fuel Pump Cavitation Characteristics . . . « « . . . . Section III' DESIGN OF AN ATTITUDE-STABLE XENON REMOVAL SYSTEM. .« + 4 o o+ - . CiI‘CU.lt DeSCI‘:Lp'thIl - s e e e e [ e s e e ’ ¢ . s w | . . o e System Pressure Control. c s e el e e e e e e e e e Prlmary Serlo s 6 e e e % e e .+ o + e ‘e s s e e s @ - ‘ Second.ary SW].I‘J.. . o . . s . . » ol, - . * & '! . * . .0 .o . . . Dynamic SealS. o« « o ¢ ¢ o o e s s 0 e e 000 e e s e " Development Tests .. e e e e e e e e e e e e e e e e e SyStem AnalySiS * . -l r » . . L) ¢ . LI * e+ v . * ¢ * . '.‘ 13 13 Y 19 . 19 2l 2 , 2k . 45 ks » 51 66 - IO . 195 . 15 go:. . Bo . B3 . 83 91> -vi- Contents (cont'd.) Appendix A SUMMARY OF WORK LEADING TO THE DESIGn;QF PUMPS FOR THE CFRE,‘ N, (0} | Appendix B | CENTRIFUGE THEORY AND C.ALCUI.I.ATIONS . “- . . s e . ’i. | . e o.- . 0 . . e . . - :’l]_':s Appendix C -Appendix D CALCULATIONS FOR THE ART FUEL PUMP IMPELIER REDESIGN . . . % o . . . . 1133 Appendix E BYPASS FLOW CALCULATION AND EXPANSION TANK HEATING . . + o » « « . . . 1145 | ,_Appendiqu | | RADIOACTIVITY IN THE ART FUEL PUMP OIL SYSTEM o o v « 4 o « + o o o o . 2I55 .Appendix.G Xe™32 POISONING + « ¢ v v v v v v v v e vt v e e e v e e e . 2165 ACmOWIEDEWMS ‘. " = ‘ ° » . B » . . » » - . * - . "‘ . 3 * ., . L J . ‘ . * . a 9‘1«73 REFERENCES . o ¢ o .‘.' v .. ‘e c‘ t e 2 e vo .c LI T ol - s e e s . c. . -‘o 17)-" Figure 1.1 1.2 1.3 1.4 1.5° 2‘.1 2.2 2.3 'e.h. 2.5 2.6 2.7: 2.8 2.9u $2.10 2.11 2.12 2,13 2.1k —Vll— LIST OF. ILLUSTRATIONS ART Bypass Fuel Circuit « o« o« o« oo o &+ o o System .« . ' Title' | Vertlcal Sectlon Through Reactor Assembly 'ART Fuel Pump , - Schematic ART Fuel Clrcuit.‘, “ i e s e W1th Model 32 Impeller ;_.; " ART Primary Fuel CLrcuit v . « o o o o Brass ART Fuel Pump Impeller e s é- 9 s @ ART Fuel Pump Experiment ,Experiment ART Fuel Pump Experiment ART Fuel Punip Experiment ART Fuel Pump Experiment ART Fuel Pump Experiment_ ART Fuel Pump Experlment_ ART Fuel Pump PExperiment ART Fuel Pump Experiment .. ART Fuel Pump - Experiment . ART Fuel Pump _ Experiment . Corrections Applied to the ART Fuel Pump " Total Discharge Pressure for Obtaining the Centrlfuge Performance CharacterlstlcsI ,No. ART Fuel Pumpf l-T&bleel..... Performance Characterlstlcs, NO. 2 -— T&ble 2 l ‘- e e " ® - @, Performance Characteristics NO.3-—T8.bl€21. e s s e Performance CharacteristiCS‘-' No. b - Table 2.1 . . Performance Characterlstlcs‘ Nos. 3, 5, and 6 - Table 2.1 Performance Characterlstlcs‘ NO T - Table 2 l - . * '; .o' Performance Characteristlcs- Performance CharacteriStics Performance Characteristics NO. 13 - T&ble 2.1 o.o - .. Performance Characteristics. NO. ll" - Table 2.1 e e 0.8 Performance :Characteristics Statlc Discharge Pressure . . . & + . Frow w2 . FulI Scale Plastlc Model of the Xenon Removal Page 10 11 12 1k 15 - 20 25 3% 133 EN 35 36 37 38 39 4o b1 ko 43 Figure 2.15 2.16 2,17 2.18 2.19 2.20 2.21 2.22 2,23 2.2k 2.25 2.26 2.27 2.28 . 2.29 P - S -viii- Tftlé . ART Fuel Pump Volute Performance Characteristics . Comblned Xenon Removal and Fuel Pump Hot Test- Loop ART Fuel Pump Bypass Flow Characterlstlcs with Model 32 Impeller at a 3-1/2 in. Depth of - Liquid (Water) in the Expansion Tank . . . . . Upper Slinger Seal with Radlal Vanes . « + « « .+ Upper Slinger Seal with an Axial. Step R ART Fuel Pump Performance Characteristics with Model 32 Impeller at a 1/2 in. Depth of Liquid (Water). in the Expansion Tank. . . . . ART Fuel Pump Performsnce Characterlstics Vith Model 32 Impellef at a1l in. Depth of Liquid (Water).in the Expansion Tank ... . . . ART Fuel Pump Performance Charactéristics'with Model 32 Impeller at a 2 in. Depth of Liquid (Water) in the Expansion Tank c e e e ART Fuel Pump Performance Characteristics with Model 32 Impeller at a 3 in. Depth of _ Liquid (Water) in the Expansion Tank ... . ART Fuel Pump Fluctuation Characteristics with Model 32 Impeller at a 1/2 in. Depth of Liquid (Water) in the Expansion Tank . . . . . ART Fuel Pump Fluctuation Characterlstlcs W1th Model 32 Impeller at a 1 in. Depth of _' Liquid (Water) in the Expansion Tank - . . . ART Fuel Pump Fluctuation Characterlstlcs with Model 32 Impeller at a 2 in. Depth of - Liquid. (Water) in the Expansion Tank . . . . . ART Fuel Pump Fluctuation Characteristics with Model 32 Impeller at a 3 in.: Depth of Liquid (Water) in the Expansion Tank . . . . . Calculated and Experimental ART Fuel Pump Bypass Flow Characteristics . . . . « « &« o0 .o « . Cavitation Characterlstlcs of ART Fuel Pump W1th_ Model 32 Impeller at a Pump Speed of 2400 rpm and a 3 in. Depth of quuld (Water) in the Ex- ' pan31on Tank . '.°'. . e s 4 e o e s e s e s s Page L 46 50 35 57 58 "55 o & 6 62 63 6L 65 .68 Figure 2 .’3‘-0- 2.31 2.32 3.1 3.2 3.3+ 3.k 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 To. - ix;- . Title CaV1tat10n Characterlstlcs of ART Fuel Pump with ~ Model 32 Impeller at a Pump Speed of 2700 rpm ~and.’a 3 in. Depth of quuld (Water) in the Ex- ~pansion Tank . . . . . “ e e e e e e eo e Cavitation Characteristics of ART Fuel Pump. W1th Model 32 Impeller at a Pump Speed of 3000 rpm and a 3 in. Depth of Liquid (Water} in the Ex- pansion Tank . . o o« o o o o s o o .6 = o o Minimum Expan51on Tank Gas Pressure Required to Prevent CaV1tatlon in the ART Fuel Pump e o o ' Attltude-Stable Xenon Removal System e o e e e Attltude—Stable Xenon Removal System Rotary Assembly Test Model No. 15 e e e e e eTe Schematic of Attltude-Stable Fuel Circuit: . . Attitude-Stable System Pressure Control . . .. . Attitude-Stable System Primary Bypass. swirl . . ' Attitude-stable System Secondary Bypass Serl . Attitude-Stable System Pump Shaft Shroud . + . . _AttitudeAStable System Centrlfuge,Sllnger Seal.. . Attitude-Stable System Swirl Chamber.Energy' Balance c s e s s s 5 e s e e-3 o o e e e v e Attitude-Stable Xénon Removal System Test Mbdel . NO. 7 . ' ‘.' I . * * . e o o o . o . o 7‘ * .o * Attltude-Stable Xenon Removal System Test Model No ll . . o 0 . @ - 9 ‘e . . .- o » O G o . * 9 Attltude-Stable Xenon Removal System Test Model NO. 15 - - - . . . ° . . . . ® - ° o s e o . Attltude-Stable System Maln Fuel Clrcult Charac- teristlc ‘. - > - . . L ] o > o . o 'D T e o e - & -Attltude-Stable System Bypass Clrcult Character-:' lstic s & s & & e o 8 o -0 e & & e @ u"o e e » fAttitude-Stable System Parallel Main Fuel and BYP&SS FlOW ’Cl'l&raC'teriS'biCS e o o5 o @ ‘e e s ! 2 Fes ‘Page 69 Y . 178 179 181 182 . 184 ;Bs, 187 189, 159 100 101 101 . 1101 Figure A.l A.2 A.3 B.1 B.2 B.3 C.1 c.2 C.3 | C.4 C.5 C.6 Dll D.2. D.3 D.ha D.kb -X - “Title | .Prellmlnary Model No. 1 of an Attltude-Stable . Xenon Removal System . v « o v o e o0 . Prellmlnary Model No. .2 of an Attltude-Stable - Xenon Removal System e s s s e s ee e-e- e Preliminary Model N6. 3 of an Attitude-Stable . Xenon Removal System « « o ¢ o ¢« o o o« & Correlatlon of Drag Coeff1c1ent and Reynolds Number for Bubbles inWater . . « i « ¢ Effects of Bubble Radlus and Exit Veloc1ty on Centrlfuge Effectiveness at 2700 rpm. W1th ‘a Cup Radius of 2-7/84An. v 4 o e o , Effects of Bubble Radius and Exit Velocity on Centrifuge Effectiveness at '1350- rpm with a Cup Radius of 2-7/8 in. . . . . . ..o o | Head Characteristlcs of a Fluid Rotatlng as a SOlld Body * L o . . . - » a . - a 0 . . . Leakage Characterlstlcs of a Dynamlc Seal W1th an -Axial Clearance of:0.050 in. . . . « & Leakage Characteristics of a Dynamic Seal‘fiith - an ‘Axial Clearance of O. 128 iNe v o o o & Leakage Characteristics of a Dynamlc Seal with an Axial Clearance of 0.248 in. . . . . - Leakage Characteristics of a Dynamic Seal W1th an Axial Clearance of O. 481 in. . . . . . Zero Leakage Characteristics of =a Dynamic Seal Pump Inlet Velocity Diagram . . + . +/¢ « o . Redesigned ART Fuel Pump Blade Trace and Fluld Passage Layout « ¢« ¢« o o ¢ ¢ o o o & 2 0 & Veloc1ty Proflle Entering the ART Fuel Pump Impe lle r ® - e ® . *® . . e . *® * . ‘o . . * Plane Parallel to Fuel Pump Shaft Centerllne Plane Perpendlcular to Fuel Pump Shaf't Centerllne .Pagé 107 110 118 119 120 125 | 126 . 127 128 129 130 133 135 136 140 140 Figure D:hg D.5a D.5b . D.5¢c D.5d E.3 E.4 G.1 G.2 G.3 - G.l _,,,:' | o B~ L i 2 -xi- Title '.Fluld Streamline along Fuel Pump Shroud .« + + + 4 Blade Trace e * s+ s+ s s 0o . o . e . c s 0 s . ¢ o e Fluid P&SS&SG o> . . . . . . o 'o . .-- -. o @ . ° . . e Distance from Tip Along Reference Radius . . . . Blade B:Lanl: . a. . . o e . s e -"' 9’ oo! o ..._,. _o' o Calculated ART Fuel Pump Bypass Flow Charac- terlsthS c - e e . o . e o * e & ~®- @ 0"'. .. . ART Reactor Fuel and NaK Temperatures at a Constant Ppmp Flow of 645 gpm o o v o ¢ o0 o o o « o o & | The Effect of Reactor Power Level on the ART Ex- - pansion Tank Liquid Level o o o o o e o sive o s Mixed Mean Temperature of Fuel in the ART Expan51on Tank as a Function of Reactor Power to Fuel and Pump Speed at a Constant Fuel Flow of 645 gpm per PUlpP o o o o o o o o 0 o o0 o0 o 0 s o a Schematic of Flow Through Sparger and.Expan51on T&I].k ... o"o ® e s e e e e e 0 & & & & o @ b & e Xenon Solublllty in Fuel-30 « « & o v o o o o o v Xenon Concentratlon at Equlllbrlum 1n Fuel 30 with a Power Generation of 60 MW v « & & o « « & ,“Xe135 Poisoning in Fuel-30 at 60 Mw Power and a Helium Bleed of 1000 Liter/day « « + « « « « « . | Page 1ho . 1h1 . Ah41 L AhT 41 147 148 . 149 150 . 169 a70 171 172 _‘-xii- o LIST OF TABIES 1.1 Design Requirements « « o o « o o o o o oo o o s o T 1.2 Design DBte « o o o o o o o o o 0 e s 0 0 oo o o u 9 . 2.1 : Conditidns and Resuits of water:PErformafice | Tests of ART Fuel Pump « + o « o ¢ ¢ 4 ¢ o o o o o _29. 3.1l Model'lS - PErforfiance Evaluafion e s s e e e 193 3.2 | Séquéntial Develdfimént of Aftitfidé{Stgble Xenon Removal System « ¢ ¢ o o o ¢ o 0 0 o o o o 0 o o gl 33 . Mode% TmffifErqumance Evaluation T:_i'g..'..; - (9. 3.4 | Modei”lfi - Dimensions . .. .1;-. . ;‘: .l. o s e e s 198 F.l Concentration and Activity df“Kryptofi, Xenon and Daughter Products in Expansion Tank and 0il _ _ System for a Helium Flow Rate of 1000 liters/day. . 162 G.1l Xenon Concentration in Expansion Tank Gas J VOl'(me 0700?_0'.0000-{1o-o:.oo.‘-oo.. I68’ ABSTRACT This report summarizes the design and development of the fuel pump and xenon removal system for the Aircraft Reactor Test. The system is comprised of two fuel pumps-with an expansion tank mounted just above énd between them, and a complex of impellers, seals, and stators arranged to be a part of the removable fuel pump assemblies. The xenon removalisystem is designed to bleed appfoximatély 1.5% of the main fuel flow rate'ihto the expansion _cham- ber through the xenon removal equipmenf; The stripped fuel is returned to the main system by way of centrifuges'mounted immediately in back of the fuel pump impellers. The .centrifuges serve to filter out any entrained bubbles and pro#ide in conjunction with seal impellers a suitable main fuel system pressure level. | A | _The initial design and development work included as an obJjective the achievement of a system insensitive to attitude. That work is included in this report as an aid to future designs. This report inclfides the following: :l. -Thé‘undérlying design precepts apd requirementé upon which the development was based. 2. The functioning and principle of operation of the finalized design. - | 3. The chronological evaluation of the finalized design and the major design changes in the development. ' 4. Discussions and calculations pertinent:to five design criteria: | . | a. pump impéller deSign' b. system pressure control c. swirl stability. d. bubble removal in the centrifuge e. seal requirements and performance, H— . A Section 1 'DESIGN CRITERTA DESIGN CRITERIA Precepts and Requirements " The high absorption cross section of Xe_35 makes 1t extremely desirable to strip xenon continuously from all or a fractlon of the fuel¥* 1n high power j density reactors. Xenon can be stripped from liquid fuels by éxposing the- fuel to a gas. interface where the xenon molecules can diffuse from the - liquid. The rate of xenon extraction from fuel at a given temperature 1s a function of the xenon concentration in the fuel the partial pressure of xenon in the . contacting gas, and the rate of exposure. Thus, a xenon removal, or Y quirements The pump should have the highest possible degree of reliability'which _ necessitates large radlal and. axial clearances for mov1ng parts to mlnlmlze the effects of thermal dlstortlon. B s The pump shaft seal and bearing must. operate at temperatures below 300° F. The pump should not constltute a hole in the shleld, 'The pump should be replaceable without removal of the reactor ShlEld. The bearing and seal should be sufflclently shielded frcm radlatlon to give a satisfactory iubricant 1ife. Cooling prov151ons must be adequate to remove the heat generated by rediation at full power and at the same tlme no areas contacted by llquld fuel may drop below the fuel free21ng p01nt at zero power. -The pump-mnst.supply theé requlred fuel system head and flow. The pump must be free from cavitatlon within the requlred operatlng | pressure level. The pump must meet the reactor geometry demands. Pump, casing, and expansion tank should: be rugged in construction, un- _ likely to be subgect to thermal distortion, and have a.mlnlmum of moving parts. ' ' Operation should not be sensitive to variations in pump,speed or to differences in speed between pumps, includiné one pump stopped. The pump shaft power required should be minimized. Fuel entry into the cool region around the impeller shaft below the seal must be prevented. The system should be ‘such as to drain naturally when in its normal position. Provision must be 'made for removing xenon with a high degree of reli- aebility and sufficiently thoroughly to provide a substential margin over the concentration acceptable from the reactor control standpoint. 16. 17. 18. 19. 20. el. 22. 25« To. effect gas removal from the fuel & fraction should be recycled from the main'cirCuit through a ¢hamber,with a helium atmosphere. To 1ncrease the dlssolved gas removal effectlveness, the fuel in the gas removal chamber should be agitated to increase the helium entraimment thereby increasing the surface area and reducing the diffusion path length in the fuel. _ | The stripped fuel must be returned to the main fuel circuit free of gas bubbles. o To prevent bubbles from enterlng the core the return c1rcu1t must be designed without leakageo .Thejgystem pressfire must be maintained at a predetermined lével§ An expansion volfime must be provided for the fuel. Clrculatlon.must be maintained through the expansion tank to avoid over- heatlng from fission product decay heat. Reactor off-gas and fuel vapor contamingtion of the pump lubrlcatlng oil should be & minimum. Table 1.2 - Design Data Fuel No. 30 Density'(#/ft3) Surface Tension (dynes/cm) Viscosity (#/ft-hr) Liquidus Temperature Water Density (#/f£3) Surface Tension (dynes/cm) Viscosity'(#/ftfhr) Pump Shaft Spacing (in.) Fuel Pump Impeller 0.D. (in.) Fuel Pump Impeller I.D. (in.) Total Fuel Volume Outside Expansion Tank (ft3) Change in Expansiqn-Tank - Fuel Volume from Fill Temperature to Operating Temperature (in.3) Bleed Flow Rate per Pump (gpm) Value = 246.4-0.03227(°F) 530°C-157 630 -132 730 -115 1100°F-21.3 1300 . -12.8 1500 =8.5 968°F - 62.3 68°F-72.8 68°F-2.42 21 2.7 3.5 8.78 568 12 Source (2) (2) (2) (2) (3) (1) ) Calculated Calculated (1) FUEL PUMP CONTROL ROD Na EXPANSION TANK NaK FUEL EXPANSION TANK HEAT EXCHANGER ASSEMBLY - Na-TO-NaK HEAT EXCHANGER 4= Na INLET REFLECTOR ASSEMBLY “ 2 - 3 THERMOCOUPLE z ~- W 2 w TILE LAYER af— BERYLLIUM | REFLECTOR-— FUEL~TO-NaK _HEAT EXCHANGER . SPHERICAL B,C TILE LAYER FILLER PLATES THERMOCQUPLE SLEEVE FUEL DRAIN {HOT) FIGURE II VERTICALZSECTIBRATHRC PSS ORNL—-LR—DWG 18291 PUMP SHAFT UPPER SLINGER SEAL SPARGING CHAMBER EXPANSION TANK LOWER SLINGER SEAL _ANTI-SWIRL VANES CENTRIFUGE SLINGER SEAL SEAL PLATE 0.050 in. _y_ SHRGUD 0.05Q in. CENTRIFUGE CENTRIFUGE BAFFLE SHROUD SKIRT CENTRIFUGE 0.198 in. BLADE FUEL PUMP VOLUTE. HUB 0.062 in. FUEL PUMP IMPELLER BLADE SHROUD FUEL PUMP SLINGER SEAL BYPASS FEED HOLE FIG. 1.2 ART FUEL PUMP WITH MODEL 32 IMPELLER -11=- ORNL-LR-DWG. 23922 He+ Xe'>° *He MINUS ' Heyl 2 psi ’ (l? B :LUPPER SLINGER SEAL i Ap~25psi } — ] | A # _— = tf&a_‘v-mss PUMP Ap~Tpsi - | ' "‘G)LOWER SLINGER SEAL ‘ = i~ £ ary L ESa g8 CENTRIFUGE | ~ =4 . CENTRIFUGE ' SLINGER S Ap~47 PSI (62.5 psi) SEAL FUEL PUMP = AP~6i psi AT EXCHANGE NOTE: NUMBERS DENOTING PRESSURES ARE ONLY ILLUSTRATIVE FIG. 1.3- SCHEMATIC ART FUEL CIRCUIT -12- iy 4 Circuit Descrlfi%lon - ‘The ART fuel circuit is most readily understood by referring to Figs. . 1.2 and 1.3. For clarity the circuit will be divided into zones and described separately. | | | | Zone 1. Primary Fuel Circuit _ ' | Circult wise ‘the main fuel loop consists: of two pumps operatlng in paral- lel with a common discharge. Half of this C1rcuit is shown in Fig.:1. b, fuel disch&rges‘intofthe core header and flowsrthrough the core and the heat exchangerslto the pump suction. o | ' The core header.serves‘as a plenum and as{a,mixing chafiber-into which processed fuel is discharged from the céntrifuge (shown dotted in Fig. 1.k4). Since the centrifuge and the fuel pump have a common discharge, it is seen that the fuel pumprdischarée pressure is regulated.by the'centrifuge discharge pressure and that the fuel pump suction pressure is regulated by - system resis- tance of the circult | Zone 2. Bypass Fuel Clrcult The bypass pump passes approxlmately 2% of the main fuel pump flow rate into the sparging chamber where it is intimately mixed with helium to dllute and remove from the fuel the Xenon—>° produced in the reactor core (see Fig. 1.5). The llquld-gas mlxture is jetted into the expan51on tank . where part of the gas is separated from the fuel by allowing the mixture to form a relatively quiet.pool from which the gas escapes to the off-gas system. The fuel in. the expangion tank, with small amounts of entrained gas, feeds by gravity into the centrifuge where the last of the gas is removed.and fed back to the sparging chamber. The centrifuge diseharges into the discharge of “the fuel pump there- by returning the fuel to the main circuit as well as mainfaining the discharge pressure of the main pump. -13- . ABY N R ORNL-LR-DWG. 23923 | | / | I - l | - - - 7 L | o BY-PASS PUMP [ ] | _ . T — — 1 P - , ‘ \i/ | L__J |—‘—::'_ — SPARGER - { EXPANSION . " CENTRIFUGE TANK | (- —a—— ' | N A | | - lf | | | | CORE HEADER x | FUEL PUMP HEAT EXCHANGER FIG.1.4- ART PRIMARY FUEL CIRCUIT “14- BY-PASS PUMP. ORNL.