ORNL-4541.txt

JUN 17 9N
B BUER et B

. Mffi%Tfifi T ORNL-4541§d‘¥
n

UC-80 — Reactor Technology

CONCEPTUAL DESIGN STUDY OF A
SINGLE-FLUID MOLTEN-SALT BREEDER REACTOR

Molten-Salt Reactor Program Staff

Compiled and edited by
Roy C. Robertson

“J

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the

U.S. ATOMIC ENERGY COMMISSION

ORNL-4541
UC-80 — Reactor Technology

Contract No, W-7405-eng-26

MOLTEN-SALT REACTOR PROGRAM

CONCEPTUAL DESIGN STUDY OF A

SINGLE-FLUID MOLTEN-SALT BREEDER REACTOR

Molten-Salt Reactor Program Staff

Compiled and edited by
Roy C. Robertson

With principal report contributions by:

John L. Anderson P. N. Haubenreich Roy C. Robertson

H. F. Bauman E. C. Hise M. W. Rosenthal

C. E. Bettis P. R. Kasten Dunlap Scott

E. S. Bettis R. J. Ked| W. H, Sides

R. B. Briggs H. E. McCoy A. N. Smith

W. L. Carter H. A. McLain 0. L. Smith

C. W. Collins L. E. McNeese J. R. Tallackson

W. P. Eatherly J. R. McWherter Roy E, Thoma

S. J. Ditto R. L. Moore H. L. Watts

W. K. Furlong A, M, Perry L. V. Wilson
JUNE 1971

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION
= = - ; : ' }fiytvf:a
; ~ | ORNL-DWG 70-14134

LR
- .,ql
1 e ) A= Nl =T
i 4

1000 Mw(e) MSBR STATION.

- L
-

Contents

SUMMARY ' si i 60 £ ocbrline Wb apsiaast 2 dnsinas drs it e Aaow wie b asian s (% a0 e e g TV o Khniae 1T 08 ix
ACANOWLEDGMEISTS. § 8 s. 42005 5% Me b 0 5. 35m Ma a1d 5\ s dtmta.ine. ausoials ads i sise s s arn fesuniy e nin xvi
I INTRODDECTION - vp s o' vas 3adad s i o s st o0x sk sndimtaahs, sha's a s sh whaie i ghe shn v i ath joepip 1
2. OVERALL SYSTEMS DESCRIPTIONS ANDFEATURES . ... ... it iriiniinncinnnrnans 3
21 ROACTOL PIIniary SYBIAIM- 5 o160 92w i@ 6t € Aol #5 Al 5056 i 674 § B8 RES 55 v 5 & 5% Gt o 3 3

2.2 ' ‘SecondaryeSalt COoulating Byatont .. « & « 5 48 Sei s b tia et Vvt e e SAG e r PI 50 2T 2 Ty & 5

2.3 Steam-Power System for the Turbine-GeneratorPlant . . ......... ... iiennrnn. 5

2k RrarDalt P BUSIBIN | i s v sl s % W d dik i & 5o K el e @ Ry it A vehasa i wh e 6

e LSRN BB e o Tovn Wi Bt 1 W0 B e TR 88 B AL S S e will N WP By e 4 4 b s 7

2.6 . Futl-Salt-Procdssiig SYSIem | 26062535 5.0 5 60 2 50s 555 5 5ga Poalh 90 €5 K540 355 S ITE-A% B AW €66 4 7

2.7 Auxiliary and Other SUpport SYSIEmIS & ¢ v v b s v o 0 sis s 0 6a sia s i 0486 & 0K ols e aq T Ewies 9

3. "REACTOR PRIMARY 'SYSTEM 46+ Bad 06 G030 ol 0l 50 by aipidt Siainied aid nh g oy o 4o o5 Sk 10
3.1 General Desolipon i v d s vivvd s ad aw e bt d aw o Vi wae A Ve T T Y S d AT W A 10
Sl Design ObISCHWR, 5 i s 50 gos 25815 0 5es 0 8 i PR AT 0P 2006 4o TS a4 L 298 W E 10