-L R-DWG. 23924 s '_"-.” T CENTRIFUGE - N . | FUEL PUMP |~ HEAT EXGHANGER' —— em—— e e— epm— | _Expansion T TANK FIG. 1.5- ART BWF&SS FUEL CIRCUIT -15- LA A . g L . P Ry Y G Zone 3. Dynamic Seals There are three dynamic seals which are essentially centrlfugal pumps, in the.bypass.fuel,circult,erranged to pump against the pressure of fuel efitering the seel impellers fron fihat.is.normallytthe.discharge end of.the blades. The interrelationship of . thesevseals is shown in Figs. 1.2 and 1. 3. The. upper sllnger seal serves to prevent the fuel from leaking up the punp shaft towards the bearlngs and to allow the helium to flow down the pump shaft and into the sparging chamber., The lcwer\slinger seal serveq. ” to prevent the fuel from leaking-from the sparging"chamfier'into thelcentri-v fuge and to allow. the gas removed by the centrifuge to flow into the sparging chamber. The .centrifuge slinger seal limlts the amount of recycle around the top of the. centrifuge. -16-’ Section II DESIGN AND DEVELOPMENT OF - ART FUEL PUMP AND XENON REMOVAL SYSTEM i ’:j?:-:(_-l v s (F240ts . o o $ YL ‘.Jy"-i { i R DESIGN AND DEVELOFMENT OF ART FUEL PUMP AND XENON REMOVAL SYSTEM A > S, Xenon Removal System DeSign The X-R system described in this section is the finalized.ART design. The system originally:enviSioned for application to the ART was one that would be insenSitive to attitude. Before the development work on an atti- tude-stable. system was completed,“it was. found that the. "north head" region of the reactor would have to be revised because of a severe stress problem. Another factor which contributed to the decision to radically modify the original system was a 50% increase in the anticipated reactor fuel volume. - The redesign of the 'north head" region and the increased fuel vélume | necessitated a change in both the volume and shape of the fuel expanSion tank. The shape of the tank was changed from cylindrical to ellipsiodial with the pump barrels. passing through each end of the oval. The original requirement for this system. to be insensitive to attitude was to be accomplished by spin- ‘ning the fuel volume in the expansion tank and us1ng the resulting centrifugal force to stabilise'the'liquid-gas interface. 'Changing:the shape of the expan- sion tank to.its final design made-.the problem of stabilizing t’he‘fuei‘gas | interface much more complex and resulted in a decision'to eliminate the atti- tude stability requirement for the sake of an expeditious deSign. The develop- ment work completed on the original design is included in Section III of this report to give a background for the final configuration and for. poss1ble future application in an aircraft reactor with an attitude- stability requirement. A Simple plastic model of the X-R system'was deSigned and built to test its gas removal ability and.to calibrate the.bypass_flow through the centri- . fuges. A cross'section,through this model is shown in Fig. 2.1. Sibsimss ey, > ORNL-LR.DVWG. 23925 FUEL PUMP SHAFT EXPANSION TANK REGION BY-PASS PUMP CENTRIFUGE FEED SHAFT SHROUD : BY-PASS FUEL HOLE CENTRIFUGE . FUEL PUMP DISCHARGE REGION ' ' FUEL PUMP SUCTION REGION FIG. 2.1 FULL SCALE PLASTIC M%fi!{fflfi THE XENON REMOVAL SYSTEM -20- Xenon Removal System Flow As -the - des1gn of the reactor progressed the estlmated fuel volume ine creased. ThlS 1ncreased fuel volume necess1tated a larger expan31on tank and a. larger bypass flow rate through the expanS1on tank and centrlfuge for the' follOW1ng reasons. "'_' ' ,",“' T e ' 1. The fuel expan51on tank volume 1ncreased in capac1ty in ""prOportlon to the total fuel system Volume T ‘é. To malntaln -the xenon concentratlon the same as for the smaller fuel volume, the strlpplng rate 1ncreased 1n 'dlrect proportlon to the fuel volume. ' _ 3. 'To keep the greater expan51on volume from overheatlng | and . creatlng a serlous 21rcon1um snow problem, the ;quantlty of low temperature fuel rec1rculated through the expansion tank 1ncreased 1n proportlon to the fuel volume. ‘ ' ' ‘The meterlng of the bypass flow taken from the maln fuel c1rcu1t had to be done by properly sizing the hole up the center of the pump shaft and the four radial holes formlng the bypass pump. A second factor affectlng the. by- pass flow-was the dlfference between the pump suctlon pressure and the expan- sion tank gas pressure. ThlS second factor could not be determlned 1n the model as it is controlled by the comblned performance of the centrlfuge, fuel pump and pump volute. “ .During the model tests the difference between the .pump suctlon pressure and the expansion tank gas pressure was calculated from performance data ob- tained from separate tests of the fuel pump and centrlfuges. From Fig. 2.1 it can be seen that there are. two pr1nC1ple volumes to the model. The volume between the two upper plates 31mulates the expans1on tank and the volume between the two lower plates represents the high pressure region at the dlscharge of . the fuel pump in the maln fuel c1rcu1t At the lower end of the shaft is a bearlng and seal arrangement to separate the | large hlgh pressure volume from the hole in the center. of the shaft. The reason for this arrangement vas to control the Dbressure in the small volume ' at the lower end of the shaft at a value near that expected in the suctlon region of the main fuel pump . To calibrate the bypass flow fhrough the model a line containing a flow meter was confiected.befween the high pressure discharge region and the low. pressure region at the lower end of the shaft. A valve in the line was used to control the flow until the pressure at the entrance to the shaft matched that exbected at the suction of the main fuel pump. If the flow up the shaft was greater than that desired in the X-R system, then the bypass pump holes were reduced, and conversely if the flow was too small the holes were opened. The reasofis for metering the flow with the bypass pump holes were that . changes . in ths holes size were simple to make and also were highly effeqtive as the pressure drop in this region is very large. } The bypass flow rate which fiaS'measured by the flow meter was that bleed from the discharge region to the'region below the seal while the measurement desired was that of the flow up the shaft. In the actual test the leakage around the seal was large (approximately 2 to 4 gpm), and amount to 15 to 30% of the total bypass rate. As this leakage could only be estimated, the first . calibration could,easily be in error by.1l0 to 15%. This bypass rate was later determined.muchfmore accurately in the single pump Inconel hot loop test rig to be described later. Xenon Removal System Gas Removal | The 1ncreased bypass flow necessitated by the 1ncreased reactor fuel volume 1ncreased the velocity of the fuel passing out of the centrifuge to such an extent that gas bubbles were carried into the main system. This in- gassing was eliminated by increasing the flGW-area.available to the fuel. This increase in flow area was aécomplished by increasing the number of holes in the outer wall of centrlfuge from 8 to 33. | A skirt connected to the lower end of the shaft shroud was added to the system after it was originally bullt to prevent the incoming liquid from falling to the bottom of the centrifuge cup and being thrown directly out- ward toward‘ths exit holes\‘ Without the skirt the liquid which is thrown out- ward carries entrained gas bubbles into the high velocity region of the exlt holes where the smaller bubbles Wlll then be carrled on into the maln fuel circuit. Two otheruitems which wereaadded,to the modelgwere centrifuge'baffle and blades. . As the direction and magnitude of the flow in the centrifuge cups are not known, the point at which ‘the smallest gas bubbles are screened out in the cup. is not known. However,-it is logical that any- direct passage through the cup would enable the liquid to. carry more gas through the cup | than if the veloc1ty'were more evenly distributed over the full height of the cup. . The baffle wag added to prevent the incoming fluid from having a direct passage to the exit holes. - The centrifuge blades were added to decrease the slippage between the incoming fluid to the_centrifuge cup and:the cup. Prevention of this slippage improves the system in two ways. The effectiveness,of'the~centrifuge_depends primarily upon the speed of. rotation of the fluid. It is obvious then that any slippage between the fluid and the cup decreases the effectiveness of the centrifuge. . I - _ L .The_second»reason.for reducing this:slippage stems from_thelnecessity' of having to balance the discharge pressure of the centrifuge against the . pressure rise across the centrifuge slinger'seal " _The discharge pressure of the centrifuge must be maintalned above that of the centrifuge slinger dis- charge to prevent the fluid from bypassing the centrifuge and carrying gas 1nto the maln system., With the blades installed it is p0351ble to predict the centrifuge discharge pressure very closely from the equation H = _H;E = 5% where : H = Head, ft 1lb per 1b U = Tangential velocity, £t per sec A g = gravitational constant 32,2 ft per sec2.1’ The presence of slippage in the cup'introduces an unknown coefficient to this equation and also makes the head delivered by the centrifuge more depend- ent on the flow through the centrifuge. . " There is one advantage to having the sllppage in the cup. During'test' with the model it became apparent that the’ discharge pressure of the centri- fuge 1s a function of the liquid level in the expansion tank. 'As the liquid level 1n the expans1on tank decreases the centrifuge pressure decreases, re- sulting in a. like decrease in. pressure at the pump suction region° This reduced pressure at the pump suction regults in a smaller flow up the pump shaft and through the centrifuge. 'The decreased flow’ there causes a reduction in the slippage’ which tends to increase the centrifuge discharge pressure. This stabilizing ‘effect on the pressures ‘was not advantageous enough to overcome the disadvantages stated above, ' The opening between the sparging chamber and the expansion tank was made as large as possible° During tests made with this opening restrictive 1t was found that the fluid was backed into the sparging chamber. When the fluid level moved radially inmward from the tip of the lower slinger seal, it re- stricted the gas flOW1ng upward around the shaft from the centrifuge to the - expanS1on tank. With this normal path restricted, the gas was forced to leave the centrifuge through the same opening 'ags that used by the incoming fluid]which-reduced”the effectiveness of the centrifuge for larger'bjpass flovs. ART Fuel Pump Development ‘The operability of a. face-type gas-seal sump pump pumping liquid metals was experimentally demonstrated in the summer of 1951(5)In the fall of 1953, G. F. Wislicenus designed an impeller (Fig. 2. 2) to produce the head and flow reguired of the ART reactor using Flinek as the fuel. A volute was designed within the space limitation of the reactor for this impeller in late 195k, ‘ Throughout this paper sections of the pump will be located by numerical sub- scripts° The'position of these.sections and their numbers are shown in Fig. 2.2. The subscript (o) designates the "eye" of the passage just before the impeller; (1) the impeller vane outermost inlet edge, and (2) the impeller vane outlet edge or impeller rim.',A full scale model was built and water performance test data were obtained in early_l955.' The model was installed in a test”loop built ofu6 in. pipe with head and flow'measuring instrumentaq tion and a throttling valve, The pump was driven by 1l5=hp variableuspeed d=-c motor.. The first experiment was: performed by using a pumpasuction con- . figuration that 51mulated the’ ART reactor design° The entrance region of the - P | 'g 2h- , :QRNL—m- . 23924 SHROUD m 1. ' ‘ BLADE _ SLINGER SEAL . l’/////// FIG. 2.2- BRASS ART oA P N G T RN AN R QLR RIS RN FUET™PUMP IMPELLER - . %é:fi$ pump was obscured by a flat plate located 1 in. below and parallel -to the pump suction. A second test was performed with the plate removed and replaced by an 8 in. pipe comnected directly to the pump suction. The_results of these two experiments are shown in Figs. 2.3 and 2.4, The pump performance as indi- cated by these two experiments was not affected by the simulated reactor en- trance conditions. Since the flat plate was very. difficult to install, the 8 in. pipe suction was used for all subsequent testing. The test conditions and results.from a series of experiments performed in the test loop are given‘in Table 2.1. The'performance data from these experi- ments are plotted in Figs. 2.3 through 2. 13h The performance data given in Fig. 2.13 are representative of the best operation obtained during this series of tests. The pump efficiencies are not considered accurate on an absolute basis,'because the motor was not calibrated and the motor efficiencies were obtained from the manufacturer's computed data. It isvestimated.that the efficiency of-the pumps, exclusive of seal and beariné losses, is approximately 75% at the design point. The impeller blade tip angle, 62, was increased to 26.5 deg from 22 deg in order to produce the reactor design head and flow and a pump speed,approaching ‘the design speed; the results of this change are given in Fig. 2.5. Data ob- tained by varying the impeller radisl clearance from 0.010 in, to_0.0hO“in. indicated a'loss in pump performance of'approximately 8% at. low-flow hithhead conditions and of approximately M% near the design point. These data are given in Fig. 2 T ' A caVitation-like noise persisted throughout Experiments 1 through 12 at - flows above 400 gpm and speeds in excess of 2000 rpm. The intensity of this disturbance increased with increased flow or speed above the threshold values. It was suspected that the noise'mightcbé;thééresultsfoffipoorlfluidflguidancefiat the leading edges of the impeller uanes, An attempt was made to improve this condition by changing the entrance angle of the blade to match that of the fluid by machining the leading edges so that they lay on the surface of a cone wnose apexlwent through the shaft centerline. The performance data from these -26- changes are given in Figs. 2.6 and 2.8. Modificatipns were;elso.made.to,the, inlet radius of the suction 'eye and.to fihe.impeller nut to improve the entrance conditions. These changes had no noticeable.effect upon the eheracter of the . ndise;;-A new:impeller\blede,was designed to-give better entrance conditions with no’ noticeable improvement. The'performance data,of,this‘impeller blade. is given in Fig. 2,10, | , It was determined that the noise was not prlmarlly due to cav1tat10n at the blade leading edge but rather to a local condition that existed at the tongue, of the volute. This was established by the use of-a'qarbon microphone probe.. A circuit was devised so that the microphone signal could be observed on a. cathode«ray oscilloscope._‘With this arrangement, the region of maximum noise was located near the volute tongue. The volute tongue was cut back to allow more area for the. flow. This modification gave a decrease. in performance with no improvement in the noise. 'The‘performance data for this modification is given in Fig. 2.9. | | ‘ The need for addltlonal flow area in the volute is evidenced by the hy- draulic unbalance of the-impellerfat the design point and .the occurrence of the design speed maximum efficiency at approximafely 20% less than the design - flow. A structural change in the design of the reactor north head region to improve the stress condltions allowed the pump volute flow area to be increased; however, this 1ncrease only compensated'for»the 1ncrease'1n-the fuel flow rate. requlred for improved heat exchanger performance. To - obtain the additional ‘required flow area, the pump volute was redesigned W1thout diffusion cones. Results of the water performance tests as glven in Fig. 2.11 with the re- designed volute indicated excessive leakage losses between the impeller dis- charge and the impeller suction. This is indicated by the sag in the perform- ance curve at high-head low-flow conditions.~'The-éhroud-radial seal slingers were extended to the impeller full diameter to increage their developed back pressure. The resfilts of this modification are given in Fig. 2.12.' To further reduce this leakage, the impeller'axiai clearance was reduced from 0.102 in. to 0.053 in. The results of thisvfiodification are given in Fig. 2.13. -27- R R The différence in the pump total discharge head and.the average impeller blade static discharge head as’a function of pump speed-and flow, as deter- ‘mined from Experiment No. 15, is shown in Fig. 2.1k. The data shown in - Fig. 2.14 allows the -calculation of -the:volute efficiency (k). This effici- ency, k, (Fig. 2.15) is defined as the fraction of the absolute impeller dis- charge velocity head which.is converted to static head at the pump discharge.. It may be'séen from‘Fig,'2;13 that the design point lies in the region of maximuh efficiency and‘that:the condition of.hYdraulid force balance on the impeller occurs near the design point:'.Thg-nOise»present with the previous" volute design was completely eliminated. Based on the fierformance of the pump indicated by Experiment:No;.l5, it was decided that‘thé redesigned volute and the ofiginal impeller with a B, of 26.5 deg would be used in the reactor. _68... TABLE 2.1 CONDITIONS AND RESULTS OF WATER PERFORMANCE TESTS OF ART FUEL PUMP Experlment |Reactor Design Impeller Impeller:: [Point at Time Suction . S Number ' Design¥* Design Point of Tegt** Conditions Remarks and Results 1 Five vanes, Y70 gpm 606 gpm | Suction box, simu- Impeller design point met at Blade Tip Angle, b5 £t head 50 ft head|-lating reactor; 1/2 in.| approx. 2650 rpm.- Reactor (B ), 22 deg 2500 rpm . ~2300 rpm radius on suction eye design met at approx. 3050 2 ' e rpm. Pump very noisy. 2 Same as 1 Same as 1 Same as 1 | Straight 8-in. pipe No appreciable change in - S L ' with four antiswirl performance as compared to vanes; 1/2 in. radius Exp. 1. on suction eye . 3 Five vanes, 606 gpm Same as"1l | Same as 2 Reactor design p01nt ‘met at | : 6235 26;5 deg .50 £t head ' approx. 2800 rpm; approx. a S - 2850 rpm . 10% increase in efficiency. o with respect to Exps- 1 with | the ‘peak efficiency shifted | towards the hlgher flows Pump very noisy ' b Five vanés, Same as 3 Same as 1 | Same as 2 -Reactor design met at B, - 22 deg : | | approx. 3100 rpm. Pump leading edges wvery noisy. - jcut back on an ' R 80 degcone , angle L 1. _ L L , 5 lseme as 3 Same as 3 | 'Same as 1 | Same as 2 Impeller radial clearance was increased to 0.025 in. from 0.010 in. No appreci- able change in pump. perform~ { ance., Pump very.noisy - ‘Table 2.1.(cohtinued) Reactor Design Suction . -0t~ Experiment Impeller - Impeller Point at Time P Number Design¥* Design Point of Tesgt** Conditions Remarks and Results 6 Same as 3 Seme as 3 Seme as 1 | Same as 2 Impeller radial clearance - . . T was increased to 0.0L0 in. ~from 0.025 in. No appreci- able change in pump perform- ance at design point; 8% loss in head at flows below 300 gpm. Pump_very noisy.- 7 |8ix Vanes, Same as 3 - 620 gpm Same as 2 | Impeller design point (Exp. B, - 22 deg ’ 35 £t head | ' 1) met at approx. 2600 rpm. - |1€ading edges cut A~ 2800 rpm Reactor design point (Exp. 1) ~ |back on a 56 deg ' ‘met at approx. 3050 rpm. Re- cone angle ! actor design point (Exp. ') o met approx. 2850 rpm. Pump very noisy. 8 Same as T Same as 3 Same as T Same as 2 except Simiiar:to Expts. 1 and 7. ‘ ' ' ' ' ‘ 1/2 in. radius - S ' increased to 1 in. on suction eye. 9 Same as 3 Same as 3. Same as 7 |Same as 8 Volute tongue cut back " ' ’ 1 ,3/8 in. Performance simi- lar to Exp. 3. 10 - |Same as 3 Same as 3| - Same as 7 |Same as 8 Volute tongue cut back 3/8 in. modified impeller nut; performance similar. to Exp. 3. B h =’"i ° "TE" Table 2.1 (continued) Number Experiment| ~ Tipeller Design* = -Impeller"" Reactor Design Point at Tlme of - TestX¥k:- Suctlon “‘“Condltlonsr 11 12 . 13 14 157 Same as 3 Five vanes;¥* e2 - 26.5 deg Same as 3 Same as 3 . |Same as 3 Design Point Same as 3 620 gom 25 ft head 2800 rpm Same as 3 . Same as 3 Same as 3 Same as T ‘Same as 7 645 gpm 37 £t head A 2800 rpm Same as 13 Same as 13 - . ‘Same as 8 - Same as 8 . Same as 8 - Same as 8 Same as 8 flow. and speed .ameter, Remarks and Results Volute tongfie cut back 30 deg; performance showed a ' 'deCrease'in_head and flow. No change in noise. Volute tongue back 30 deg; j:Performance similar to Exp. 11 New volute with increased flow area. Impeller hy- draulically balanced near reactor design point. Head is decreased at constant . Pump very quiet. Extended sllnger on lower shroud to full 1mpelle* di- Increased hea® at constant flow and speed. '-.Eff1c1ency peak at de51gn - p01nt Impeller axial shroud clear- ance reduced to 0.053 in. from 0.103 in. Increased head at constant flow and speed. | Increased efficiency peaks at design point. o :‘E{F{: . b Sl ' ¥ The impeller blades were redesigned to provide entrance angles more nearly aligned with the fuel. *% Due to the time lag between the impeller design and the 1mpeller testlng, a different impelleii ée51gn point and reactor design point resulted. - ze.. 100 90 Head (.i:'ra__lh) ' omw‘ao_mr WAL E Dol wew Pump Efficiency (%) .- e Pressure Balance Line Reactor Deslgn Point A 2700 mpn 300¢ rpm Pump Speed . 600 700 800 CAPACITY (gpm) FIG. 2.3 ART FUEL PUMP PERFORMANCE CHARACTERISTICS EXPERIMENT NO, 1 - TABLE 2.1 ORNL-LR-DWG, 23924 £ _ : olutel Pressure Ballance Line / \ <--- Pump Efficiency| (%) y ) 75 /’ - 70 " Reactor 7 Design Point / \ P 4 /0 \ L/ ' 3300 rpm Pump Speed 1 [ Ciq." . . L. 60 i ”~ A NN\ 3000 rpm 2700 rpm © 2250 rpm ' 2000 rpm 00 rpm 100 - 200 300 L0oo 500 600 700 800 900 1000 ’ ’ CAPACITY (gpm) ‘ ' FIG, 2., ART FUEL PUMP PERFORMANCE CHARACTERISTICS EXPERIMENT NO.-A:;?‘x"Il“ifiEllfiliE 2.1 ORNL-LR-DWG. 23929 _ I . Pressure Balance Line ---- Pump Efficiency (%) 3000 rpm Pump Speed \ 2700 rpm pm 1500 rpm Tm L1000 Tpm 400 500 500 700 800 900 1000 CAPACITY (gpm) ' ' *T.f'n':..‘s’fi FIG, 2,5 ART FUEL PUMP PERFORMANCE O ARACTERISTICS EXPERIMENT NO. 3 -~ TABLE 2,1 SRR -ge- ORMNL-LR-DWG. 239!' 90. 80 Head (£2.1D) === Pump Efficiency (%) Plressure Balahce Line _—Reactor Design Point 70 60 504b ’- 3000 rpm Pump Speed 700 rpm g 500 rpn 30 LOO 500 600 T00 800 200 1000 CAPACITY (gpm) 0 100 200 FIG, 2,6 ART FUEL PUMP PERFORMANCE CHARACTERISIICS EXPERIMENT NO, 4 - TABjE:Zulosswes 90 ft 1b AR ORNL-LR-DWG. 23931 80 kl\‘--\ ] . 70 = ::_: - N Experiment No. 3 . < : Hxperiment Ng. 5 <+~ - \‘\ 60 \\ I NN Experiment No.b\ 50 5 \. F,Reactor Degign Point F \\ ® ko - \ O = 30 > z 20 \ 10 T00 200 300 LG 500 %00 700 800 960 1000 CAPACITY (gpm . ) S o | S FIG, 2,7 ART FUEL PUMP PERFORMANCE CHARACTERISTICS AT 2700 rpm EXPERIMENT NOS. 3, 5, AND 6 - TABLE 2,1 o -Le- ORNL-L R-DWG. 23952‘ 90 80 === Pump E fficiency (% 70 € Pressure Balance Line Head (EI%E) tor Design Ppint 3000 rpm Pump Spepd pm ® 1000 pm 300 Loo 500 600 700 800 900 1000 CAPACITY (gpm) FIG. 2.8 ART FUEL PUMP PERFORMANCE CHARAGCTERISTICS EXPERIMENT NO, 7 - TABLE 2.1 RRNRRERY . QRNL-LR-D¥G, 23933 8 ---- Pump Efficiency (%) 7 Velute Pressure Balance Line -88- .Head (EEI%E) actor Design / N\ Point 10 \ 2700 DO rpm rpm Pump Spe ad 0 100 200 300 400 500 600 700 800 900 1000 | CAPACITY (gpm) _ FIG. 2.9 ART FUEL PUMP PERFORMANCE CHARACTERISTICS EXPERIMENT NO. 11 - TABLE 2,1 -uan’E% 90 80 ---- Pump Efficiency (%) 70 Head (%}E) 10 - Volute| Pressure Balance Line /7 ‘?\Kflemtor Design Point 200 7 ~ - )/ETJU rpmPump $peed Ve A s _ ~ e 7 < s 24p0 7 rpm N ,/ o~ ” ’/' 70>\ ~ ~® 2000 rpm 6§ 300 L0O 500 300 700 800 CAPACITY (gpm) FIG, 2.10 ART FUEL PUMP PERFORMANCE CHARACTERISTICS EXPERTMENT NO; e~ TABLE 2,1 -ov- Head (%F]'E) 70 50 /) ORNL-LR-DWG. 23935 ~~== Pump Ej Priciency (% 40 r'\ rpm Pump Speed 30 20 P~ \ | . / 2D00 rpm ) ‘\7\. 150 o .\ / ~ ~ / s - ! | \ \ .S 7 100 rpm 1] e >+ ~e .~ { Volute ! [ \ N\ \ g \J\\\t /7 ® 1500 FIG. 2.12° ART FUEL PUMP PERFORMANCE Ci ARACTERISTICS EXPERIMENT NO. 1l - TABLE 2.1 " CAPACITY (gpm) 00 rpm =% 55 rpm 100 200 300 400 500 700 1000 -z~ Head (£, - 1% 8o 70 50 30 20 10 IRy CAPACITY (gpm) \\\\ 5 ~~w= Pump Efficiency (%) | 7 % ‘\ II I’ \/i _ o / \ / .’ ! ~..® TN / / 7'\ . ' \,\'\ / . / / \ / 1/ . ~ . [ 7’\.\ / / L. 80 < I’ / T~ /// 81\\\.- \ / / / ' TN \L\l l/ / / } ll 75 - es 70 T~ o/ I | Po%/ 7 N * | , / 3000 rpm Pump $peed | [/ / || / \L\! // / > N // / i O 7 D2 NI { i /| N | , \ \ \ \ / // P /™ 2700 rpm “\ \‘ SN \ / N\ - 7 \K \\-.-:' \/./ a4 2400| Tpm ‘ — — _ \ ™~ —— {__.—- 2000 rpm ~ oluteé Pressure Balance‘ L;Ee > — U I'pm 100 20 50 5 70 8 90 10 . FIG. 2,13 ART FUEL PUMP PERFORMANCE CHARACTERISTICS EXPERIMENT NO, 15 - TABLE 2,1 pumussii™® It Pump Total Discharge Pressure Minus Impeller Blade Static Discharge Pressure (—-.I#) - o Y ORNL-LR-DWG. 23938 100 m8 uoo _.r8 moo 80 ...8 moo woo : CAPACITY (gpm) FIG. 2,14 .oowwmo,fiozm APPLIED TO THE ART FUEL PUMP TOTAL DISCHARGE PRESSURE FOR OBTAINING THE CENTRIFUGE STATIC DISCHARGE PRESSURE -43- FIG., 2.15 Els g 8 qm..% / & = . 8 E - E fl‘ / o | | g & S \ o o, m \\ o um. ° TN B g g" \ . o H B / M (ud3) Lousto}srd : N0 wrgrixey 38 moTj dumg \ \: o \/// / M \ "~ AN N & L / S \. / / A,%. N N & \/ ‘ / g . B \ %// NS .// o / . m /\/ N m / _ / B \;///// _ N .mm S g ,/;/////////, < / m. / o E ,//, £ S B " 3 2 '3 R ‘S A & Q (g ur £ouetoryyg eynfop dumd Temg) ¥ -l ART FUEL PUMP VOLUTE PERFORMANCE CHARACTERISTICS Combined.ART Fuel Pump and Xenon Removal System Development B The £inal development work on the X-R system was’ done with a gingle pump Inconel hot. 100p test rig (see Fig. 2.16). . This’ final work was originally planned to ‘be -done- W1th a twin pump test loop very similar to the ‘north head region of the ART. Construction difficulties with the twin pump loop delayed its completion and the develoPment work was shifted to the single: pump loop.“ This looP combined the’ Latest X-R system tested with the plastic model and the final pump design determined from .the pump. water tests.“ All development work done With the Single pump rig used water as the working fluid. System Pressure Fluctuations At high liquid levels in the expansion tank (above 3 in.) it was found that the pressure throughout the main: system was subject to rather Large and irregular fluctuations. These fluctuations decreased as the expanSion tank - level was decreased from 3 in. to about 3/h in. where the fluctuations again’ began to increase.‘ This increase at low levels was expected.' The fluctuations at low. expanSion tank levels were observed in the plastic model and attributed to a fluctuating liquid-gas interface in the centrifuge.. The flow into the centrifuge is by graVity from the expansion tank and any decrease in the level in the expanSion tank will decrease The flow to the centrifuge._ As the main ' fuel - circuit is filled with an essentially non-compressable fluid it is necces- | sary’ that the flow to and from this system be equal, thus, . any. change in the flow to the centrifuge will result in a change in the pressure of the main system and therefore a change in the flow from the main. system back to the expanSion tank° It can then be.seen_that as the liguid level in the expansion tank decreases, the flow‘to and'through'the'Centrifuge-decreases‘and the main system pressure‘decreases.- As shown. previously the main system pressure is controlled by the discharge pressure of the centrifuge and any decrease in main system pressure results from a decrease in the centrifuge discharge pressure.: In the case_beingidiscussed the centrifuge-is running at constant speed and has - a Tixed outside diameter. Therefore any decrease in discharge pressure must Lo : . - s B UNCLASSIFIED PHOTG 27742 7 o l; = B! - 3 . . " - o i g i T il = 3 aL o : il r = A - — - e 2 - " - ia b = - i . - - - §"F ~ i R . . V } ¥ g b Ly . - - 2 3 - 2 =t - L o -3 e - = i L) - I 1 5 . = B [ i Sy S - A : 5 : o : i g e - T " . FIG. 2.16- COMBINED XENON REMOVAL AND FUEL PUMP HOT TEST result from a decrease in the suction pressure of the centrifuge or increase in the effective inSide diameter. .In the case of the low expanSion tank level the effective inSide diameter of the liquid—gas interface moves radially out- ward decreaSing the pressure rise through the centrifuge._ This change in liquid- gas interface was observed in some of the earlier tests with plastic models._ For very low liquid levels this condition becomes. unstable with the centrifuge alter- nately trying to becomemfirst empty and then as-the,outflow decreases to become' full.‘ ;.. : - : - o . A S :i . . The cause of ‘the fluctuations at high expanSion tank liquid level was not nearly 50 obvious. | During the development work each of the various flow paths were investigated and the difficulty Mas found to be caused by a fluctuation_ in the pressure at the inSide diameter of the centrifuge The cause of this fluctuation was traced to the lower slinger seal . At very high liquid levels"‘ the centrifuge will run in a flooded condition and it is: probable that slugs.' of liquid as well as gas Wlll flow upward between ‘the shroud and shaft to the inlet of the lower slinger seal ‘This gas and liquid flow through the slinger impeller would cause it to .surge with the surges being transmitted back to the centrifuge as irregular periods of high and low pressure. These surges would then be transmitted through the centrifuge to the main system. ) ~To alleViate these surges the vanes were removed from the lower slinger seal By remov1ng these ‘vanes the head of the lower slinger seal is reduced and any flow surges in this region Wlll be dampened. System Pressure Level A difficulty encountered in the. early testing of the combined system was an excessive system pressure level which results in 8 high stress condition | being imposed on the plate. which separates the expanSion tank and ‘the fuél pump discharge.- Since this pressure level is fixed by the centrifuge pressure ‘rise, attempts were made to reduce the centrifuge pressure rise by two different approaches._ First, the inside diameter of the centrifuge cup was increased and - second, the exit-area_fromflthe centrifuge was decreased., There were several ad- vantages to the first'method; The change in the'existing centrifuges were very “h7- simple to-make; and also it was possiblento calculate ver& closely the new.in-e side diameter of the centrifuge cup to get the desired change in the main system pressure level. Another advantage is that the Cross sectional area of the en-. trance to- the cup is 1ncreased and will 1ead to a more stable system pressure at lower expan51on tank liquid levels.‘ One disadvantage is that the radial depth of the 1iqu1d in, the centrifuge 1s decreased. This thinner 1ayer may enable gas bubbles to pass into the main system under operating conditions where they would have been stopped with the thicker layer. In the case of the small exit=area from the'centrifuge the major disadvantage was the difficulty of -making the change. _Any decrease in the exit‘area byudecreasing the nunmber of exit holes or the diameter of these holes'will increase the-velocity through the cup and thus reduce the degassing ability of the centrifuges.' To overcome this difficulty the holes would have to be tapered with the hole area at the inner wall of.the cup remaining the same and with the ares at the exit in the outer diameter being reduced. Not only'would the construction of these holes be difficult but their size would have to be determined by trial and error. One advantage of this system is that the liquid level or liquid gas_interfacef . in the centrifuge‘does’not change. .The .second and’ more important advantage was never confirmed by test. This second advantage would show up only in the case where one of-the two reactor pumps_had stopped while the other was running. ‘For this-one.pump out operation it is very desirable that the centrifuge have a head flow characteristic such that the head decreases very rapidly with any in- crease in through flow. All of'theuclearances around the centrifuges are large and the leakage is controlled by a d&namic seal. When one pump stops the seal- ig also lost so.that the. leakage becomes very large unless the pressure drop across this clearance is decreased.. In. the case where the exit area is small,r the‘pressurezdrop,across the_exit holes of thepoperatinglpump isclarge and will increase rapidly with the increased flow resulting from 1eakage around the in< . operative pump. In the case of the enlarged inside centrifuge diameter the ‘ head of the centrifuge must also drop off rapidly as’ the through flow 1ncrease, but in this case, the decrease in head must be accomplished by the centrifuge, the centrifuge becoming partially empty and'increasing the effective inside diameter even more. This increase in the diameter of the gas-liquid inter- face leads to a breakdown in the effectiveness of the centrifuge. Bypass Flow Determination and Control One of the most important purposes of the combined system tests was to determine and control the-bypass flow. As. previously noted the plastic model served only. to give an approximate rate because of leakage of . the lower seal. In the combined single pump--system the bypass flow passage was 1dentical to that for the ART. A disadvantage of this‘system is the lack of a method of directly measuring this bypass flow. The procedure used was to plug the hole in the center of the pump shaft through which the bypass flow normally passes and substitute an external line between the pump suction region and the expan- sion tank. This external line contained a valve and flow meter to regulate and measure the flow. During these tests runs the pump speed and main system flow were held constant'while the flow through the external line was varied. The expan31on tank gas pressure and the pump suction pressure were measured. This" method yields a -pseudo head-flow curve for the centrifuge and is used as a calibration for test runs where the actual flow up the pump shaft cannot,be determined -directly. The use of the pump suction”pressure rather than the centrifuge discharge pressure for calibration purposes 1is arbritrary. The use of the suction pressure does simplify the immediate reduction of the data, and - more important the pressure differences between the pump suction and the ex- pansion tank is one of the factors controlling}thehbypass flow and is of more interest than the pressure difference.between the centrifuge:discharge and the expansion tank. -After this pseudo head-flow curve was established for a series of operating points, the plug was removed from the pump shaft and the external line closed. The operating point used in the above calibration runs were then repeated and again the pump suction pressure and the expansion tank gas pressure were determined. The data for the runs with the pump shaft plugged were plotted as shown in Fig. 2.17.. The'difference between the pump suction pressure and the ‘expan51on tank gas pressure for the pump shaft open runs were then checked against the identical calibration run and the bypass flow up the shaft was read from the curve, -.)4-9"' ORNL-L R-DWG. 23940 Bypasrs Flow through the Expansion Tank - Bhaft Pluggefl . ' (LO0% = 17.6 gpm) , -og- FUEL PUMP SUCTION PRESSUNLE MINUS EXPANSION TANK GAS PRESSURE 1z 10 ~2 . Expansion [Tank Liquid Level - Shaf L Open / 100% [ Pump Speed) 5 A RN Ty S A, fll ~\ A . O ~N « | 2400 rpm & Y [y \ ) LR b0 rpm 3300 rpr . FIG, 2,17 ART FUEL PUMP BYPASS FLOW CHARACTERISTICS WITH MOOEL 32 IMPELLER AT A 3-1/2 IN, TEPTH OF LIQUID (WATER) IN THE EXPANSION TANK The flow through the external line or up the pump shaft is not the total flow through the centrifuge;"Part of the centrifuge flow recirculates over the centrifuge slinger seal and back 1nto the cup. Thisfrecirculation flow" is not ‘accurately known but s1nce it is not a function of the manner in which - ‘the, flow gets ‘from the pump suction to the eXpansion tank, but instead is only a function of the flow through the" centrifuge, it does not enter into the cali- bration. = Fuel Leakage Past Upper Slinger'Seal. One item 1nherent in the X-R system design was the cause of much concern. This was the poss1bility of fuel passing over the upper -slinger seal and being forced or drawn up into the narrow annulus between the pump shaft and the bar- rier shield. plug If any ‘fuel from the system were to. rise into this region it could freeze because of the cooling of the lubricant circulating in this region. 