3.1.2 General Description and Design Considerations . . . .. ... ..ot cvivinres 10

3.2 - SPeOial MAtalElk - - i1 s h 53 Dok ach Gorinis no si il WE A0 Bl 45 & @E A5 9% RS RYE T Bk B 20
321 POl BRI o 1.5 5 5 aorBe 5 A B A e R o SR SR T e 5 20

3.:2.2 'Co0lafit SAL, « s S visid 55 Tad 2l B plnd aii i S poal BG4 5L A S SRS MEE T AT 4T B e a2 22

328 - Rector GEaphlle ! o ¢ b v.us oxrmr e 1 siamm s e opDve vs 4ig 240 38 SaEiave & o o ey p ¥ 22

et (Hatelloy N .o oW & AR Ak 8 s s e dTnbd vod #5238 5 piie o aas o pse e (s &35 825 38 26

3.3 +Nuclear Characterlstics i ¢ s sni sasndivaiidmbin s Padwa sardbas X dad b aow e »e as 6 28
3.3.1 Selection of MSBRCOT IRSIZIN 2. i G siwnis #2 o 5 ssph o 5 o5 gt 85 % 848N 5 0 & ke 28

3.3.2' “Optimizatioh o \COre DesIBN ....c <5« 2 5 F8.0% % sv.avn 5 0cs ¥ niidmn e min & 048 Fiv' Siminss 29

3.3.3 Effect of Changes in the Fuel-Cycle and Core Design Parameters . ................ 31

3.3  Regctivity Coefficiensts ... « & i afnc s amen s o s valed e 5 sl @ 1 MK &a w36 Vel ataae i 35

3.3.5 Gammaand Neutron Heatingin the MSBR < . % oo oo mrccnvid s abaie oot aba 36

336 FissionProduct Heating in the MSBR. .. ..ci6 i dssvabrsnrsnvspunreehidbohagds 37

3.3V “THttum Froduetionand TRSDUHANE o 5% 5ow a bl iain S % resid i e s Ko s Gingdieis s 38

3.4 Thermal and Hydraulic Design of Coreand Reflector . . . . ... .. ¢ i ioinenann 44
BRI, BEORE: o 0% 2w 5, 40T 1 DA 3 A SR B PR R b T O RYR ACBY V WOR 7 ST S Pl sy e il 44

342 Rodigl RTIEOLOL : c <555 08 edis nig® 4 T alavie B 55 d i@ o kit Qs SeoPs .09 00 o 49

343 Accial RETISCIOTE < o4 5 b 5 i oivie A FRES A Vam ta a'n 50 a D ke Gion o s e Tumsd R s S und s 50

35 Reaotor Vel Dosipll. 2555533005 Pe barmt n P vid B2 s 43 nsmsotatis wB o2 2ud 24 a1
35.1. Réaitor VEsse]l DeSCAPHOM « et 35166 55 4 6675 aehomd mie-s1srion pasesra ¢ S66550sbe500 4598 51

3.5.2' Reactor Vessel Temperatires caaieeas a-ijae o se ba i s $5 @i sm €6 &o 5060w &6 e 52

353" Nonotof Wassel SUORIE: .. w5 rs B 1 15 570 o oih! dia vty e e G5 o0l FUd ik & 29 5 & &8 52

iii
iv

3.6 Peitoary Systanm SSEPIG 15 5.8 S sl ie s it wbs mrefodlinars 595 4 o afr 3cp i o €5 A e dh ks 53
37 Prary Heat BXCHAnEOIT v o vuieivina 0550/ 0,0 4% U9 8 Po sie o'k d v B o ol 5 5% & 015 4438 7 54
SE { Dealg ReHISMBIEE . cholilicis ins o it ui s i o mirama sty oy (o v sy Ao e G § O 54

IR TR SICITEED I 5rv- 3 v 9 e I s . et e i AN 91, 0 o AN AT e 1) 7, 20 54

2:1:3 , Denan Caleulations: o .:v 6% 03 3 vE5 s i mPsItdd s 0GR Ta BOCRG I b TGRS 57

3:74 ‘Relisbility of Designr Calealations' . a .« +w5 ma & w hsrries o v 5edw be s ia Lo wd oo 58