'The fuel on freezing becomes hard and could do damage to the shaft 5 and shield plug can ‘and could. pos31bly freeze the shaft to the stationary shield plug can. This would necessitate a change of the pump rotary element. If the reactor had been at power this task would be unpleasant. - In the early test 1n the Inconel hot. loop, the upper slinger seal design was the same ag used in. the plastic model. This design had twelve radial vanes about 3/16 in. high for. the full radial distance. (see Fig. 2 18) In these early tests 1t was found that the gas pressure in the expansion tank was. greater than that in the oil catch basin region -of the rotary element. In many cases, this. pressure difference was as large as three to four feet of. working fluid- head. TFor the first de51gn this pressure difference was found to depend upon the pump-speed,,the gas flow down the pump-shaft,annulus, and the 1iqu1d level in the-expansionftankc“vThereaWas*a‘noticeable?"break“pointfi.in‘this.pressure difference as the expansion tank liquid level passed 2 in. This is aiso~the~ “level at which the liquid. will uncover the ports between the sparging chamber and the expans1on tank. As the liquid level was reduced from a. ‘high. level to ‘about 2 in., there was no significant change in the pressure difference. At approximately 2 in. the pressure differerce dropped sharply and was then again fairly. independent of the liquid level as it was reduced further. -The level ‘at which the break point occurred was not fixed but was-dependent on whether .. the liquid level was ‘increasing or,deereasing. The.leéel.for'this break point ‘'was higher for the:case when the liquid level was.increasing-than»when the level was decreasing. It was also noted even at the highest liquid levels: that as the rafie-of.gas flow down from the eatch basin was increased the pres- sure difference was decreased and for extremely high £low, several times the normal'SOO-liters~per day, the pressure difference would reverse with. the. . cateh basin gas pressure -being greater-than the gas'pressure in .the expansibn tank. It was also found that this undesirable pressure difference increased as the pump speed 1ncreased.» - Undéer: normal operating conditlons & gas pressure reversal in this region Wlll not cause any damage. The potential head of the slinger is greater than the pressure difference 80 that no liquid will be drawn over the vanes. The real danger of this condition is for;the‘case'where one or both pumps stdp suddenly‘asxinwthewcase;of a power failure. If the slowing down rate of the pump is so .rapid that the inflowing gas camnnot build up the pressure in the lower pressure -region at the same rate the dynamic head of the seals is deQ creasing due to.the decreasing speed, then the liquid will be drawn up into - the annulus around the pump . shaft and probably require a replacement for the 'rotary assembly. ' ' All of the above .observations show that the upper slinger seal is.acting as a gas pump_ The actual gas pressure buildup in the expan51on,tank is. far greater than a pnmp'of-the size and type used should be able‘to'n;rmallj pro- duce using gas as the working fluid. It appears that the pressure buildup is due.to.thefreCireulatiOn of 1iqnid'through-and'over-the:_vaneso The upper slinger seal acts' the same-as an-open‘face eentrifugal"bump’eperating under no flow conditions. Due to the rather large axial clearance over“theztop of the seal and recirculation is large,\ This recirculation appears as a flow outward through the vanes and an inwardfflow through the clearance above the vanes. - The outward flow carries gas bubbles to the outer tip where it then escapes into the sparging region and then into:the expansion tank._ As this process continues the pressure difference Wlll increase and the length of Vane which is covered by liquid will increase as-the liquid is drawn radially inward to produce a. large head to offset this rising pressure difference. As this radial dimen31on 'of the 11qu1d increages, the centrifugal effect of the spinning 11qu1d begins to screen or centrifuge the gas inward toward ‘The pump shaft. These counter- acting effects thus 1imit the magnitude of the pressure difference the upper ‘;slinger seal can produce. This also explains the observed dependence of the _ pressure difference on the 1iquid level, speed and gas flow. As the gas flow is 1ncreased it will reach such a flow that the upper slinger seal cannot en- train and pump -the gas as fast as 1t enters.. When this flow. is reached the pressure difference will reverse.. In. the case of the effect of the liquid level the observed hysteresis becomes clear. Before the upper slinger seal can act as a gas pump it must become primed. When the liquid 1eve1 is in- 'creasing the upper slinger seal will "throw". the liquid avay from its tip | until the level has reached a. point higher than the upper slinger seal or un- til the 1iqu1d which obstructs the ex1t ports from the sparging chamber causes the pressure to build up -in the region of the tip of the upper slinger seal. When the expansion tank is filled to about 2-1/2 in., the liquid 1eve1 is above the . upper slinger seal and 1t is primed when the pump is. started To break this prime the 1eve1 must be reduced until the exit ports from the sparging.chamber are open and the pressure:at_the outer_radius of the upper slinger seal is reduced. This als0'accounts'for the sudden ghange in pressure when.these points are‘reached."The ability'of the'upper slinger_seal to en-~ train and pump gas is obviously a function of the speed. | In trying to overcome th1s problem, the basic ided was to reduce the large recirculation over the upper slinger seal which with the large axial clearance required is impOSSible to completely stop.‘ The first attempt was »to replace the vaned upper slinger seal with one whose top surface was smooth The axial clearance was held the same as for the vaned upper slinger seal. This reduced the unfavorable pressure difference but did not reverse it. S "u! . -53- Severa150ther methods were tried with some improvement but were not completely Satisfactory. All of these changes were attempts to reduce the outward flow . potential at the rotating gsurface. ' ' ' The de51gnwwh1ch finally gave a pressure difference in-the desired direc-. . ‘tion, i.e., the catch basin’ pressure greater than the pressure in the expansion tank, is shown in Fig. 2 19 The top surface. of the upper slinger seal was cut back for a radial distance of 3/h in. This configuration gave a catch basin pressure higher than that in the expansion tank by about 10 to 14 in. at design 'eondltlons and high llquid levels in' the expansion tank. This incréased clear- ance reduces the.potential.head of the upper slinger seal, but as it has far | more head capacity than required to protect the rotary element when operating normally there is no deleterious effect in making this change. The outSide.. diameter .of the upper slinger seal could not be decreased to obtain the fabor- able pressure'gradient.' It was found that a liquid seal must be maintained be- Atweenfthe face of the upper slinger seal and thé¢ sparging chamber. With this . Pavorable pressure gradient,.the»fuel.cannot enter the shaft annuluSJduring the slowing down ‘of the pump. Also once the pump has stopped, the gas flow : . to the expan51on tank through the shaft annulus malntains the favorable pres- sure gradient. ' ' ‘ Experimental Results : ' The test results of the final design configuration are shown in Figs. 2.20 through‘2,28. Figs. 2.20 through 2.23 give the-head flow characteristics of the sYStem as a-funetion of speed. The value of the difference hetween the pump suction?and expangion tank pressure is also shown. The small variatiens in the head~flow curves frem'ene liquid level to another are probably not sig- nificant and are within the*aceuracy~of the data. The values of the difference ‘between' the pump suction pressure (PS)'anduthe expahSion-tank-gaS'pressure'(PHe) - are definitely a function of liguid level at lower levels, especially at 1 in. and below for reasons given previously. -~ S S . 5o . UNGL ASSIFIED . ORNL-LR-DWG. 23941 3}“ 3“.‘ - “ 76 xl_e V_ANES | "FIG. 2.18 'UPPER SLINGER SEAL WITH RADIAL VANES | ) FIG. 2.19 . o UPPER SLINGER SEAL WITH AN AXIAL STEP 3 , — The reason for the scatter in the data can be seen by examining Figs.. 2.2h.through 2.27. -These figures show the pressure fluctuations throughdut the -test loop as a function;of speed, flow, and liquid level. . While these fluctuations are not large enough to cause any damage to the reactor, they did makejanalysis.of the data difficult. The reason these relatively small fluctuations_wére so important can be seen by studying Fig:.2717. If the _ valueof-v-PS "PHe on the figure are shifted.by 0.2 ft, the error in.deter- mining the actual bypass .flow at 645 gpm amounts to about .5%. At lower liquid levels the fluctuations become larger and the experimental values .of the bypass flow erratic. For this reason, the bypass flows for the lower : liquidqlefels are calculated (see Appendix E). | Calculated .and experimentally determined flow up the pump shaft or that amount which bypasses the main fuel pump is shown in Fig. 2.28. This is not the total reéirculation through the expansion- tank. In the test loop in which the data was obtained there was é leakage .of approximately 0.6 gpm from a point near the pump suction to the expansion tank. This leakage should be added to the bypass flow to get the actual total flow th;ough the expansion fiank{ In the réactor there is also a leakage very similar to the one in the tést loop. This leakage is around the island where it paSses through the floor of the expansion tank. The amount of this leakage is estimated to be about 4 gpm or 2 gpm per pump and like the leakage in the test loop will depend upon the operating pump speed and flow. ‘ ' The total flow through the centrifuge cups is equal to the sum of thé bypass flow, leakage and recirculation over the top éentrifuge'seal vanes. The leakage over the vanes is large, about 6 to 8 gpm at design conditions. This large leakége_does not effect the degassing ability of the cup. The reason for this is that when the flow hasiincreased to the point where in- gassing occurs the heé@ of the centrifuge has dropped to the point where it: is less than the head of the vanes. Under this cpndition it is probable that the centrifuge is not completely full and that the inside diameter of the vanes receive gas only. The vanes then act as a slinger and there is no flow through them in either direction, the location of the liquid-gas interface being determined by the head required to match that of the centrifuge.- -56- - Ls— (LL1b HEAD 70} 50 Lo 20 ORNLM N b S 0L S 10— 160-rpm /“"/ 7 /f / / /.r\//x//////\x.{llBlLl } 1800;\\/‘://///'//\,.\;\///,>( 12 1,4“)0 rpm )./ /__\'\ /\./. > 300" o0 900_1;9‘:/-;'-—{.- -;..,___.; //\//\ X — N A N AN 2100 3 - mPs L 1800 npm ~Z ™\ Y~ \\ / 15001‘pm-.\ | ‘ / | T~ /32 1/16 =~ . 350 0500 &0 ——750 8O0 %0 —mte —T CAPACITY (gpm) ' FIG. 2.2l ART FUEL PUMP FLUCTUATION CHARACTERISTICS WITH MODEL- 32 IMPELLER - AT A 1/2-IN, DEPTH OF LIQUID (WATER) IN TE EXPANSION TANK R - @@W £ 1b 1b -29- ZAD ( H 70 60 50 40 onne . jel Pump Suc Fluctuations . \ . N\ N\ N tion Pressure N/ TN /'\'./ ) ' | /8 1t N/ /58 ek S N 7 ISL NG £ Oy 1/32 \'\/ g i 1209 // 1/16 - \'\_-.‘ \(‘/ "--" 300 600 rpm | ' ' TPMmey,_ | , 0 100 200 300 Loo 500 600 700 800 - 900 1000 1100 - CAPACITY (gpm) . f » ' FI3, 2,25 RT FUEL PUMP FLUCTUATION.CHARACTERISTICS WITH MODEL 32 TMPSLLER ’ ‘ AT A 1-IN, DEPTH OF LIOUID (WATER) IN THE EXPANSION TANK — PR -€9- ORNL-LR- . 23948 70 ; ; T ‘ - - 3300 rpm Pump Speed 50 : . ' - 07671~ S , \i-mel Pump Suction| Pressure Fljictuations 50 ‘ - / Y _ ; ) \ ) . ) 2700 rpm ‘\,//7 Lo +— / N N 3 g S N HEAD ( ~/ I 7N ~— /A 0 §-68 3 2100 rpm :\'w / , \/ . ‘ .\. /7 . : ‘ . — | / ~C 10.28° £t 20 : 1600 rpm_ . ' ,/ ~ 0.2L £t | = 7 - p— < 3/8 : 5/16><' WV = . | : \§V—F\m1 Pump Suctiop Pressure Fluctuations —— ) ft 1v HEAD ( -v9- / SN 210D rp:/ €< AN ‘Y _ / | ‘\ 5/15 2100 rpm 2700 rpn \/ / N N /7 _ 1716 4 1500 rpm / 7 ~ 1800 krpm \//\/N\ / / 1200 /\1/59/ > \ | . = Y N | 1/’ - TAe | [ /16 "\ R ' | S-.-‘\-. 300 %OTHR___ 100 200 300 500 500 600 700 800 " 900 1000 1100 , | CAPACITY (gpm) ) ] _ FIG, 2,27 ART FUEL PUMP FLUCTUATION CHARACTERISTICS WI T MOTEL 32 IMPELLER AT A 3-IN. DEPTH OF LIQUID (WATER) IN THE EXPANSION TANK -g9- ORNL-L R-DW¥G. 23%0 e Cplculated = o =« = =" = = Experimental o . . 55 2ho; " .-1"/26 Expahsion Tank ' X Liguid Lewvel 2100 rpm | 50 33 ™ v \ | 3/‘hw_..\ \\ \ \ o !{ 3300 rpm B R 1 I _‘.\' |- ‘\ /| L A AT Ay 7] 2700 rpo VoL \ | ,./‘? 100 rm e s T "~ P100 rpm FIG. 2.28 CALCULATED AND EXPERIMENTAL ART FUEL PUMP BYPASS F;L.OW CHARACTERISTICS ART Fuel Pumpwgavitation Characteristics ) | Cavitation refers to conditions nhere,.because of a~local increase in velocity as in regions of either boundary curvature or‘eddy.generation, vaporization of the flowing liquid creates vapor filled cavities. These cavities collapse . in regions of higher pressure.elseuhere in the system. In order to form such vapor cavities,‘the pressure first has to drop to the vapor.pressure of‘the working fluid; a condition that can be realized as a local condition without a change in the average system.pressure. For a pump oPerating at a constant-speed.a local pressure drop results from separation and contraction of flow and deviation of streamlines from their normal trajectory such as takes place in a turn or in passing,an obstruction to the flow. The evils of the cavitation phenomenon are essentially three in number: (1) reduction in efficiency due to constriction of flow and loss of energy, (2) objectionable - if not structurally dangerous - vibration and and noise; and (3) possible extensive pitting of houndary materials in the zone of bubble collapse, due apparently to failure through fatigue after countless stress reversals. ' The local pressure drop caused by the difference in the pressures on the leading and trailing sides of theVimpeller blade is.the ma jor cause of cavitation in a centrifugal pump. This local pressurefdrop is a function of - the relative inlet velocity. (W), the nuiber of blades (Z), the thickness of the blades (t) and the local absolute velocity (C ). The.tegt loop from Which the original pump. development data was obtained was not constructed in a manner to allow operation at an elevated temperature or at a. suction pressure below atmospheric,' For this reason the determination of the fuel pump cavitation characteristics was delayed until the construction of a suitable loop._ The loop designed for the hot testing of the fuel pump was used for obtaining the cavitation data. The Thoma cavitation parameter' ((f)(T)(B) was determined for the final 1mpeller design as a function of the percent change in the pump maximum head at a constant flow and speed. The parameter o is -66- . L egEAme s . ' .y where H is the pump total suction pressurenmeasured‘close to the. pump eye.in‘ order to avoid frictional losses, Fvap,ls the vapor pressure of the working fluld -and H max is the pump-maximum totel head. - The results of these’ experiments are shown in Figs 2.29 through-2.31. . From these cavitation data and the data previously obtained on the:difference between the expansion tank pressure and the pump suetion pressure, Fig. 2,32 was made on the assumptlon that the suction pressure requlred t0. prevent cavie tation was independent of the liquid level in the expansion tank. Fig 2.32 e afi&“ - gives the minimum expansion tank pressure required to prevent cavitation as & function of the expansion tank liquid level and the pump speed:and flow. -67- -89~ Change in Maximum Head (%) 10 15 20 or I % '——'?fl'-.—fll-“———-&—o—o—b 2 2 =‘i"""'—: —h—a— d—A—gb———h ;’*.— e . a )/— . h . -1 / A . n ‘ / 4 ° 2 . 8/ , & / /1 & & | 600 gpm g = 0.61 = ® | 645 gpm. 0. = 0.825) A 700 gpm oo = 1.40 0 0.5 0.6 0.7 0.8 0.9 1,0 1.1 1.2 1.3 1l LS 1.7 1.8 VALUE OF CAVITATION PARAMETER @& , FIG. 2.29 CAVITATION CHARACTERISTICS OF ART FUEL PUMP WITH MODEL 32 IMPELLER AT A PUMP SPEED OF 2L0O rpm AND A 3-IN, DEPTH OF LISUID (WATER) IN THE EXPANSION TANK 1.6 -69—- Change in Maximum Eead (%) ORNL.L R-DWG. 23952 0 = +———F a4 4 ° | L‘/ . A 4/ 5 / i w a o 10 | ‘ I. ® @ | 600 gpm O:':O.hé. ® |65 gpm gz = 0,51 4 | 700 gem g7 = 0,785 15 20 = 0, 0.2 0.3 0. 0.5 0.6 0.7 0. VALUE OF CAVITATION PARAMETER O FIG. 2,30 - CAVITATION CHARACTERISTICS OF ART FUEL PUMP WITH MODEL 32 IMPELLER - AT A PUMP SPEED OF 2700 rpm AND A 3-IN. DEPTH GF LIQUID (WATER) IN THE EXPANSION TANK it gt 2 ‘fl‘:ji' -0c- Change in Maximum Head (%) 10 15 20 onu“s ® | 600 gpm -, = 0.L412 @} 645 gom L . = 0.Lh3 A| 700 gmeNy = 0,583 0.1 0.2 0.3 0.4 0.5 0.5 0.7 o; ' ' VALUE OF CAVITATION PARAMETER O FIG. 2,31 CAVITATION CHARACTERISTICS OF ART FUEL PUMP WITH MODEL 32 IMPELLER AT A PUMP SPEED OF 3000 rpm AND A 3.IN, DEPTH OF LIQUID (WATER) IN THE EXPANSION TANK _lé- Expansion Tank Gas Pressure Minus V:)apor Pressure of Working Fluid (£t abs 26 2k 20 18 16 1L )2 10 ORNL-L R-DWG. 23954 700 gpm Kn. Expansion Liquid Level 3000 rpm 1 in, e , ‘ 1/2 in. 600 2 in. 1/2 1in, /\ gpm _ 3 m.]\ 1 4n 1 in, 2 in, : 3 in, . 00 2 in 3 in, 700 gpm r 700 gpm ] /] 600 / 645 gpm 6L5 gpm T~7500 gpm 3000 rpm 2700 rpm- _ 2L00 rpm FIG, 2,32 . MINIMUM EXPANSION TANK GAS PRESSURE REQUIRED TO PREVENT CAVITATION IN TYE ART FUEL PUMP SECTION IIT DESIGN OF AN ATTITUDE STABLE XENON REMOVAL SYSTEM DESIGN OF AN ATTITUDE-STABLE' XENON REMOVAL SYSTEM In addition to the design precepts and requirements of a stationary up- right xenon .removal system, one suitable for application to aircraft must meet the additional requirement of attitude stability (9)_ An over- all picture of” the original attitude stable X-R system is given 1n Fig. 8. 1. Illustrated particularly are the swirl chamber in relation to the rotary elements, the interconnecting parts, and the direction of fuel passage The swirl chamber acts as an expansion tank in the fuel system and - as & processing tank for. xenon stripping. All other’ required functions ‘are carried out in the-rotary assembly,valcross section of which is-shown‘in'Fig.’3.2, which shows a full scale test model of the rotary assembly. The nomenclature of parts as used in the description is identified in Fig 3.2. o Fig. 3.3 depicts the system as a. hydraulic circuit : The numbers are only illustrative tonp01nt out the pressure interdependence between the X-R system and the main fuel circuit. Circuit Description The bleed circuit is most readily understood by referring to. Figs 3.2 and 3.3. In the description below the circuit will be described in zones, as follows: Zone 1l: Centrifuge, core header region and nozzles (pressure control) - Zone 2: Nozzles, swirl chamber; and flow split stators ' (primary swirl) Zone 3 Flow split stators, swirl pumps, and swirl chamber (secondary swirl) . | Zone l. .System Pressure Control Circuit.wise, it is best to consider the centrifuge as a radial vane centrifugal pump (see Fig. 3.4) discharging fuel to the core header region, which serves as a mixing chamber into which processed fuel is discharged 60 ORNL- TR YA TG 5 PRt iy a0y System Fig. 3. Attitude-Stable Xenon Removal ORNL-L R-DWG. 23955 ? - ‘UPPER SILINGE{R SEAL ? AN\ Y, / N _ /§ Y - / I > K / | T 2 AT 1Y I / \ o 1 \§ TR : LCENTRIFUGE LW © / | | AN (e SLINGER / | ‘ | seac SRR % CENTRIFUGE 2 9 (I % {.I,llmll-'l-'l_s % o _ 2 CENJ(I-‘;‘II_FEUGE FIGURE 3.2- ATTITUDE-STABLE: XENON REMOVAL SYSTEM ROTARY ASSEMBLY -77- | rw B - [1 ~ _— | Dp~60psi I | : . SLINGER 777 2 _ AL , 7 . ; : wl D p~30 psi Z 513 ‘ 1 — & ‘ ' : | 2NN XENON- REMOVAL SYSTEM | 7 & ; : | SPLIT BLEED. CIRCUIT < | g - , | 7 SPLIT /A L - Z N SWIRL STATOR * _ 2 ) x L CHAMBER L 2 . W NOZZLES 8’ °© 4 CENTRIFUGE 4 ; 512« I SLINGER L/ “ 2 a W SEAL Ap~ 41 psi * > ‘ - . Ap~ 43 psi —e | 40psi b . . | C A CENTRIFUGE N~ Y O (” 7 - FUEL PUMP L HEAT gxcHANGER REACTOR CORE -~ MAIN FUEL CIRCUIT \ ‘ NUMBERS DENOTING PRESSURES ARE ONLY ILLUSTRATIVE. FI1G. 3.3 SCHEMATIC OF ATTITUDE-STABLE FUEL CIRCUIT e 78 -..~;_~_.,._, E | : UNCLASSIFIED ORNL-LR-DWG. 23956 EXPANSION TANK - GORE HEADER (PLENUM) S EEEE G ISR IS TS AN TeE—" e e ST S— FIG.3.4- ATTlTUDE'STAw,STEM PRESSURE CONTROL -79- and from which an equal'volume of fuel is bled for processing. Two nozzles direct the‘bleed flow into the expansion - tank as'tangential jets to induce ‘a swirl, These nozzles represent the major resistance in the circuit.and, | together with the centrifuge speed, determine the bleed flow rate and the header region pressure. Since the header region is also the discharge region for the main fuel pumps (shown dotted in Fig. 3.4) it is seen that the bleed circuit in effect determines the main pump discharge pressure and thereby the reactor system pressure. By making the pressure drop across..the nozzles equel'to the mainfsystem fuel pressure drop, the main fuel pump suction pres- snre can be maintained at swirl chamber pressure (controlled by the off-gas system). Zone 2. Primary Swirl This part of the bleed circuit serves the primary purpose of agitating: the fuel. The fuel in the swirl chamber is made to spin (reinforced by the circuitsof Zone 3) by the tangentially directed discharge of.the fuel from. : the nozzles (see Fig. 3.5) and careful contouring of the swirl chamber walls. The resultinguhigh peripheral’velocities‘induce violent agitation; mixing, | . and gas entrainment in the spinning fuel. The spin also serves to make the fuel level insenstive to attitude. As seen in Fig. 3.5, ducting designed into the .swirl chamber wall directs fuel into two protruding ports in the | pump hou51ng. The - fuel scooped up by the ports is directed to the flow split stators which divert such fractions of the fuel axially downward into the 'centrifuges as' is required to balance the centrifuge discharge’ (see Zone l) ~ Under normal operating conditions the centrifuge is simply flooded, the ex-' cess being diverted upwards into Zone 3. Zone 3. Secondary Swirl | The upward directed excess of Zone 2 is passed to two auxiliary centri-- fugal "Serl" pumps mounted on the main fuel pump shafts. The discharge of these pumps is directed by suitable porting and contouring (see Fig. 3. 6) to reinforce- the spin in the swirl chamber. This additional energy is.not re- quired until the level in the swirl chamber rises to about 40% full. : By -80- - 18- . UNCLASSIFIED - . ORNL-LR-DWG. 23957 - FIG.3.5- ATTITUDE-STABLE SYSTEM PRIMARY BY-PASS SWIRL -zg- . UNCLASSIFIED ORNL-LR-DWG. 23958 FIG. 3.6 - ATTITUDE - STABLE SYSTEM SECONDARY BY-PASS SWIRL virtue of its elevated position the-SWirl pumps automatically cut.in when the level in the swirl chamhéfrises sufficiently to prime the swirl impeller; The swirl'impellér aisomserves to vefit the gas'liberated from fhe fuel in the centrifugeé. A gas passage conneéting_the swirl impeller suction,side with the centrifuge center is formed by.é shroud around the shaft as shown in Fig. 3.7. | ‘ Dynamic Seals | - To complete.the.circuit.descriptioh two important seals will be described briéfly. Both seals are essentially radial vane centrifugal impellers arranged to pump agalnst the pressure of fuel enterlng the impeller from what is normally the dlscharge end of the blades. ' In the case of the upper slinger seal, its function ‘is to prevent any fuel "leaking past the swirl pump from moving up the pump.shaft toward the bearings. The upper slinger seal impeller radius exceeds the swirl 1mpeller radius in order to prOV1de some safety margln and to permit 1nverted operatlon.» The centrlfuge sllnger seal is an integral part of the centrifuge (see Fig. 3.8) and serves to limit the amount of recycle around the top of the centrifuge. The blade diameters must be such as to .develop nearly centrifuge pressure but never to exceed it. A more detailed.discussion regarding design and performancé is appendéd. Test Development The chronological stages of the development of the ART-XR system are summarized diagrammatically in Table 3.2. Excerpts of a summary by A. P. Fraas of the development of the first six models are abpended to this report. These modeis first indicated the need'of_emfiloying a centrifuge and pointed out several instability problems which led to the decision to build the swirl chamber as a separate physical entity.. ' ' Denoted as Model No. 7 in Table 3.2 and 1llustrated in Fig. 3.10, such a swirl chamber was tested independently of the pumps by supplying feed from the plant sanitary water supply line. Tests were run at 2.6, 4.6, 8.9 psig - -83- UNCL ASSIFIED ~ ORNL-LR-DWG. 23959 < 7 — FIG.3.7- ATTITUDE -STABLE SYSTEM PUMP SHAFT SHEOUD, .fl‘-&‘r | -Efi'gfi 'UNCL ASSIFIED ORNL-L R-DWG 23960 el S RS f ////////// N ///////// Fl1G. 38 ATTITUDE STABLE SYSTEM CENTRIFUGE o SLINGER SEAL \ - inlet pressure with various resistances in the discharge lines simulating lift and. friction losses., The test data are summarized in Table 3.3. In conducting the tests it was observed that the 1iquid 1evel in the swirl chember sought 1ts own equillbrlum 1eve1 for any combination of inlet and back pressure. Efforts to change liquid level independently.of pressures,. by adding water directly to the swirllchamber, were'unsuccessful'-either.the level reverted promptly to its original equilibrium level or a non-equilibrium ' condition resulted with the level continuously rising to overflowing. The nature of this swirl instability was-subsequently analyzed.and is schematically represented in Fig. 3.9 in terms of'an.energy balance. The sequential dependence is as follows: 1. The amount of liquid retained by the centrifuge controls the centrifuge delivery pressure. 2. The centrifuge delivery pressure acting on the nozzle deter- mines the flow rate out of the centrifuge and into- the swirl 'chamber. N ' 3. With the flow rate and pressure established the energy input to the fluid is established ' L, For equilibrium conditions the input~energy,must equal the losses. The losses can be broken down intos ,a.'-nozzle loss - a function of flow b. swirl chamber discharge loss - a function of flow . c. 1lift - fixed d. swirl chamber agitation loss - a function of swirl chamber | liquid level ' :. _e. swirl chamber drag loss - a function of swirl chamber | liguid level. 5. Since the swirl chamber also serves as the expansion volume for the fuel systems, & rising fuel temperature will initially cause a rising level in the swirl'chamber. . This changing level by way ‘the 44 and e dependence above, affects the system equilibrium in a manner dependent on the initial equilibrium liquid level. . -86- —18- ' UNCL ASSIFIED ORNL-L R-DWG, 23961 r--—-—~~— -~~~ ~~—"~""~>"~—>"~>"™"™"™>"™"™""™>""""“""™"™"™"">""""" """+ [ | LIFT > | jo z | I &lu | T — , |2 | ) . SWIRL |GHAMBER : ccoom )% ’ | DELIVERED o ‘ I - FIXED 2 Jjo L_{GENTRIFUGE ————w=< NOZZLE | | ST e A | - WALL | . aciTATION o DISCHARGE FRIGTION | ’AGITATION , ~_ T - ;Loss ) = L0SS HIGH LOW = LEVEL LEVEL - @ ___________ I ‘i I °/6 FULL - 1 9. FULL 1 | A . | _- WALL AGITATION | |- "HAaFRICTION N3, 1 N o A | a \ o N\ I o R | S %% FULL .. | I ' ‘ I L — | EQUAL"Q" FOR EQUILIBRIUM}— — — — — ] FIG. 3.9- ATTITUDE - STABLE SYSTEM SWIRL CHAMBER ENERGY BALANCE ) From Fig. 3.9 it is seen that the agitation,losses_decrease with rising level and .is the predominant loss at "low" liquid levels. The drag losses, however, rise with 1ncre331ng level and predominate at "high" liquid levels._ Instabillty.A. _ With the swirl chamber at an initially "low" liquid level, a rising level (due to a rising fuel temperature);reduces'the _ swirl chamber loss and thereby inereases the flow out.ef the swirl chamber. With the centrifuge cup initially full this. does not, however, . increase the flow into the swirl chamber. In effect, then, the swirl chamber reverts to its initial - stable level, transferring the extra inventory to the space 7 - above the centrifuge.. As born out by test experience, the swirl chamber "refuses to act as an expansion chamber". Tnstability B: | With thé swirl chamber at an initially "high" liquid level, - & rising level increases the swirl chamber loss and thereby decreases the flow out of the swirl'chumber_andvinto-the. centrifuge;cup. With ite supply reduced tne centrifuge 1iquid ‘_level'drops'thus reducing the energy supplied to ‘the swirl chamber. With energy 1osses increasing {due to rising 1iquid level from (a) the extra inventory due to temperature rise and (b) the volume no longer retained by the centrifuge) and input " energy decreasing, the system is more and more unbalanced, re- sulting in a run awaylcondition ae bOrn out by test experience.' The. swirl chamber may overflow and the centrifuge cups will lose their primes.\' At pressures above 7 psig suff1C1ent flow was obtalned to give good spin and agitation. However, the tests indicated definite 1nstability tendencies for certain combinations of liquid level, back pressure and inlet pressure. -88- ORML-LR-DWG. 23962 NOZZLE S SN NSNS N | | SN SRS AN OOV NN NNNN i 2z &l Ty - / T T b [ —— g FIG. 3.10- ATTITUDE - STABLE XENON REMOVAL SYSTEM TEST MODEL NO.7 gx L Ze e .'; B -;14 IR The 1nstab111ty tendency was more clearly apparent in Test Model No. (Table 3.2) when the swlrl chamber was placed’ in series with one centrifuge. The analysis of the swirl chamber indicated a marginal energy balance be-_ o tween the energy in the fluid issuing from the jets .and the fluid friction los=- ses in the swirl chamber. At some minimum swirl chamber level a metastable 0perat1ng condition was attained. ;Any.small-increase"in frictional resistance unbelanced the system to the point where the centrifuges lost all prime and freelylpassed air. It was not possible'to reestablish stable Operating con- ditions. In Test Model No. 9 every effort was made to reduce connecting line friction losses. Furthermore, the swirl chamber was raised such that the centrifuge could be supplied by gravity. However, even under these conditions, the system could not be stabilized over the full range of liquid levels. It was, therefore, decided to augment the energy supplied to the swirl chamber by an;auxiliary impeller as shown in Model No. lO,.Table’3.2. The centrifuge discharge was passed upward through the clearance space between the centrlfuge cup and. the wall where it joined the auxiliary bypass fluid discharged by the impeller mounted on. top of the centrifuge cup. . Although the centrifuge flow rate could not be determined, this model was able to remove bubbles from the plenum chamber below the centrifuge. Swirl chamber operatlon was stable. However, no appreciable system pressure could be developed. ‘ : The performance of Model 10, Table 3.2 and the anslysis leading up to it indicated that the system was. sufficiently'well understood. to make an attempt at an improved de51gn appear promlsing. A two-pump integral plastic model was, therefore, de51gned incorporating the following features: 1. Locating the SW1rl chamber above the centrifuge cup to permit priming. 2, Incorporating an auxiliary impeller wh1ch by superimposing a re- C1rcu1at1ng Tlow supplied excess energy to the swirl-chamber. !3. _De51gning a configuration such as to make passages between swirl chamber and centrifuge assembly as short as possible to minimize friction losses. ) 4. Introducing the centrifuge delivered flow-into the swirl.chamber . through highly restrictivé nozzles to maintain the system pres-- surefand to provide .spin at low swirl levels. The assembly as originally built is 1llustrated in Fig. 3 11. The development of this unit to the final working model involved changes' in the rotary assembly only These are illustrated in Table 3 2 Models 11 - through 15. The problems encountered were primarily sealing problems to prevent reciroulating short circuits. This development work is outlined in Table 3 2. Finally, the entire assembly wags mounted on-a rig to permit inverted operation. The performance is summarized in Table 3.k. The final critical dimensions for Model 15, Table 3.2, are listed in Table 3.4, The final design'is illustrated in Fig. 3.12. ' System Analysis (Refer to Fig. 3.3) The manner in which the X-R system controls the main system pressure is most easily seen from a study of the characteristic curves of the two systems. Figs. 3.13 and 3.1k illustrate the characteristic curves for the main fuel pump circuit and bypass flow circuit respectively. The ordinates.are—drawn to the same scale but intentionally do not have their origin defined. The labeled points have the following physical significance: (a) fuel pump discharge pressure (b) centrifuge discharge pressure (c) centrifuge inlet pressure (d) fuel pump suction pressure. As previously described and illustrated in Fig. 3.4, both the centrifuge and fuel pump discharge into a common header and therefore points (a) and (b) must be.at the same pressure. Secondly, since the centrifuge inlet, because of . its gas interface is_essentiallyrat swirl chamber helium pressnre, i.e., since point (c) is at p = O (reference), all the other pressures are fixed relative to (c) as shown in Fig. 3.15. .-91"' fic‘gi Ivlv,i - The system pressure, particularly the pump suction pressure, is therefore controlled by the helium. pressure and the . bypass flow pressure drop. The bleed flow Ap is primarily controlled by the nozzle resistance. This nozzle is made highly restrictlve 1n order to keep the pump suction pressure high enough to - prevent cavitation. Any pressure drop across the centrifuge cup holes reduces the syétem pressure by an equivalent amount. Thus any condition that simulates . an increased flow ;hrough'fhé céhtrifuge-cup holes effects a reduction in the - - system pressure level. .-92; Hnas Table 3.1 Model 15 - Performance Evaluation. Test Condltlons A. Speed range 700 rpm to 1800 rpm 'B. Storage volume range 20% full. to 80% full 'C. Attitude: 1) upright, 2) incllned, 3) inverted. Performance Criterion Test Condition . Evaluation 1. Expansion Volume Stability | A - Completely stable : , : Completely stable c ~ Completely stable 2. Agitation | A | Varies from fair to excellent as speed’ " increases. B , Best at low volumes - generally good. c Good. 3. Upper Slinger Seal Effectiveness A Complétely effective B Completely effective C Completely effective L. System Pressure A Within 5% of theoretical ' - ‘ expected. B o Drops with dropping vol- "~ ume when less than 20% - C ' . Drops off completely at low volume. 5. Bubble Removal Effectiveness A © Good - best at lower : o . "~ speeds. B ‘ Good - best a high liguid ' -+ levels. C o Fails at low volume and ~low speed. 6. Power Required A - Increases with speed. B - Increases markedly with o N : : high liquid level. o e 7. One Pump Out , | .. Loss of system pressure; poor agitation. _93_ SEQUENTIAL DEVELOPMENT OF ATTITUDE-STABLE XENON REMOVAL SYSTEM MODEL . o - o MAJOR CHANGES FROM NO. REFERENCE FLQW PATTERN . PREVIOUS MODEL EVALUATION. N o SLINGER 2 Fig. A.1l 1) Swirl insufficient to STORAGE remove bubbles. | | : PUMP SWIRL 2) Agitation good T - r TN 3) Slinger effective. r—— ==~ i /T ] T 4 | . * ' | L 2 Fig. A.l Jet added to assist swirl., Swirl stili_insufficient to remove bubbles. 3 Flg. A.2 Centrifuge added above pump Insufficient 1ift avail- impeller. able for getting fluid into centrifuge. 4 Fig. A.2 Scoops added in swirl Effective up to 100 rmm ’ chamber to assist lifting only. fluid into centrifuge. 5 Fig. A.3 1) Peripheral airfoil scoops | 1) Unstable operation. added. 2) Passed air bubbles. 2) Expansion - and swirl ' chamber built as unit chamber (baffle omitted)}. 3) Centrifuge discharge into impeller discharge region 5 Fig. A.> Lebyrinth seal added between | No improvement, . punp volute and swirl chamber, Decision: 1) To separate centrifuge and swirl chambers. 2) Hew models not to include simulat.d main circuit. Table 3.2 (contd.) MODEL MAJOR CHANGES FROM 10. REFEREICE FLQW PATTERN - PREVIOUS MODEL EVALUATION 7 Fig. 3.10 Isolation of swirl | Test of swirl chamber chamber from pump |with city water presswure impeller. . . supply and variably . o restricted discharge. ‘See Table 3.3. SWIRL CHAMBER C——= 8 Fig. 3.10 __+__ Swirl chamber - 11) system unstable as - | connected to single liquid level is i I centrifuge unit. changed. " )‘, 2) Centrifuge loses prime. [fifi' r_~1 _ (& 5 | g rig. 3,10 = H Swirl chamber Marginal performance due . o raised with respect| to insufficient spin T B to centrifuge. energy . . =] ) . R . : ’(’ Flow reduced by - _|plocking one nozzlel _F _’_ " - * . . - t ‘ ! 10 G Discharge through |1) Effective bubble Fige 3.10 P J | jj] ) connection opposite impeller mounted on centrifuge. separation. 