38 Salt Clrchlation Pumps «u s« v b i d o hsdnd nora o mSEua/a 2o 0 el ok e &85 i 34 W EA0E ENaDA 58
38,1 Euel-BaltFPUNIDE 2 020 r Pt aale snsd i oG8 Sl 0 R DE G4 R oTTus B rid o s oo 58

382 'Coolatt-Sai CIrCBoN PURDE 45 § : 5% §+ 2528 s st 499005 2 5 av st s ot neh 61

35:3 + St Franster PUMD + ¢ » ot tturn, b Bwi i piu 1850, il s € 0003 3 0 T uBross 4t oG d s B a 61

29" Bubhle GeidPaton and T DOPRIAIO . » 4 5 & £3-54 ¢ AR S A A & 08 Fw shag oo a e @ nsiiie wreraa s ¥ @ 61
591 EoauCHON:. i v o o eh e I e Wade i e R Wt Ga B8 ca TS ST g e e I B B 61

3.9.2 DL CENIEIEE 4w i o7afiate .00 8 p oo™ Eip malwy 50 G0 D AT A P m s DY, v E AL AN 61

F93  BUDDIE SCIRTEIOT . v6s 3 rouhai’s §.075 2.8 5 Wons” snm S1a nis GRS et 5 £ S 00 @ £ 63

3.9.4 Bubble Removal and Addition SyStem  « i v ve e e ss o v evesscos oaiass nsiessn ona 63

4. COOLANT-SALT CIRCULATING SYSTEME .. 5 ;o 85 5628 oo dan ot o b a4 b ot sagdod snkudtas 65
BT TGEHOIAD e 350 5 2 090 B0k e T5 o4 V55 08 BRSO bt e Oih YL EE B E aio I b s wiod sl 65
4.2 et GENAIMIONE . o 5 i-iin iiv Baigh 5.8 s Bre ddvalalh O 8 S AR @ CRAY b dalih asw OsbiEwa 65
2 IGERBTR] ' it ol ale o tin e 2L WS N R My B ik e B R bR B T v e % 65

. TRCHBUON o ot i v B 53 WAL, S8 BRI o R ok W I el 5 e wrg #oF 5 66

4:2.8  Design CAeUIaON . o5 o 5.0 4 sob g0d Son L ys A8 5557000, 56 2 B IE D o #o ShE CR0cOdN o5 0 68

4.24 Relability of Dezign CalcHIBlONE v v s « 0 et aodnn v e v ad 1o gos sad pk ol ad s v 68

43, " Soanl REheatery & ..<v: iu % 15 RPE .« U avvh v AL 2B TR 40 WE A8 RV 0L S T Bt 16 558 s 69
AT L GETBRAE % T B S BN S PR WE ¥R D SR R A VSR AR AT ACSEAFA AN IS RO 69

G52  BIERCHDUOM .« .- 20ks 51 ivsacer Do o 500 Pra s FRA FaB% 5 v b eanhs & u & Mo i g & pw & 69

&3.2 . Doiltn Caleiiation® v & <o 5.0 vad e Kb agind'd ool b e S e s e N Wiy s S § 71

434 Religbility of Degign Colculations’ st:. . s % 5% s wie e v o, a5 sy @ mmhody 4 A A Pdre &8 o0 o s 71

A4 CoalaDi-Sall SYEteIti PIPIN = o 5 s o5 a8 5 P08 v igssnss sub o agsen sciib il 354 o1 b #0680 3% 57 56 71
4.2 - SecOmIary-BVs il ROPIOTIURKS : o2 < - tiaon By i § 1 S o A SRS da s ar i B DS G el 71
LG ~CODIMIDSEIEINAMTBVEIONE o i3 T s sid e P e afs i wie it AR A e 1 s i A o4¢ 72
o BEBAMPOWERSYSTEM. o005 boive soim s doie b g b i luily £ 9 Badiir & oIe B e i s arengis & 3 5§ 74
Dk ERBERL % 5 2tk S i Dl v hd Bl e T RS AT W Bd WA At A D . B £iris O B8 S lE Al 74
5.2 Description of the Reference Design MSBR Steam-Power System . . - ......., ...c...0.u.. 75
' . MSBR Plant Thrminl BINCISRCE . » 5.5 0008 et v sl 5,0 60 v aig logo i rsbols b B ov0as B0 006 o2 86 ik o293l 77
5.4 Selection of Steam Conditions for the MSBR Steam-Power Cycle . . . ... ................. 77
55 'Usé of Rehéatin the MSBR SIeantCyeIe. & v« o s ol s 00 ol via i 6r g b ehipe & ibne aiaie aia g o eds £ided 79
5.6 Effect of Feedwater Temperature on the MSBR Steam-PowerCycle . .................... 80
5.7 Pressure-Booster Pumps for Mixing Feedwater-Heating System . . ... ... ..., iiuron. 81
5.8 Mixing Chamber for Feedwater HOating ... ...« o v ciaevias cumaes soneis sorns s o siv o908 81
3.9 Superheat Control'by AttempParation « .. ..« swaw s advoi s o d s edis a Gk adhas a awas e s 81
S8 Robeut Steamn PYhBaRtols. 2. 5o Suils Fa el % d @ B st bb smieie e Too s R 0o 509 Sk vl e BE. £ 81
2101 - OondralibesctipoN o < & & e . 0R e F5-8 0 5% @ B o5 Nt s R25 58 a3 4 TVEEFE 00195 81

5102 - Detiid ConfluatAlionS. oo wesls i siingides s fpus fu aimngn O350 RS nis ook #1 81
6. FURL-SALTDRAIN SYSTEM . s s i ooaidals smisls & i o 9o olnyife 9% o & s iy oy 3 o8 4o 84
673 Gentral Doslgn CoRBIAOTatOnS « o « o.v ssvomnrsdin 3.5 sMGE LFYA € SEED 3 2EE0h SEh3E 575 S D5 3655 84

6.2 Fuel-Salt Drain Line . .. ... .ccc it vrimueecennoceesasansevoersansasassssssesass 85

6.3 Primary Drain Tank with Salt-Cooled Heat-Disposal System . ......................... 85
ISR s T A AR -y T N e (PSR- PR G A S S G NIy 85

632 et Sonrchs I Dy Tank'. 5.5 55 %5 5wea e 2 o hare 8 sl 0d Sesndhag. SEivs o 90

6:3.3'  Heat Transfer in Drain Tank'Walls: < . v s v o i svets 5% 563 60a 3% 95 omie o s 75 ols 544 9 90

054" " EeNReIGSRD OIS | v icr wiu e 3 B s 5 ol d'n w% ARANE Pt e g sl s & ¢ Sront o B 91

64 Fuel-Salt Driin Tank With NaK Cooling . .. ... iivcisesssionsteibosessis vaaiasss 94
QAL INEONIICHION ¢ 3. a6 2t s alannats Baramats g ai o G an ike hob S04 50 28 1 R aras BiYit 94

T S Y [T O O TN e P O e o PR T SR S N L A Y X T 95

6:8 LSl SIOTIgR ERIR. S8 v ii R os Fo RS R T F e e Bt d bm s waeSuepet B S P aiknd abe 97

& REACGTOR OFE-GAS SYSTEM . i i5: oi606 00 iociua 0 51 iake 4 5. w1505 viiho (poosn 0675 60 A19 7008 5@’ Koot 16 98
ol CODEEAN S & 75 5% 5 G & a oA s DB SR b B e b B v e T i S e e P Al i 98

7.2 Basic Assumptions and Design Criteria for Steady-State Operation . . .................... 99

7.3 Summary Description of Off-Gas System . . . .+ v e cveevpvvsevsronvonssssssonssnses 100

7.4 The 475 Xepon HOldup SYSEID « & & 4.0 535w § daoniaic d 000 a0s Diika § Wie b st % B 15otd £9 517 102

Fud - Eong-Peladt ChartOal BRd . oi ¢ ovisi bRstmud $oanwe o as 2o m Gbhe Al Dol s 2 et 104
$.6 The GagCleanip Systen 15 o e b il T 2 I B R L A T e R A S e Sa RS SsTatae 106