2) No system pressure. %) Good spin except at very high liquid level. L) Centrifuge can lose prime because of ‘ limited spin energy and excessive fluid friction losses. Decision: .To build plastic model contaihing’twb centrifuge assemblies with integral swirl chamber between shaeft centers. B ‘ ‘ ' -95- Table 3.2 (contd.) MODEL MAJOR CHANGES FROM i oV NO. REFERENCE FLOW PA'I'I‘ERN PREVIOUS MODEL EVALUATION 11 Fig. 3.11 1) Two centrifuges Slinger seal ineffective. : feeding common swirl ' chambers. 2) Auxiliary impeller to reinforce swirl “at high volume levels. 12 Fig. 3.1 Slinger replaced by 1) Slinger impeller ' slinger impeller. effective. 2) Low pressure in main systen. ] A— 13 Fig. 3.2 | ' Impeller edded above 1) System pressure ' ~ h centrifuge. satisfactory. ?;/ f ' 2) Bubbles Ly-possed or %% passed through 2 7 centrifuge. YLllls 74 \‘\ E /y_ 1h ' Special test rig to Labyrinth seals in- investigate sealing effective. i . arrangements. ' ‘ gZzzz23 : [ 1 [ \ 1 15 rig. 3.1 Same as Model 13. Identical with Model 17| Satisfactory operation 3.2 but with rebuilt centri- with respect to: 3.12 fuge and carefully a) Dubble scparation " |verting rig provided. - machined inpeller. In- b) Stability c) System pressure d) Attitude changes Static Head Table 3.3 Model 7--!Perfdrmance Evaluation Spin Velocity Swirl * Inlet Back Measured 3/4 in. above Chamber Agitation Stability Pressure Pressure scoop § ft/sec Flow Efficiency Evaluation Evaluation £t £t _ cTs 'J_i'% — 6 0.52 5.25 0357 '1‘16;7 | fair stable 0.605 ' 5.78 .0357 -18.2> poor stable unstable B - | | - unstable 10.6 0.52 3.38 .0k88 13.4 good ' stable 0.605 3.02 W73 13.7 good " stable 0.687 322 460 1h.1 fair stable 0.77 3.0k 0473 15.2 poor stable 0.855 5.25. OLT3 16.1 noné stable unstable unstable 20.6 0.52 - 8.9 .067 10.8 -éicellent stable 0.687 8.181_ 067 . 11,6 . excellent . stable 0.935 - 9.62° 067 12.9 excellent stable 1.27 _9.62 067 1k.5 excellent stable 1.40 9.62 067 15.2 fair stable unstable unstable * Swirl chamber efficiency = Total Head of Leaving Fluid Total Head of Entering Fluid _97- Table 3 A - . Model 15 - Dimensions Bléed Flow Rate - .0297 ft3/sec Swirl Chamber Nozzle Orifice - 1/16 in. x 5/8 in. Expansion Volume required N 0.45 £43 Expansion Tenk Dimensions. _ 0.D. 1k in. | I.D. 5 in. Height 5.75 in. Centrifuge Cup Size B o 0.D. 5.75 in. ’ Lo I.D. 3.50 in. Centriffige Holes ' : + © Number | 8 | | | | Diameter - 1/4 in, Centrifuge Slinger Seal -~ radial vane I.D. 3.50 in. | | ' " radial vane 0.D. 5-5/8 in. —66— ORNL-LR-DWG, 23!63 SECTION A-A = .. : .SECTION B-B L FIG. 3.11- ATTITUDE-STABLE XENON REMOVAL SYSTEM TEST MODEL NO. I - iz A B S SECTION B-B S % =001~ N\ 5o o -Et-r; -!;'! UNCLASSIFIED ORNL-LR-DWG. 23965 - PUMP DELIVERY PRESSURE | l o | ~ SLIPPAGE | 85 4r HoLeg r5-———=———-- : | CENTRIFUGE | DELIVERY AP AP : PRE SSURE | | | 0 .44 Q 0 0268 Q FIG. 3.13 | FIG. 3.14 ATTITUDE -STABLE SYSTEM MAIN ATTITUDE - STABLE SYSTEM FUEL CIRCUIT CHARACTERISTIC BY-PASS CIRCUIT CHARACTERISTIC FIG. 3.15 ATTITUDE -STABLE SYSTEM PARALLEL MAIN FUEL AND BY-PASS FLOW CHARACTER!STIC | -101- Appendix A SUMMARY OF WORK (PO JULY, 1954) LEADING TO THE DESIGN OF PUMPS FOR THE CFRE B SUMMARY OF WORK (TO JULY, 195k) LEADING TO THE DESIGN OF PUMPS FOR THE CFRE | (Refer to Models l through 6 Table 5 2)* “The face-type gas-seal sump pump is attractive for use with high tempera- ture liquids because, by inserting a heat dam, the seal can be operated at - conventional temperatures with conventional materials, for example, hardened steel'and-graphitar~in‘a petroleum ‘0il bath providing both lubrication and cooling. . An experimental unit was prepared in the. summer of 1950 on the thesis that the free surface.could be maintained in a centrifugal field and hence could be made insensitive to -the vériations_in pump attitude and negative "g" loads to‘ be .expected in flight. A particular feature of the design was that it should: act . to de-aerate the system -- a feature considered essential in a full-scale aircraft type pump. . The principles of operation were demonstrated by operating the pump with water in the fall ofkl950~ap@fapprokimatélyLSOersofqurther@testing with sodium at temperatures up to 1000°F was carried out in fhe sumer of 19510. . As the design:of the ART progressed it became evident that there is a strong incentive to remove xenon, and that this might be done by agitating the fuel in the -expansion tank so that it would entrain bubbles., This would serve both to: increase the surface area by & large factor and to bring a high profiorf tion of the fuel close to a surface within a short‘period.of'time.- Thus ‘1 or 24 of the fuel could be by-passed through the expansicn tank and the bulk of the xenon stripped before returning it to the ‘main stream. . The arrangement had the additionsl advantage that it would avoid overheating of the fuel in the expansion tank as a result of fission product decay activity. While a separate agitator could be provided, it was felt that the number of moving parts should be minimized and that the pump impeller shaft could serve a double purpose. The notion of utilizing & centrifugal field in back of the impeller to stabi-~ lize a free surface there had Worked-so#nicely in the pump.mefitioned'earlier *A. P. Fraas, abstract taken fram "CFRE De51gn and Development Program.Handbook, Prcject No. 3. . . P | , C . . ~105- that a modification of it to effect xenon removal seemed promising. A simple model of work, glass, ‘and plastic was prepared to 1nvest1gate the possibilities. A section through this model is shown in Fig. A.1l. The model was designed to operate with the fluid in the expansion tank swirling under the impulse of fluid accelerated and thrown off by the impeller. A coneiderably higher swirl velocity vas .expected -in the annulus just above the impeller, and a diaphragm was inserted to separate this region from the main part of the expan51on tank above. It was hoped that the swirl veloc1ty‘1n this lower region would be high enough 50 that the outer periphery_Would‘operate_free of bubbles. Thus, fluid leaking upward through the labyrinth seal between the impeller and the. casing would pass into the expansion tank where it would be violently agitated and mixed with»bubbles, " This bypass flow from the high pressure region at the impeller discharge into the low pressure region in the expansion tank would be returned to the impeller inlet through holes in the periphery of the swirl chamber under the expansion tank. It should be noted that the resistance of the main fuel system was simu- lated in the model of Fig: A.l by a set of twelve l/h-in.-dia orifices in the ‘ lower of the two disks boundihg the annulus into which the.impeller discharged. Thus the main stream flowed directly from the impeller discharge through these | orifices back to the pump inlet while the bypass flow through the expansion " tank followed a more devious route. A slinger was mounted on the impeller shaft :just under the roof of the expansion tank to prevent fuel from entering the clearance between the shaft and the heat dam between the expansion tank and - the gas seal. | When the model was built and tested, it was found that it performed substantially as desired except that the swirl velocity in the expansion tank was not adequate to'centrifoge.the bubbles out of the f£luid returning from the expansion tank to the impeller,inlet. Tangential jets in'the floor of the swirl chamber were drilled from the high pressure reglon at the pump discharge to give a higher swirl velocity in an effort to get better bubble separation. No improvement was obtained. Two features of the pump did perform surprisingly well, however. The slinger was eminently effective in preventing the entrance of fluid into the region between the slinger and the seal, and the agitation and bubble entrainment of the fuel in the expansion tank gave good pramise of - -106- =01~ . ’ o . Omjvu PLYWOOD ~SLINGER GLASS POT ORIFICES TO SIMULATE. EXTERNAL SYSTEM (2 EQUALLY SPAGED. - '- ’ ] | | —— . m— i — pm— e mmmaniss e — cv—" S e e— e m—— e — A e — — — ——— —— o— g —— - —— e — it e RUBBER SEAL FIG. A.1- PRELIMINARY MODEL NO.1 OF AN ATTITUDE - STABLE XENON REMOVAL SYSTEM BB removing the xenon. _ The results of this amalytical work clearly showed the need for a centri- fuge operating with at least as high a.tip'speed as the imreller; On'this basis the impeller was modified as indicated in Fig. A.2 to include a centrifuge cup on its re#r face. Holes were drilled in the periphery of the cup so that fluid would’discharge'into the swirl chamber. This arrangement was found to be much superior to the first model, but it proved difficult to get fluid to flow into the centrifuge cup at speeds above about 1200 rpm. Various scoops were tried, including airfoil-shaped radial structs. The latter were hollow with inlet . scoopé in their leading edges near the outer periphery so that ram pressure from swirl invthe expansion tank vould force fluid through the strut and into the centrifuge cup. The best of these scoops workedlfairly well up to about 1400 rpm, and would give positive de-geration at speeds fram about 800 to 1400 rpm. A careful re-exemination of the system led to the decision to make several major changes. Since the first model had deteriorated considerably in the course of the several modifications, it was decidedlthat a neW‘fiodel should be built as shown in Fig. A.3. The baffle or diaphragm at the top of the centrifuge cfip was cmitted and the fluid allowed to discharge from the centrifuge through 3/16-in. holes into the regions behind the impeller vane tips. It was expected that cmitting the diaphragm would increase the liquid level range in the expansion tank for which it might be hoped that the pump could operate. The centrifuge cup was made as long as possible in the space available to minimize the radial through=flow:véloc¢ity. 'A new type of scoop was designed to fit in the bottom of the expansion tank to lift the fiuid through the annulus between the tank wall and an internal liner, and deliver it to the centrifuge. This arrange- ment was found to be. satisfactory in every respect except that, at high speeds, if the liquid level dropped so that the free surfacé of the expansion tank vortex dropped below the top of the centrifuge cup, the fluid would be thrown off the sheel intermittently and long streamers of air would extend down to and into thevlabyrinth seal gt the top of the impeller and severe bubble entrain- ment in the main stream would result. -108- ' Separating the expansion tank from the pump impelier seemed;to bé a pramising'arrangement if sufficient'agitatibn could be produced by jet action to take care of both xenon removal apd pumping the bypass flow from the expan 81on tank to the centrlfuges. A c1rcular tank between the two putps as shown in Fig. 3.10 was designed to accomplish both functions. Tangentlal Jets in the bottam‘of the tank were de31gned to allow fluid to discharge ffom the'high‘ pressure region at the pump discharge into the expansion tank, which would run at essentially’pump impeller inlet pressure. The large scale swirl induced in the expansion tank could be made fo entrain gas bubbles through the use of aspirator tubes. The ram.pressuré_on'tangential7scoops in the outer walls should servé-to pump fluid through thé scoops to the centrifuge chamber. | -109- ORNL-LR-DWG, 23967 PLYWOOD SLINGER ' CENTRIFUGE CUP LASS POT ORIFICES TO SIMULATE EXTERNAL SYSTEM 12 EQUALLY SPAGED. . RUBBER SEAL FIG. A.2- PRELIMINARY MODEL NO.2 OF AN ATTITUDE-STABLE XENON REMOVAL SYSTEM ORML.LR-DWG. 239468 =LLL- FIG. A.3- PRELIMINARY MODEL NO.3 OF AN ATTITUDE- STABLE XENON REMOVAL SYSTEM Appendix B CENTRIFUGE THEORY AND CALCULATIONS CENTRIFUGE TEEORY AND CALCULATIONS The centrifuges serre two purposes essential to the sys‘bem° First, they remove the mixture of helium and enert fission gases from the fuel before it -zreenters the main fuel system; and secondly, they serve as pumps to return the fuel to the main system under pressure ' The abllity of the centrlfuges to remove the ges bubbles from the fuel’ depends primarily upon four factors: the speed of rotatlon of the centrifuges, the radial dlstance to the p01nt of dlscharge, the velocity of radial flow 'through the centrifuges, and the size of the bubble to be removed. | The relatlonshlp between these. factors may be shown as ‘follows: The drag force on an obJect W1th motion relative to a fluid ‘is given by the equatlon h o | : | pve | Fd = _cd_E-é—A where - F, = drag force (1b) Cd = coefficient of drag f = density of the fluid (llkft 3) - V. = velocity of the obJect relatlve to the fluld (ftcsec‘;) | _ _ g = gravitational acceleration (ft-seofe)_ A = area of object projected to the flow (fte). The centrlpetal (flotatlon) force on a bubble in a rotatlng fluld (assuming dens1ty of obJect is negligible) is given by the egquation k3 w9x Fc = 3 r f’ where S ‘ = centripetal force (Ib) = bubble radius (ft) . . 1 ',{-Q"-r -(Sec ) = radial dlstancj from center of rotation to center of bubble (fi). -115- angulargye % < = R i For the relative motion between the centrifuge cup and the bubble to be zero, the two forces must be equal. _»Cd_‘zg__j_ A= ll/sflrp-—*‘- and V = 3C The value of C; is a function of Reynold's Number* and is as shown 1n Fig. (10) a B.l. Figs. B.2 and B.3 are plots of radial liquid velocity versus bubble radius for..zero relative motion between the bubble and the centrifuge cup. The radial distance to the bubble is assumed to be 2~ 7/8 in. in each case. The above analys1s indicates several desirable des1gn features, namely, a large diameter centrifuge to increase the centripetal force on the bubble and also reduce the radial velocity, a tall cup to reduce the radial velocity and a high speed to increase the centripetal force on a bubble. The diameter and height of the cup are limited by shielding considerations and the speed is limited by the main fuel pump design. Another method of increasing the effectiveness of the centrifuge is to increase the area of the holes that feed fuel back to the main system. This reduces the radial velocity of the fluid at the periphery where the centrifuge is most effective. This also reduces the tendency for the radial flow through the holes to dlsrupt the.pressure field set up_byrthe rotating.liquid. It has been calculated that any bubble smaller than approximately 0.003 in. radius will passfout of the ART centrifuge if it reaches the exit holes when operating at 2700 rpm on Fuel 30 (see Fig. B.2). It is probable that there exists a point. at & smaller radius and lower radial velocity where bubbles smaller than 0.003 in. radius will be screened. The- number of these small bubbles which enter the system is a function of the surface tension of the liquid. With clear water the number of bubbles entering the system was very small,. By adding a small amount of soap, thus decreaSing the surface tension, the system became very cloudy from the air bubbles pa551ng through the centri- fuges. - All of the fuels presently being ‘considered have a surface tension ap- proximately twice that of water. The second function of the centfifuge is to déiiver‘the-sfripped fuel to thé system under pressure. The preSSure delivered-by\thé centrifuge is a" fuhétion of the outside diameter of the Cup_and either\the,diameter of the .free surface when'thecgpis not.running full'Or'the inside diameter of the. éentrifuge top when the cup is flooded. The pressure buildup.in a rotating fluid with no flow is identicel to that of & radial vane pump. As the thorough flow increases, there will be some'Slippage between the cup and the fluid;-giving a pumping characteristic similar to that of a pump with backward swept vanes. There will also be'a’pressure drop across the outlets frbfi the éentrifuge de- pending on the through flow and hole sizes. -117- -gi 1- Drag Coefficient Reynolds Number FIG. B.1 UNCLASSIF IED ORNL-LR-DWG. 23959 8 1 1 CORRELATION OF IRAG COEFFICIENT AND REYNOLDS NUMBER FOR BUBELES IN WATER =slL- ‘Bxit Centrifuge Velocity (ft/sec) FIG, B.2 BUBBLE RADIUS (in.) o _ o ' 1/ \ F\lel-ll«h -'-'-'--_.______; Fuel—_‘ifi - Fuel-] '\\ 4_—______—___________4—-——"""'— / Water /’ —\ -__-__-—__—____‘ / lé// é 5,00 0.0 5.01 0,01 5,020 .02k 0,078 5.032 EFFECTS OF-BUBBLE RADIUS AND EXIT VELOCITY ON CENTRIFUGE EFFECTIVENESS AT 2700 rpm WITH A CUP RADIUS OF 2-7/8 IN, -0z1- Exit Centrifuge Velocity (ft/sec) ORNL-L R-DWG. 23%1 - e —— ____—-__“-_'—'_‘—-u—_ Fuel ~ : Fuel-3) Ea— ‘ Fuel-1l '—-—._______________ ; Water ‘7 ——__--_'_'—-—-_.__ _/ A / A / / 74 / g / 0.00k 0.008 0.012 0,016 0.020 0,02l 0,028 0.032 BUBBLE RADIUS (in.) FIG, B,3 EFFECTS OF BUBBLE RADIUS AND EXIT VELOCITY ON CENTRIFUGE EFFECTIVENESS AT 1350 rpm WITH A CUP RADIUS OF 2-7/8 IN, Appendix C * DYNAMIC SEALS DYNAMIC SEALS In the X=-R system seals'are required in two critical locations, namely, on the shaft to prevent‘fuel“léakage'up into the bearing housing and in the region of the pump to prevent high pressure fuel from leaking around the centrifuges. In the césg'of’the’uppér'Slinger séal, no leakage can be tole- rated, while in the ‘centrifuge seal a small amount of leakage, properly directed, does not seriously affect the system. The centrifuge seal is essential to the system for two reasons: 1) to determine the pressure in the main fuel circuit, and 2) to control the amount of leskage. "All fuel which passes through the. centrifuge seal is recirculated through the centrifuge, thus increasing the radial velocity and consequently ‘the size of the bUbblés-which'will escape to the main fuel circuit. o - - = A very similar sealing aerrangement was used for each location. The shaft seal used was a slinger attached to- the shaft (see Fig. 1.2). The potential delivery~pres$ure of this slinger was greater than the pressure of the fuel in the region of the seal. As the theoretical pressure rise across the seal was greater than its discharge pressure, a gas pocket exists around the shaft. In the case of the centrifuge seals, radial vanes were attached to the top of the centrifuge cups (see Fig. 3.8). For this arrangement to stop all leak- age, the discharge pressure from the seals must exactly match the pressure from the centrifuges. This condition would be impossible to maintain in operation. In the actual system a radial vane impeller was designed to deliver a pressure slightly less than that from the centrifuges. The leakage around the centri- fuge cups will then be so directed that no fuel bypasses the centrifuges. As the head potential of the seal and the centrifugecare very close, the leakagé is small and the additional flow through the centrifuge is held to an accept- “able amount. Figure C.1l shows the head delivered as a function of vane length or cup diameter for various speeds. The net head delivered may be determined by subtracting the pressure corresponding to the inéide diemeter from that for the outside diameter. _123- Flgures C.2 through C.5 give the amount of leakage through the centri- fuge seals as a function of speed and the difference in pressure delivered by the centrifuges and seals for several axial clearances. These axial . clearances are the distance from the top of the seal vanes and the stationary plate above (see Fig. 3.8). | . -As the speed increases, the difference in pressure delivered by the centrl- ‘fuges and the seal impellers will increase and the flow delivered by the centri- fuges will also increase. In order- to reduce the bypass leakage caused'by the larger pressure differentiasl, it is necessary that the holes in the centrifuge cups be slightly restrictive. This restriction will also idd stability to the system by enabling it to sense and counteract any variations in leskage. . It was found experimentally that the ability of the system to screefi gas bubbles depended upon the precision with which the centrifuge seal impellers were'fiade. 