Gt ORGSO & 4008 QiR FSATTET RaRERIZ T4 11 §50.5 48 0k 300 % b Bodaintiauple. Sves 9 109

273 T B T T T T Tt P Ty e 109

8. FUEL-SALT PROCESSING SYSTEM . i< 53 a0a 2 aals s0irdm 6 0.5 685488 36 8 Coea 4o is oiainss 110
Bl GEHOERL i 55 a0 St 5 I . e B A B 0 A VA e W AN R e R 110

B3 {Protaciinium ISGIAtON' ; & 5iid v 1ed & LA bp AN SE 5 U QM aE e RS R TR 5GP &2 111

8.3 'Rare-Barth Remioval .« ..vc.s ofs wos 60 534 83 5r 8% <ra ' sor aedei fad & 5.5 5L AT 46 48 b Fite Sadan 112

R4 Tifgraed PIant FIOWSRREE & ;. < fintaliis & wl svensiin S ausre o6l s & vis B o ,a00 alodie o, ale = wdut & o/ 5% 112

BS ' Sali-Bismutl COntattond « «. o2 fa'irils 5 & aa's vdidd arahic w48 @43 ki G Ve §iFd s i 55ROk 114

5.6 . PRIOTIDBOIE A 0. v 02 0t s 55 S AL w A TE A o, S i 5 W, S e iy i TR B 9 AR s 208 114

87 el ReconStutlon. . 58 v wad D8 o rtisvi L ns S 6§ DB TH LSS, 5 0 ) DA o 114

8.8 SaltCleanup :1.:vsceatinirmIfavispmidossaadPaisas PEavalBatessdd s sd@ensma 114

8.9 PURDEL .o 67 5 .Fvs Rebivn A ABTDE e § 50T s g waE S sivia s d e vioesernshie s e o sl K yS 20 115
B0 MEROAINE . 55 ; = o4 3% bdrsd e svb B s PRI EE st esis ol Fufos S SLstaopte S ki d saachss o od o 05 115

0 LIGUIDWASTE DISPFOSAL BYSTEM oy 5 ands 75 siidie s5v o wolp 5% ooty e o p 908 e d P8 bris sit ™ ss 116
10. PLANT OPERATION, CONTROL, AND INSTRUMENTATION .. ..... ..ot iiiennnaiinnns 117
3OL! ERNBIBL: < ¢ 5 T o R 5T 2000 it i Die Call s Bk o0 M bimy e Sl TS SR AP R Y44 AT 117
102 " MISBIC RoacRVity COIEAL 7. 20/h 0 seid & 2.5 8o s v o aod il wm Bl 00 n e 0 ok olphie & scdad 0.8 117

10.3:  Reactivity Control for Emergency SHULAOWR . .i 4 i sogivo aisasrnaa 65 oo as v oo s oo oo 119
vi

10:4° FPlant Prolective SYSEM 1.v.0 58 65 iy rnLirin avy S dvhedell o 3 va v 78 &5 v s P syt ca 3 119
04,0 “GEDBEAD 450t et v W 48 56 R fba s D BB WER @ B2 B aE Fo 50 Basim Sll3us 0 0kt s 119

102"  Reactivity RoGUEHION = ok ah 3 chili it Gk a5 @R ah S W5 Shon: e 0 R €5 M8 5508 2 119

1043 LoRdRodneHON - ; : 58 s ¥ AT dns i ala ol 3% 2 b i R g b 4 s e MES 4l 40 § 119

JOM4.4 " FURLIRAIN 57 5.5 660 0 TR wASA T man 5 AR 30 Voo v, 0 Wy S0 i v (h s A 4 119

10.5 Availability of Instrumentation and Controls for the Reference Design MSBR . . ............ 120
106, Allowable Rites o Load ChanBes: ¢, .« 25 8o i a2 0k b os TEwdi s a0 st ank e anlaaie diapaas 120
10;7. Contiol of Full and Paral Load OPBIAtON + 52 9= 3@ b vt vemayens s s netavsboesss e 121
10.8 ' Copteo! fhrEast SERAOWN 5.5 satd e 5 % citeterns’ ip oivisimet ifod 250555 dhe st gtote Bialin tane1n 9.9 e 122
10.9 Startup, Standby, and Shutdown Procedures . ... «...cuiiiiaetensosesam novesnsaianre 123
1005L) ASDTELEL 5 w0 wwie e 9% At 50 n¥ dlige 58 56 Bam ol i o5 v B 3000 B v A Brmie 45 sl 123