'One unit in which the blades were glued in place passed bubbles very badly. Upon replacing these seals with new ones in which the vanes were milled into the top of the centrifuge cups, the system operated satisfactorily. Figure C.G_gives head and speed characteristics of a seal at no flow conditions. -124. - UNCLASSIFED -L 239;.'2 e 7 FIG. C.1 HEAD CHARACTERISTICS OF A FLUTD e ROTATING AS A SOLID BODY 2000 o 2250 8 2500 2750 — ‘0 - @ . - .8 o — 1N B 100 SEAL LEAKAGE X 100 (ft3/sec) 3.k 3.0 2.6 2.2 1.8 . 006 0.2 UNCLASSIFIED ORNL-LR-DWG. 27973 SEAL SPEED (rpm) FIG C.2 LEAKAGE CHARACTERISTICS ~126- \ ‘ -S¢al Dimensions-- 5 1/k in. 0.D., ' 2 7/8 in. I.D., ® o \ e \ * o | o ft (Seal Prlessure Drop) . b \ .\.__\. \.‘\ .-_1;5;-.-= 1.15 |ft ® o — ot @ o 1500 1700 1900 2100 2300 OF A DYNAMIC SEAL WITH AN AXTAL CLEARANCE OF 0,050 in, SEAL LEAKAGE x 100 (ft.5/sec.) ORNL-LR-QWG, 23974 UNCL ASSIFIED .'SEALVS?EED (rpm) AN AXJAL CLEARANCE OF 0.128 IN. -127- 3J+\\\ ‘ , . “ ) "\ . Seal Dimensjons: -5 l/h in. OD . N\ | " 27/8 in. ID \\\ - 1/2 in{ High 300 ‘ \ - 4.61 pt. ® * 2.6 . - o 2.2 \ '\ ’ \ v. 1.8 | L.h 5 2.3 ft. (Sedl Pressure Prop) . 10 ‘-:.-"'":--..._.__ ¢ . . @ - 4 0.6 —— B —o : - L < 0.2 . - 1300 1500 1700 | 1900 . 2100 . 2300 " FIG. C.3 ‘IEAKAGE CHARACTERISTICS OF A DYNAMIC SEAL WITH 3.4 3.0 2.6 o . N '—J . & SEAL LFAKAGE X 100 (ft/sec) |_l ™ 1.0 0.6 0».2 UNCL ASSIF [ED ORNL-LR-DWG. 23973 Sehl Dimensiong-< 5 1/k in} 0.D., 2 7/8 in} I.D., 1/2 in} high . . \\\ L.61 ft * - N Ao ° ° \. \ | ¢ . _ ~ ® . \ \. \ K\LB £t (Beal Pressure Drop) 5.““‘=-_ 1\ O —— ° \.&ft ’ ) —~———|_ .o -N-.-—._— 1300 1500 1700 1900 2100 2300 SEAL SPEED (rpm) FIG C.L4 LEAK CHARACTERISTICS OF A DYNAMIC SEAL WITH AN AXTAL CLEARANCE OF 0.2L48 in. -128- SEAT, LEAKAGE x 100 (£t /sec.) UNCLASSIFIED SEAL SPEED (rpm) AN AXTAL CLEARANCE OF 0.481 IN. -129- .\ 3.6 1 ‘ — \ Seal Dimensjons: 5 1/4|in. OD o e ] - | 27/8/in. m . \\\\\E:6l rt) oo 1/2 i%,'High | | ‘\\\\\: Sl - 2,4 }— N \\ . .6\ * o | | ® 2.3 £t. (Seal Pregsure Drop) ® i ' \ 1.2 , . ‘ o | " _ S ——— o.8j:\‘\‘ | T~ 1.15 tf. . o 0.k o T 1300 1500 1700 . 1900 2100 2300 FIG. C.5 LEAKAGE CHARACTERISTICS OF A DYNAMIC SEAL WITH UNCL ASSIFIED ORNL-LR-DWG. 23977 Seal Dimensions-- 5 1/l in. 0.D., 2 7/8 in. I.D., 1/2 in, high Axial Clearances-- 0.957, 0.837,_0.71h, 0.481, 0.2h6, 0.128, 0.050 (ip.) P~ 9 > i + 4 p 5 | 3 w0 w—- Calculated @ Experimental’ 7 8 © SEAI SPEED (rpm) FIG C.6 ZERO LEAKAGE CHARACTERISTICS OF A DYNAMIC SEAL -130- | | B Appendix D CALCULATIONS FOR THE ART FUEL PUMP . IMPELIER REDESIGN - DRy g R CALCULATIONS FOR THE ART FUEL PUMP IMPELLER REDESIGN Assuming . no prerotation, Fig. D.1 is the inlet velocity diagram for the impeller at the outermost inlet radius from Wthh it can be seen that 2 2 ' W= U+ Ca (1) where . . . D, = Impelier blade diameter at Elg. D.1 - Pump Inlet Velocity Diagram } 'the outermost. inlet edge - ‘N = Impeller speed, revolutions per sec. Referring to Fig. D.2, assume the fluid streamlines entering the blades.. curve in an axial plane with radii, R, which is a funetion of the distance from the shaft centerline, r, according to the'fo;lowing reiationship R = a +br - - (2) where a and b are cohstants evaluated'as_followsf’ 0.25 £t when r, = 0.09637 Tt R, = r, = 0.15ft Ry, = 0.08333 £t b = -3.108'and a = 0.5495 or R = 0.5495 - 3.108 r - (2a) The blade leading edges lie on the surface of a cone having a 28 deg half angle. This surface is approximately perpendicular to the throughput component of the inlet flow, Ca, therefore h | A = %%E%%%Tfl;' A = area perpendieular (3) & to throughput com- ponent of flow - As the fluid turns the corner from the inlet eye 1nto the 1mpeller blading, it follows the law of constant moment of momentum, i. e.,‘ Ca = where K = a constant (L) (ka) or Ca = and from the law of continuity " Qp = | Caan . o (5) Qp = Total flow into impeller . (includes leakage) r . 27K 2 T : QT T sin 28 deg - j a + br ar , | (53) rll . : 5, : | Ty - T ~ a '+ br, | - Qp “sin 28 deg b 2 T a+br, | Gy sin 20 deg 1 | j\ (6a) or K = —————————— 2 r. -r : a + br - 271 a 1 b + '—é In a + br ; | _ P S | o Substituting in equation (6a) K = 2.405 - 2.405 - and Ca =g 5495 =3 o6z L - Tt 28.86 . r in or Ca_” 6’59h - 3.1007 Fig. D.3 is a plot of Ca as a function of r which is the velocity profile enter- ing the impeller blades. From Flg D.3 it can be seen that at r = 1.8 in. .y = 28.86 ft/sec. ‘ U1‘.= nDlN = %26 # 45 = h2?39'ft/sec N assume to Ee'hB rpé wi = Ca’ + ui = 28.86° + li2.39° . - | Wy = 2629,8 wi"fi 51.28 £t/sec o a1Ca o, -128.86 _ ., -l . ?l' = tan 'U—l -= ‘tan m —-A! ta.n. 0.6808 {:faf\-‘ P‘n[‘"d'vq anp e SIS 13‘— &3 ~GE1- ORNL-LR-DWG. !J'WH 7 R - 1.0 1.0 - N, l 2,875 & R ‘ 2.3619 = 2.75 2.79| 1.15658 L3R f Y i Fluid Passage A 2 - 26.5 2,875 E / : s | o - 3 A o 1.0 7 B - g . ‘u < n . ., . Blade Trace 4, - 26,5° ,E . \ Five Bladed Impeller / . . A1.0 At 0 0.5 1.0 1.5 2.0 2.5 ' Inches Circumferentially A Fluid Streamline glong Shroud i | ] | B FIG, D,2 REDESIGNED ART FUEL PUMP BLADE TRACE AND FLUID PASSAGE LAYOUT AXTAL VELOCITY, Ca (ft/sec) © ORNL.LR-DWG. 239%9 L0 30 N 1.0 - 1.2 1. 1.6 RADIUS, r (in.) FIG D.3 VELOCITY PROFILE ENTERING THE ART FUEL PUMP IMPELLER =136~ 1.8 ] ek 8 2""‘&’”‘ oy d 34,25 deg ag3h.deg Py = ‘Bmv = 24,1 deg AJQh'deg' By = 19.32 deg ~19 deg - - ‘In order:forithe impeller blade to be aligned-with the fluid, the angle formed by the 1ntersect10n of the blade mean line and the tangent.of its termi- natlng circle whose center is common with the shaft centerline must equal the fluid angle. For ease¢of fabrication the trace of the blade:mean line on the hub-firbjected on a plane perpendicular to the shaft centerline was made. circu- “lar, - Due to the fluid fiassage curvature the angle that the flow sees is not’ -the’same angle ‘as:.that seen in the plane perpendicular'to;the-shaft centerline. A-correction is made.for this deviation with the following equation: tan fi cos 7 = tanB' 7- S _ | _._. (N where - B = angle fluid sees "y = angle between the tangent to the fluid passage - "~ and the plane perpendicular to the shaft center- line , . B! = angle seen in the plane perpendicular to shaft centerline. To allow for contraction through the blades a Bi of 20 deg was used for the blade entrance angle on the hub. From Fig..D02‘7 was measured and found to be 33.17 deg. | ‘ tan 20 deg ces_33tl7 deg = tanB{ B,' = 16.94 deg From the experimental work, the blade. 52 was determlned as 26.5 deg° Nb cor~ rection is requlred for the fluid passage curvature at the discharge since for ithls 1mpeller the tangent of ‘the passage is parallel to the plane perpendlcular to the shaft centerllne. The radius of therhub'tfaee is calculated from the e uation a2 a2 ‘ ' ‘ q o | R2 - R_i K o (8) 2(R, cosB,-R; cosBiAHEE -137- Substituting in equation (8),f3 is calculated to be 2.3619 in. The locus of the centers of curvature of the blades was found from the law of c051nes to lie on a 2.600 in. diameter c1rcle. The hub blade trace is shown in Fig. D.Z2. To.align the fluid along the shroud with-the’blade, the streamline of the fluid along the shroud is assumed to follow a portion of a two to one ellipse. ‘The relative entrance angle of this streamline is equal to Bl and the relative exit angle equal to B2 The ordlnate of this partial ellipse represents the arc length of the fluid passage at the shroud in a plane parallel to the pump centerline, and the abscissa represents the circumferential .distance along the shroud blade trace in a plarne perpendicular to the shaft centerline. The arc length of the partial ellipse is the arc length of the blade trace in the plane perpendicular to'the'shaft-centerline. The fluid passage arc length is obtained from the pump layout as shown in Fig. D.2. The angles Bl and Bg are determined by W and U. The tangents of Bl and 52 are the tangents of the ellipse at the blade entrance and exit. With these values, “the portlon of the ellipse between Bl and 62 is defined as is shown in Fig. D.2. | The blade trace on the shroud is constructed from the partial ellipse .and the shroud contour in a plane'parallel'to the shaft centerline. This construc- tion is done by plottlng on-the plane perpendicular to the shaft. centerllne concentric circles (see Fig. D.4b) whose radii are the distance from the shaft centerline to a set of arbltrarlly chosen points. on the shroud contour (see Fig. D. ha). o . ‘ ~ Starting at the dlscharge end of the blade, measure a chord length (C) ob- -tained from Fig. D.4c on an average circle as shown in Fig. D. 4, ‘and connect its termination to the shaft center. The point of intersectlon of this radian and the next smaller concentric circle lies on the shroud blade trace. This process is continued as shown in Fig. D.4b until the smallesteconcentric circle is reached. The points of intersection obtained determine the blade trace as shown in Fig. D.2.. | | ‘The shape of the blade blank is determined from the blade trace and the fluid passage, Figs. D.5a and D.5b. Plane eect;ons perpendicular.to the refer- ence radius are taken through the blade as shown in fig. D.Sa, as sections A, B, C. etc. These sections are construoted as shown in_Fig.~D.5o with the distance "a" obtained from Fig. D.5a and the distance- "d" obtained from Fig.. D.5b. From the plane ‘sections shown in Flgs.-D 5a and D. 5c, sectlons -D, E, ¥, etc., parallel-to the reference perpendicular plane are constructed‘ The center“Of curvature andfifadii of these sections. is-also'the center;of curva-.. ture and radll of the blade blank The blade mean: llne and blank is laid out from these radii as shown in Fig:.D.5d. Since this 1mpeller 1s of welded con- structlon, it was desirous to make the blades and fluld passage parts of sur- faces of revolutlon. -139- “-{{ et -::'-‘ i.:(“.‘:‘b: Erdye S b, UNCL ASSIFIED R ORML-LR-BWG. 23980 Fig. D.4a Plane Paralle/ to Shaft Centerline Fig.D.4b Plane Perpendicular to Shaft Centerline 9 fe £8 e @ 7 | o 64 < e | ¢ Q 5 T . . : S 4 @ E 3 v o 3 4 Q. '1-_7:: 21 /9, A7 ;"1‘”??1 s, — RS e X 3 3 w | 1 BT Circumferential Distance Fig.0.4¢ Fluid Streamline Along Shroud -]49_ =lyl= Section Parallel to Reference Perpendicular Plane Center of Biade Curvature Distance From Reference Perpendicular A Plane F‘ig, D.5¢ Distance From Tip Along Reference Radius —a Fig. D.5a Blade Troce Referenpe Radius Reference Perpendicuiar | Plane Qo UNCL ASSIFED ORNL-LR-DWG. 23981 Shroud Mean Line Leading Edge Shaft Fig. D.5b Fluid Passage Fig. D.5d Blade Blank Appendix E BYPASS FLOW CALCULATIONS AND EXPANSION TANK HEATING Lo ok g TS ,--:;3;:53;;;%x‘a-xi’sr?‘%-.:;j%efl;, BYPASS FLOW CALCULATIONS AND EXPANSION TANK HEATING In examining the experimental data it was found that the accuracy required to closely determine the bypass flow rate directly was greater than the accuracy of the data that could be obtained. zThis is espécially true at the 1ower expansion tank 1iquid levels where the system (pressure) fluctuatlons were quite large. : An equation was derived to determine the pressure losses in the bypass circuit, that is, from the pump suction to the expension tank. This circuit was then tested separately with water to determine the validlty of the de- rived equatlon. In the water tests the shaft and spool piece were not rotat- ing so that the test was only a check of the assumed loss coefficients. It - was found that the predicted value was within 4% of the experimental value. Also a cheéck of the experimentally determined value of bypass flow at high expansion tank liguid levels agreed very well with the calculated values (see Fig; 2. 28)'"'The total loss in this circuit was comprised of the following: 1) the loss entering the pump shaft, 2) the loss entering the four radial shaft holes, '3) the loss enterlng the four holes in the spool piece, and ) the loss leaving the spool piece. The driving force or pressure differ- entlal was taken as the difference between the pump ‘suction and the eXpan51on tank pressure plus the head developed in the radlal ‘holes which were assumed to be acting as radial impellers. ' The eqpatlon used is as fbllow5° e .2 '43 = j"O..5(ll) ;g— + 5. 5(12) B2 +‘O,23(ll)';§— +l.ve ; where: , : . v, = velocity in the hole up shaft, ft/sep, ve = velocity at exit of radial spool holes, ft/sec, AP = _(PS -:Pfie) + QP'(pgad developed by radial holes);‘ff. ~145- Bl Substituting in the value of the areas o - 19u (22, where Q = bypass ‘flow 1n gpm Calculated bypass flow to the fuel expan51on tank through each pump shaft is shown in Fig. E.l. - The temperature of the fuel in the expansion tank will depend upon four factors: 1) the fuel volume in the tank, 2) the total recirculation rate through the tank, 3) the power density of the fuel in the tank, and 4) the temperature of the fuel entering. The temperature of the fuel entering the tank will be approximately that of the fuel leaving the heat exchangers and | will depend on the manner in which the reactor will operate. Flgure E.2 shofis the entfance, exit, and equatorial temperatures of the fuel in the,core.- These temperatures ere based on an operating procedure where the system is started from an isothermal temperature of 1200°F. As the power is increased, fhe cofe equator temperature is held constant until the inlet NaK temperature drops to 1070°F; at higher operating power the NaK temperature is held at 1070°F while the fuel temperature Increases. The power shown in Figs. E.2, E.3, and E.h is that which is received by the fuel and not the total nuclear power. The design power to fuel is about 55 Mw. ' ' Figure E.3 gives the expansion tank 11qu1d level as a function of the power to the fuel. 1In this figure it is assumed that the reactor is filled at 1200°F to a level of 1/2 to 3/4 in. The final level was based on the mean fuel.tempefature and is.approximately 25°F less than that of the core equator temperature at 55 Mw of power to the fuel. Thé élight rise in liquid level from O to 20 Mw is due:to a shrinkage in theireactor voiume. It has been . 3 estimatedclf} )thé.,t the volume decreases by about 70 in.” from assembly tenipera— ture to the design operating conditions. 'ifi'Fig. E.5 this was assumed to be linear fiith pOWET | | ' | The mixed mean fuel tempéiature‘in the ekfiansion tank is shown in Fige. E.bt. In determining th;s temperature the initial f£ill level was assumed - -146- =Lyl- ART Fuel Pump Bypess Flow (%) 1008 = 17.6 gpm 65 60 55 50 | L5 3300 rpm Pump .Speed 2400 rpm 700 gpm 550 gpm . FIG, E.1 CALCULATED ART FUEL PUMP BYPASS FLOW CHARACTERISTICS 3-1/2 in. 2 in.. N 1/2 in, Expansion Tank Liquid Level 1700 1600 1500 11,00 TEMPERATURE (°F) = W 8 = no o . o A 1100 1000 ORNL-LR-DWG. 29983 Fuel to / | Hes? 1t Exchangers Fuel af Core Equatd pd '/ Fuel Fy Heat Exclw/ ] - \ _— / . NaK { .o Heat Exchd inger g ~. 10 20 30 - Lo "POWER TO FUEL (Mw -148- . FIG E-2 ART REACTOR FUEL AND NAK TEMPERATURES AT A CONSTANT PUMP FLOW OF 6L5 gpm 50 60 '(in;) EXPANSION TANK LILUID LEVEL / : 3/l in. Fill Le 2 in. Fill Level Y 16~ 20 30 ) £o POWER TO FUEL (Mw) FIG E.3 THE EFFECT OF REACTOR POWER LEVEL ON THE ART EXPANSION TANK LTIQUID IEVEL -149- 60 —— Lo ORM- -OWG. 23984~ % -051- Mixed Mean Fuel Temperature in Expansion Tank (°F) 1375 1350 1325 1300 1275 1250 1225 - 1200 1175 ORMNL-L R-DWG, 23985 2100 rpm pump Speed NOO rpm & . had ~d » e s gy 2700 TP T % 3300 rpm T 60 Mwi? A 2 50 Mw /) Lo Mw (Reéctor Power) 30 Mw FIG. E.4 MIXED MEAN TEMPERATURE OF FUEL IN THE ART EXPANSION TANK AS A FUNCTION OF REACTOR POWER TO FUEL AND PUMP SPEED AT A CONSTANT FUEL FLOW OF 6i5 gpm PER s to be 5/8 in. at 1200°F and the bypass flow was taken from the calculated value shown in Fig. 2.26. The.leakage around the island through the floor of the expansion tank was calculated to be 4 gpm at 2700 rpm and is e function of the pump speed. The temperature of the fuel entering the expan- sion tank was obtained from Fig. E. i.' The heating rate in the tank is the sum of the after-heat rate (11 w/cmelh) -and . the fissioniheat release (10 /cmifl'l These heatlng rates are for 60 Mw of nuclear power or 55 Mw of power to the fuel. The healing rate was assumed to-be linear with power. The hlgh degree of turbulence in the fuel expan31on tank as revealed by high speed movies taken in this region indicate that the mixed mean temperature “will probably be the true fuel'temperature,' -151- gg’éffi z?&@@ T ~ Appendix F . 