092 | AU PTOBEUNEE . v vcv n s o rvmin & r mis i v s o - ey SV W WA oy 8w 124

10.9:3°  NormnalSORBAOIWN 513 5 alani #75 ack bt 5 R 3 s v a9 Fore i e e R 2 i g 125

115 AMKILTARY SYSTEME. o 2t 0 S0t urte S o2 S 07 018 A58 308 8 SETT ISR B A0 JVNAE L RS 126
LY. AT EIRtItC POWOD o6 via i i o o8 A abin ol e s §8 wascdele Mid b o8 A3 S SRR 3557 2 126
112 Coll Eleotric Hokting SYBIBIER. - 5-: 3 s s bra aass s N alsr s 58 m ik d 4o 8 ora 55 ook s 91 ige i xes 126
11.3 Radioactive Material Disposal SYSteml ..c « i vus sc sia hpseionsinmgie sa an aad oo onees aoaa 127
12, MAINTENANCE AND REPAIR SYSTEMS o+ ¢ iii:6 cn i saunie s sdra o s 4518 65070 8,0 9. o o/355 SO0 129
BT S R T v Tt e o T Wl A S A W W TN A R M R T 7B i W i i e 129
12.2. 'Semidirect Maintenance PYoCedUres. . « & : o1 g 5l syeic s s 3.8 o B sF Fad bl S0 afed e 4B 38 89 s 130
12.3 »Reifibte MaltenRaics ProCeauies. o . ; sn.v.o8 5300 is s v s s b S LNE 25 RURS IS s s a Botis 130
12.4 Graphite Disposal and Alternate Reactor Vessel Head Reclamation ..................... 133
125 1DeconiamiiBON: ;%0 i wessiei s 088 a0e s a6 5 i B ins s Syl i G5 Rk e e g i (. o a0 134
13 BULLDINGS - ANDCONTAINMENT ). : o5 e 5a® o b Ko and ok 1o osi e ambe ¥ o dndvhal dnysd 135
13UL" ALBEIEE o5 5 75 S8 3 s ATa R h ATAE & T A B o T Tl £ G e Ghararie i 35 e g PRGOS 135
132 Repotor Bullding . ;o a2 Sesa s i bfais.a sl Sumiem 8 s 0k i as a8, 308 0 Wa alerSuis oa kv 135
135 [ RSO DOIL R < 30554 Mo G Tl wl ot Mo fha® ¥ S B n sl I 2ot M TadiA 3¢ s 08 138
134 Paniaty-Bram TaKCEIL ..o sevividtie o 5 atetels Svi e o ats 3 vivonss €4 FE Feotd €5 26 M §oinie 144
13.5 “Proedic-Valve Ol 5 le o0 wesahiv ot p ilp wBddd 6 65 5500 W5 oo difa Wi araiea & so0 e 145
13.6: Spent Reactor Core and Heat Exchanger Cells ...; cvevvr e svmvrvs trnysvg uvdvangeey i 145
1T WEBIE SIOrame Gl .-c . 5.« < 3 22000 U0 5,505 agd-aall Al oudracs By 38, 000- s B Bis. kAN 5, SEorR 3 Airovy, 510 O 146
138 - Cherleal. Processing Cotl ' i 004« wivids & i naig Mi afewand sahel & a o 2% aei i o & avasind & ¥ 5 146
139 O SYSIBMITBHL 302,708 3wninm b e s A 55 A P R D8 STt T DSl S T S e Y S 5o 4 146
1'3.10- Miscellaneons Reactor BUllding Cells' ., . . . < cnevocrsnnsnomancasss smeneesssesseis 146
13.11 Steam-Generator Cells and Service A8 i« « s sireh s ain s vbsaistsdass pbniasssseis 146
1 3.1 PestwaterBaterand Tardineg BiltIogs ... « Dload tro v 51 miph o aie 3ohimm sislf ob 9 o0 5k Foaiwonrs 147

4. SITE PESCRIPTION | -3 10015575 aibiai-one: oo et inle b ma B 5.0 Amiats® V6 dv asTe s saBss 8D 36 Hacd 606 3 148
vii

15, COST ESTIMATES FOR THE MSBR STATION .. 4 a a5 ouere v sbedinsisiona osnie g ataism s sd s ol vl & » 150
151! CRpHal CoRt EsBmble ot - o i ive ice Slmain s brm o i ok 5 whaliiate o s otade ata o ivid s, a'a i v ik 150
192  Tower. Produchion Cosl . o0 f*7 6 a5 sle 0 . 00 R a3 6 w5 50t et K 4 Laneialk v sdlors 4% e 151

16, UNCERTAINTIES AND ALTERNATIVES, AND THEIR EFFECTS ON FEASIBILITY

AND FERPORMANCE ©.ii% a0 5560500 308 01756 5 BEnle D 88 et siete iub aidisns & 53 4 dvw Bu Aom s a/in 154
Y61 OB S cwr s SR e T b T e dt DS eE ¢ b AN e B B aate e Fe el a R L5 300 o B LS 2. 154
162 NIABHAIE . e ndinimt Bk $I aaehe Be 50 a win o8 Suapsdaln B0 WA 518 Seidde §it Agevs w.ae B9 5S 154
L0221  BUBLSEIL 36 A0 b ncio i b BiBiiis 656 ornion 1 b i dieiion's eIV Oa S A Salsbils 8% 154

36.4.2" . SeoonNAREY TINMIA . an.i'e sis o6 miteare it sy min e by’ 4 Wiaw shie stels @.a e MELGR e it o 155

16253 "TRESEILOR I £ « o w9 ¥% M 75 5 e v 0 S R AR A SR s e 156

1624 GTAPIULE: o 4 50 s B as 7 5.5 4 3 A0 s 5 W G5 T WM o T T P B T B By AL W 157

16:3 Systéetis And CoORMANEIE 53 5vusi 5 3 o5 GBI WK a5 DVraED 037 abrd o DG BPE 57s 024 Bes Imbinis 158
16.3.) VEMetOP .. .15 e b adaniat ake e vah e rod sl de Bt & i a5 sitazt o @ et DA 158

1632 Primary Heat EXChangers : ;.8 i« s 0 0300 8 dradac o 55 0 s ofea b & @ o wn avoebon 570 158

16353 " 1St Choilatioy PUIMDE £ 505 6 s it S0 895 & s RS2 L4 R ATS DB 48 F5 + B AF AN puibacd ¥ 158

DE:IH,  INER TAUK -4 3% 05 msirniebe @oitrets oin & oliidajechirsh. HPe Sk o6 5855000 o5 Bl ma ¢ 159

16:3:5) PoolaSTLINEa VR . o5 s £ee dasen s o vk s asw 56 s gl o 8% 37 Pk rein & 159

16.3.6 Gaseous Fission Product Removal System . ... .....cc.cininariinnncasncenns 159

1637 OIFGAS BB ... 5 0% 50 wia s 695 6 5Pe8.Eig n b Ble §16%.906 753 & 5od s7& 0l 5 Wa lm jud o 00a s 159

DOLIE" "R GODBRIATONE: b i v 3 Soie & & W MIE W M. o SN Sr BRa e e B el SN Py 55 159

1639 ‘Ihstrumentation and CODIrOIR .55 o6, 6 am i 00 55 €0 2 ked ¥)a e soiies o1 sr 1 160

16.3.10 Piping and Equipment SUPPOTts -« : caaraa o si caman o Cansas bo vad-sead ss s 160

16.3: 11 'Coll Conictlon: 2. . 054 300002 £ o o U B TR 5 Sonrda 1008 et miaimn ot Sefl en aiara 160