'RADIOACTIVITY IN THE ART FUEL PUMP OIL SYSTEM RADIOACTIVITY IN THE ART FUEL,PUMPIOIL SYSTEM This section treats the transfer of radiocactive gases from the expansion tank region of the ART reactor to the fuel pump lubricating oil system. A report by W. K. Stair(lS) covers tests on the back transfer of helium against argon in the fuel pump to give some indication of . the back transfer. of xenon and krypton agalnst helium. Two numerlcal values from this report were used for the calculatlons of this sectlon, i.e., an attenuatlon of lOu between the concentration of radioactive gases in the expansion tank region and the con- centratlon in the 0il catch ba31n, and a leakage rate of aas from the oil catch basin to the 011 reservoir of 0.8 1n.3/day. In determlning the leak rate of 0.8 in. 3/day, it was assumed that the 0il was completely saturated with argon at the beginning of the test run and that all leakage across the seal was col- lected in the oil reservoir. The value of 10 for the attenuation in concen- tration between the fuel expansion tank and the pump seal is & minimum. The helium concentration in the argon buffer gas used in the experiment reported by Stair was 10'”, hence the actual attenuation in\concentration betfieen the tank and the seal mayihave'been:greater. Because of the shertflhalf—lives of the fission:product gases the attenua- tion .in the amount of activity between the fuel ekpansion tank and the oil regervoir is greater than the attenuation in gas concentration. Thus the rate at which the fission proancfi gases enter:the.oil system is of significant im- portance., Method of Calculation In the ART as well as in the above mentioned test, 500 liters per day of buffer gas flows down through the clearance around the pump shaft into the ex- pan51on tank and 50 liters per day Wlll be bled directly from the oil catch basin. The attenuation of 10 represents the back transfer of helium against the 500 liters per day of argon. This back leakage was determined from analy- sis of the 50 liters per day bled from the catch basin. | ~155- The equation representing the numbér (N) of nuclei of any nuclide in the catch basin as a function of time (t) may be written dN i =S - (A, + A )N where Sb = source or leak rate into basin - ' AP = reciprocai of average dwell time in the catchA basin Ag = decay constant of radioactive nuclide. Assuming an attenuation of lOLAL and a purge rate of 3000 in.3/day (50 liters per day) the value of S“b and AP may be calculated as follows: assume X concentration in expansion tank and o 10“)+ X concentration in catch basin 10 in.3 = volume of gas in catch basin = leakage of X into catch basin T X leakage from catch basin 8, = 10 'l‘ x 3000 in.3/day 3.47 x 10 6 3/sec of X concentration < ¥ (_expansmn ta.nk) Ap T BE,L00 x 10 S, = 3.47 x lO X Volume (expansion tank) = 3.86 x 10 -8 N/sec 3000 3.47 x 1073 sec™t and N may be ex;presséd as - Sb lil-e'-(APJr Ad)t:] vos . xp*')tdz : . -156- .. The leak rate into the oil system from the'eatch basin is O.8-in.3/day or 9.26 x 10"6 in.3/sec. To convert this rate to'nuclide/sec the constant . 9.26 x 10'6 must be multiplied by the nuclide/in._3 in the catch basin. The volume of the catch basin is 10 in.3~and the leak rate is: : . ‘ 9 56 x 10 6 3/sec x Ngcatch~b251n)‘=_9456-x 10-7‘Nu:izde 10 in. Nuclide Concentrations in Expansion Tank In the calculations for the ART the Tollowing assumptlons are made: Helium flow rate down shaft = 500 standard 11ters/day Gas volume in expansion tank 90 1n.3 | ‘Entire reactor fuel volume passes through xenon re- | moval system every 190 sec where all fission gases are removed. . Sr = rate of formation* of a particular nuclide in fuel.at - 60 Mw = 186 x 1016 X the yield of that nuclide per fission - | | dt fuel = - (_AS + ?\d)N - Assuming complete removal of the,gaseous nuclide in the course of its transit through the fuel expansion tank, then reciprocal of average dwell time in the reactor’ As s T 1%0“527"103 5. = 186 x 107 16 x yield at equilibrium | - - 16 186 x 107> x yield N, . = fuel 527 %2073 4 A, Se = _seuree of fission gases into expansion tank = ;Kstuel _ . . T S S ; - _ta‘rfli . ' — - A 3t AgNpyer = (Ag + M )Neary ¥ There are approximately 3.1 g W A purge rate of helium 1nto expansion tank at tank temp. and pressure . fi, . _ . - - gas volume in tank _ = 1;96 X 10"2 sec™T at equilibrium o Astuel' Ytk T A WA After the fission gases are separated from the fuel, their decay products are then treated as gases. This assumption is conservative but the actual be- havior of these nuclides is not known. | Nuclide Concentrations in Oil System The following equations give the total number of each nuclide in the oil system feeding each pump: 0. N o l --e 1 R S (X7 + A5 ‘ | SALt T : | LT 2) n° = ¢ {L--e °) )% + AL ) 2 A”B Ao | u2+hg TR +2) (k-+XT oAt =Xt T _ +cCyfe T e Z) M - - ALt T T (3) n.° = ¢,c (1-e °) > ) )5 + [ ’\2 e | - ) T. . l+'11' L M 7\,2. NCL "Ry T TRy ) TAp Ay | TARFA) At =Agt (T .\ + C,Cp (e 7; < 2 5 F1 4 T e ALt - ALt . Ao :INl N C e '3 Ao Ry A | TApFag [ 7% T =7y (3 A (v) Il Ca + -+ ) .+. P -A’-l-t A NT . C (1-e ) ll' + [l + 3 ] 3 B 7\)_|_ (A)_fi') ) : | (7(‘)1'_+ AP) .(A3+2P), [i + A2 o+ A 3 » ] R ) (?ufi?\) (l+)) ) (A2+A ) (23+)P) (Ay+2a) T 1 Az Ap A N 1 ()4+A ) (A3+A ) (At ag) (Al+APS (e 37 A c.C DR + |1+ 2 W TR | TR Ty o) o o | Na | "Al‘ Ao A1 . 3 _ , (X o~ A3)] (A Ag) [1 TRt A) TG A Tagr ) A3 As ; AL ° Hksl. | Ao T (7\2-23)(71”) T G g - oy | Ag 32 1N | (31 A3) (g 7 )] vy -A,t -A _ T 2" _ y° A3 N | (e ) ] T2 | Ay CA TR A R g (R Ay [ TR ) A ) G [ e -2t - zt - | T €,Cy (e 1te 2 Ag e N | T T ay) Tag A Tagr i) -159- where N = fiotai number of any one nuclide_ - ( AP = .purée:constant for oil catch basin = 3.&7_x 1073 sec L )J = decay constant for nuclide_undef'cohsideration C, = leakage constant to .ca.tcl‘l basin = 3.86 x 10~ (n“:i;dej)a S, Cp = leakage conetant to oil system = 9.26 x 10 7(EE§%%99) t - time (sec) superé. 0 =-oil eystem T = expansion tank subs . 1 ; parent nuclide (fission gas) 2 = first daugptef 3 = se‘condA daughter 4L = +third daughter To calculate the dose rate near the oil reservoirs all of the radioaetive nuclides are assumed to be concentrated as a poiht souree in-the oil reservoirs. In the calculationé it is assumed that the 3/8 in. thick steel wall of the reservoir will give no attenuation for gamma, energies above 0.03 Mev, an attenu- ation of 10 in the energy range of O. 01 to O. 03 Mev and 75 for energies below 0. Ol Mev. In cases where the exact percentages of the gamma, energies are not known, the maximum possible percentages are used. The data for yields, ener- gies, and percent gammas are taken from Blomeke (16). The data for dosages are taken from Rockwell(l7) CalculatiqnmResults The results of the calculations are given in Table F.l. Columns 6 and 7 give the B and y energy release rates in the fuel expansion tank. These are equal-to 15.7.and 0.1 Kw, respectively. Column 9 tabulates the disintegration per sec of each nuclide and Column 10 converts the B energy release of these ';;607 disintegrations to Mev/sec. Columms 9 and 10 refer to the activity and é%ergy release of only one of the two oil systems. Column 11 gives the y dose at a distance of 1 ft from either’of the two fuel punmp o0il reservoirs. This ¥ dose rate is about 30 mr/hr at a distance of 1 ft ‘and 1s inversely proportional to _the square of the distance from these reserv01rs. : - The approx1mate volume of oil in each pump lubricant system is 30 gal Iflthe total B emitters in the_oil_are assumed egually dispersed throughout the system, the dose to the oil is founduto‘be about 20 rads/hr. This dose for 500 hr of operationvat-GO erafter-equilibrium is reached amounts to lOh rads, which is insignificant compared to,the.lO8 rads which the oil is capahle of ab- sorbing Without serious damage. _ - o If one distinguishes between the fission gases and their daughter products, it is found that the fission gases account for about T0 to 75. per cent of the total activity in the Oii system but only 25 per cent .of the y dose calculated in Column 11. -161- 10 8 9 Table F.1 1 6 2 L 3 CONCENTRATION AND ACTIVITY OF Xr, Xe AND DAUGHTER PRODUCTS IN EXPANSION TANK AND OIL SYSTEM FOR A HELIUM FLOW RATE OF 1000 LITERS/DAY 1l 2/ £0°0< 3 103 '33%® ON . o ' o o g0 o @ Sovm%2900000%00000700&0%3&000000&00.ulfi 0 Sl=1P® | s mo n o o ooHA. m tcoA T0°0 03 £0°0= 1 203 0T="39% " 43 1 9% esoqg 4L m|OHH, 295 /Al ; : = ” wno M ~ wn g~ ueysds 110 ut OTM%BB&@fi@&?i% = 010M0%M9M91%2M0NMM . - - - - . - - . o 9PIIOoTW WOJIJ TO A SdoncinieAa~c e o~ R 88884538 4dd4dR o3 (d)a TeI01 ) . L S S R B S e oD B R oo oo oY dodd R OG0 0 00O CoO0O0CO oBo © . woqsLs 1111mLW111111111111111111111111 823 TIoup oo qod | KM X KK H KN KA KRN KRN R HANKKRRKHRKRNRNK | KKK Mmoo ~ wmnao OO M~ A0 N N ONND -~ SUCT4eI3SUTSTQ flu)abuusmlmfl.flfllo o@D 9”J9935~flm5mm3h5mmafl “M%O e« ® @& ®» & @ 4 8 8 o o @ " % & e & & & & @ ® € & & s @ . .. a doNMmAD. Famu AN AmAAIAA~ASA3AASINAA AT 4 - O\ =0 M, Oy O = ] Oy O =0 000@000%@00000@00@@@@@0000000@& &3S 3:8umfls8m1111111111111111111111111111111.111 Toqge upqehg KM M N MM MM KKK N K NK A REKREKRMRRE RN N R XR N o O o P OVHM~WNMNDOOANNO wn N =0 - 170 ut epTFTouN %@flOB%%%hM%JBM%Z16h955138M8h92_OQM - - -* * - - - - . - . - . - - L] - - . . . - e . . - - - A AR R I AR A O NN T TANOO =N HD A HOD OT X g € Y = 3 8% ~ 9, - - & s/ [T T s noo v 0000000 (o | Snoo000oonoo 1 S| (p)ajeocgo o o H 7 — 160 & 1 g 4 ~ s m A o M.HIOHK wno o © o W o~ co m n oo | R GH AR e de 185 10 10H0RRRSS 18080 1 - - - * . - "~ (g)a CROQNRLS AT AR 1001 | & Al tinkiA ot £ep/6ITT o0 D@~ OOV = =0 =[O M =DM D Q@D b= b= == b= PO\ = O 4 2 e —~ 11191@ D8R A AAm A — *P35 000T J° 6D D D0 0 000D L oo 0o b e b I L d L LD DT DLS 0qey MOTg MOT{o [H A A AAAAAAAAAAAAAS AR AR SRR S A A S A A J0F cunToy Sep | K MK M KM MR H MK KH R KN KR H MR KK HRKK KN x.M MMM KK Nt~ O~ onoD (2o %0 g N VN ND -~ MU\ OO O~ juel uotsued | MR SN S N R R AT SR RNRBRRR AT R REZF v . & @ e ® o 8 e & & & 3 S g & B & B B 4 ¢ » B S 3 & s & S 8 “XT c°UT 06 UT | Sl Mt unem i . Moo — a0 i 0 0 0 ot o N0 o o 8OpPTIONN Te30] ..D...Ué.m.z ~ 4 R 3% %0 A8H (4)¥ Teq0l fllSfiOOOOAOOOOOBOOSOJNSO00000@00&023 e . - » - . * - . - . cooQ00O o °© o oo ™~ 000w _ a9 ~or~awn om o S Y 8m & o o BY aBl (d)F ‘Bay o8RG ool NnarorodocRmormAGmoo e ® ® & @ a2 * o & ° 8 ® ¢ & 5 @ & . . e o 4 s e & 4 e @ o Ty 2L N LA T I DL O LI TN 00O oo oo Q CCO0O08 OO0 99000 O Y AAAAAAAdAAAAAAAAAAAAAAAAAAAAAAAAAAA ufiw”cou MK KKM M MO K KM R RS MR KM KN M KKK KKKK KN o Qg m O o OISO o N O HY O MOy Y ~ a m_u.w..h961w ~ . 191123.&5&197@36@2@%%2& SPTIoMN SECRET 31.6 3753 87.68 (0.1 Kw) 9795 (15,7 Kw) TOTAL ~ Appendix G Xe”” POISONING oy Xe™ > POISONING This final section iS'COncerned with the xenon poisoning in the ART. In the calculations which follow the removal of xel35 by neutron capture is neglected. As the neutron capture rate of removal is only about 5% of the total removal rate, the calculations will not be greatiy affected. Also; the results for poisoning will be comservative in that it would be smaller were the burnup rate considered. The derivation for the equilibrium xenon concentration in the fuel can best be followed by referring to Fig. G.l. The total fuel flow through the expan51on tank including the leakage around the island is about 23 gpm or 1450 cmB/sec.- The helium flow rate is treated as a parameter and is varied fram 1000 to 5000 liters/day at standard temperature and pressure. The solubility(la) of xenon in fuel is shown in Fig. G.2. - If the value of the sparging efficiency is. defined as N | ‘ (gg gol)ln (gm mol) cm 7 = 2 (l) (*‘3"m moly o - (solubili‘ty), P, where P is the absolute pmessure of the xenon in the expansion tank. ' Then T—ELB’ -ill-_—s—o-"- 14-.3 x 10 Px . \ - Z = 6.23x10°p +L . (2) For equilibrium conditions, ¥ must be equal to the rate of formation of xenon minus its decay before reaching the sparger and is the rate at which xenon enters the sParger and thns the expan51on tank This term is the same as ‘the source to the expansion tank as defined in Appendix F, and is shown in the last column of Table G. lf The value of P&h can be determined -165- L from the total xenon nuclides in-the expansion tank. The value for the total nuclides p;esént are taken for tfie case where the free gas volume.in the ex- pansion tank is 90 in.3,'and the method of calculation is again the. same.as Appefidix Fo _ | The nuclides of each xXenon which is presént in any appreciable amount is shown in Table G.l. - A total pressuré of 2 atm and a temperature 1300°F were used to determine the partial pressure of xenon.-;The'paftial pressures at helium Plow rates of from 1000 to 5000 standard liters per dey are given. in Table G.l. - Substituting the values in Table G.l in Eq. (2) gives the xenon concentration in the fuel. Figure G.3 gives the total xenon concen- tration in the fuel for helium flow retes of 1000 and 5000 standard liters per day. Also shown in Fig. G.3 is. the Xe 155 concentration in the fuel. This was obtained by teking 14% of the total xenon (the percent shown in Table G.1 is 13.63). o | The Xel35 poisoning(lg) in the reactor is the number of thermal neutrons absorbed by the Xe™®? to those absorbed by the fuel or LR ) For a fuel with 5 mole % of UFL the U235 concentratlon is about . _ 8.2 x 1020 atoms/cm3. The. ratio of the xenon-to-fuel cross section G~ fifF Por a Maxwellien distribution at ART thermal is h 9 X lO3 (20) As approx1- mately 22% of the xenon cross section is scattering, the value of 0* Ar is multiplied by 0.78. Therefore ' P o= 466 x 1000 e, | (4) The results of Eq. (h) are shown in Fig; G.4 (upper curve). This equation is only for the poisonlng effect on the thermal neutrons.- In the ABT about 4O% of the neutrons reach thermal energy. The lower curve of Fig. G.4 shows the poisoning based on the total neutrop den51ty of the reactor and is assumed to be 40% of that for the thermal neutrons. This is a gross simpli- fication of the actual case, but in the present reactor in which the purge -166- rate is high compared to the burn_up rate the ass_u_mption shofild be fairly accurate. | | - | | ' | The effect of the p01soning may be more easily understood by comparison to the. temperature coefficient, which for the ART is expected to be about 2x 10 5/ F. Thus, & poisoning of 1 x 10 =5 1s equivalent to a temperature change of 50 F. ‘ ' ’ -167- Table G.1l XENON CONCENTRATION IN EXPANSION TANK GAS VOLUME Nuclide in Expansion. Tank x-lo‘;7 for Helium Flow Rates of Decay Constant Source to Tank Nuclide | ‘Sec—.l) n ti(x)'(sx}day litzgcs)?d.ay .ngg?aay 1i tle*gg‘/)day 1 t_zg(s)?day Nu'c.:jlfid.e/sec‘ % 10-16 xel?l 0 28.2 1.1 9.40 7.05 5.6 5.5%" 132 o 43.8 21.9 4.6 11.0 - - 8.76 - 8.58 133" 3.49 x 10°° 1'.5'2' , 0.76 . 051 0.38 o.3§ 0.30 - ?155 : 1.52 x 10'6 62_;_8 .. _31.l+ | - 20.9 157.7 | 13.0 12.3 - @f 134 0 ™2 27.1° 2h.7 18.6 . 11».8' o 1k.5 135% 7.#0 x 107 | 13.7 6.95 .h.67 3.51 | 2.81 2.78 135 2.11 x 1077 62.5 3.0 20.7 15.5 124 12,2 136 0 | 60.9 56.5 20.3 15.2 S 12.2 - -_ | 1.9 137 2.9Tx 1000 3.7 17.5 11,9 ; 9.05 _‘7.29' 1% 138 6.79 x 10‘1,’ 49.9 . 25.3 17.0 12.8 | 10.2 - -10.1‘_-' 139 1.69 x 10°2 7.63 4.97 3.68 2.93 2.43 o 2.78 1l+o‘ h.33 x 1072 1.95 1.48 . 1.20 | 1.01 0.87 . : 1.23 Totals 1—5_9.80 222,56 149.56 - 112.73 90.70 T 89.59 Percent Xelo? S 1k 13.9 13.85 13.78 13.68 13.63 P 0.00kOk . 0.00205 0.00137 0.00104 0.000834 | . ORNL-LR-DWG. 23986 X liters/day He | | an X IiterS/day of He Wam mOIS/sec of Xe SPARGER ~ AND f _emd. EXPANSION 1450 /sec of fuel| TA NK (Z-Y)gm Mol/eae of X 3 1450 “M7sec of fuel Z gm m’ol‘/sec of Xe FIG. G.I- SCHEMATIC FLOW THROUGH SPARGER AND - EXPANSION TANK 169 ) gm nol cm3 atm SOLUBTLITY X 108 ( ORNL-.LR-DWG. 23987 1100 1300 FIG G.2 - 1500 TEMPERATIRE (OF) XENON SOLUBILITY IN FUEL-30 -170- 1700 . - . ORNL-LR-DWG. 23988 100 FIG G.3 XENON CONCENTRATION AT EQUILIBRIUM IN FUEL-30 WITH A POWER GENERATION OF 60 MEGAWATTS & Q ™~ 0 £ 3 (11} b O —~ 4 _ 10 1 o & § _ 000 liters/day 5 7 Helium Bleed 5 000 liters/day : Helium Bleed % ' < ' 000 liters/day 1 EFFICIENCY * % Helium Bleed -171- 100 L e 2 = o N . - o Q‘ . =3 10 — . b AT a L'H] b 1 Neutrons FIG G.L Xel3® POISONING IN FUEL-30 AT 60 MSGAWATTS POWER AND A HELIUM BLEED OF 1000 liters,day tal Neutrons 1 10 EFFICIENCY (7 ) % .. 100 -172- | & “ . VACKNOWLEDGEMENTS A pumber of people and groups played 1mportant roles in the de31gn -and development of the ART fuel pump and xerion removal system. The authors' are 1ndebted to the follow1ng people: ' ' ' ' ' W. Lowen contributed to the design of the attitude- stable xenon removal system. E. T. O'Rourke was responsible for the assembly and functlon-. ing of all eqplpment 1nvolved in the plastic models of the xenon removal system. ¢. D. Whitman and J.J.W. Simon followed the assembly and construction and subsequently operated the iron fuel pump and the combined Inconel fuel pump and Xenon removal system test loops, respectively. , We G. Cobb and L. Wilson were responsible for the layout of the fuel pump and xenon removal system. ' The authors also wish to thank the members of. the Puwer Plant Engi- - neering and Experimental Englneering Groups of the Aircraft Reactor Engl- neering Division for their contributlons and assistance in the’ de51gn and development of the ART fuel pump and xXenon removal system 1734 1. 2. 10. 11. 12. 13, 1h. . 15. 16. 17. 18. 19. . ORNL CF-54-12-52, REFERENCES | J. L. Meem, The Xenon Problem in the ART, ORNL CF-54-5-1 (May 3, 195L4). S. I. Cohen et al., A Physical PTOPerty Summery for ANP Fluoride Mixtures, ORNIrElBO (Aug. 23, 1956) AREDbANP De51gn MEeting No. 55-h ORNL CF-55 1-209 (Jan. 26, 1955) ANP Progect Drawing E50-J4U7-4, Rev. 8 (Dec. 9, 1954). A. P. Fraas,ACFRE-Design and Development Handbook, Project No. 3. H. Rouse, Fundamental Aspects of CaV1tat10n, Proceedings of the National Conference ~ Industrlal Hydraulics, p. 31, 30-7 (19h7) D. Thoma, Bericht zur Weltkraftkonferenz, London, 192h Ze Ver. ‘deut. Ing. 79, »- 329 (1935). . e ’ D. Thama, Verbalten einexr Krelselpumpe beim Betrieb im Hohlzog Bereich, Z. Ver. deut. Ing. 81, p. 972 (1937). o ' W. Lowen and G. Samuels, ART Xenon Removal System, unpubllshed (FEb. Ts 1955). F. H. Garner and D. Hammerton, Circulation In51de Gas Bubbles, Chem. Eng. Sci. 3, No.. 1 (Feb. 9, 195&) _ N Flow of Fluids Through Valves, Fittings and Pipe, Crane Company Techni- - cal Paper No. 409 (May, 1942). J« B. Baker and J. W. Michel, Pressure Changes With‘thE‘FlOW'Of‘WHter Through Tees, K 998 (Mar. 16, 1953). L. A. Mann, personal communication. A:. M. Perry, personal communication. W. K. Stair, Back Transfer of Gas in ART Pumgs, Experlmental Report Task No. Th18, Exp. Ho. 1 (Aug. T, 1956). _ Jd» O Blomeke, Nuclear ITo@erties of U235 Fission Products, T. Rockwell III, Reactor"Shieldihg Design Manual, Fig. 2.1, p, 19. ' ORNL Special Progress Report - Program 4400, ORNL-2249, p. 11 (Dec., 1956). S. Glasstone end,M{ C. Edlund, The Elements of Nuclear Reactor Theory, Ps» 335, D. Van Nostrand Company, New York, 1952, A+ M. Perry, personal communicatiohgu