164 " TRUGN CORBHSHERt A ¢ ¢ wis vy 55 5.080E b w2 G o ga o500 TPV (RS R .G v S 160
16.9 ' ‘Chemicsl Processibg SYRLAIN 4 & < 5/n'e s 5 in @k d F o & @ i o i AtRE B4 208 4o swibde 216 bty 6l toe 162
16:6 Fission PLOdUCt BEhAVIOE .0 5/s 1-c v 55 4% o 60606 0 Bt aih Fas g be as ne i ths il snshanl ife 163
16.7 Steam Conditions in the Thermal-Power Cycle .. ... v v eentinnnenvnneremncrssrvs 163
16.8 Maintenance Equipment and Procedures ... i .« cuswii swis sen aveinesms s aoradsns e ie 164
169 Safety Studies .. ... ..coivteii ittt entnereneonsosooneescseassancesssensos 165
REPERENCES. 58 5 ar ex s D% bodih ot aft shind oo A <) 4gtan mevi do $F Ton o ae e fine 5 sbe i e 166
APPENDIX.A: THEORY OF NOBLE GAS MIGRATION 's. ¢ o 545 moiy oo sipe v o8 40 modvd 494 r s seres 170
APPENDIXB: NEUTRONTHYSICS <40 p 05 Ch s 2 0ets & s - ide s el sa s i e ing mi aisa s 175

APPENDIX C. EQUIVALENT UNITS FOR ENGLISH ENGINEERING AND
INVEBRNATIONAL SYBTBME . 0ot drsn-eabed bablevasndden st is it ie s B0 0 es et 176
Summary

Preparation of a conceptual design for a 1000-MW(e)
single-fluid molten-salt reactor power station has given
confidence that such a plant is technically feasible and
economically attractive. Successful operation of the
Molten-Salt Reactor Experiment and the substantial
amount of research and development already accom-
plished on molten-salt reactor materials and processes
indicate that after the technology has been extended in
a few specific areas, a prototype Molten-Salt Breeder
Reactor (MSBR) plant could be successfully con-
structed and operated. Studies of the fuel-salt chemical
processing system are not as far advanced, but small-
scale experiments lead to optimism that a practical
system can be developed.

The reference MSBR operates on the Th-**3*U cycle,
with both fissile and fertile materials incorporated in
a single molten-salt mixture of the fluorides of lith-
ium, beryllium, thorium, and uranium. This salt, with
the composition LiF-BeF,-ThF;-UF; (71.7-16.0-12.0-
0.3 mole %), has a liquidus temperature of 930°F
(772°K), has good flow and heat transfer properties,
and has a very low vapor pressure in the operating
temperature range. It is also nonwetting and virtually
noncorrosive to graphite and the Hastelloy N container
material.

The 22-ft-diam by 20-ft-high reactor vessel contains
graphite for neutron moderation and reflection, with
the moderating region divided into zones of different
fuel-to-graphite ratios. As the salt flows upward through
the passages in and between the bare graphite bars,
fission energy heats it from about 1050°F (839°K) to
1300°F (978°K). Graphite control rods at the center of
the core are moved to displace salt and thus regulate the
nuclear power and average temperature, but these rods
do not need to be fast scramming for safety purposes.
Long-term reactivity control is by adjustment of the
fuel concentration.

The core neutron power density was chosen to give a
moderator life of about four years, based on the
irradiation tolerance of currently available grades of
graphite. The specific inventory of the plant, including
the processing system, is 1.47 kg of fissile material per

ix

MW(e), which, together with the breeding ratio of 1.06,
gives an annual fissile yield of 3.3%. The heat-power
system has a net thermal efficiency of over 44%, which
makes a reactor plant of about 2250 MW(t) ample for a
net electrical output of 1000 MW(e).

A simplified flow diagram of the MSBR is shown in
Fig. S.1. The primary salt is circulated outside the
reactor vessel through four loops. (For simplicity, only
one loop is shown in the figure.) Each circuit contains a
16,000-gpm single-stage centrifugal pump and a shell-
and-tube heat exchanger, Tritium, xenon, and krypton
are sparged from the circulating primary salt by helium
introduced in a side stream by a bubble generator and
subsequently removed by a gas separator. A 1-gpm
(0.06 liter/sec) side stream of the primary salt is
continuously processed to remove ??3Pa, to recover the
bred 233U, and to adjust the fissile content. A drain
tank provides safe storage of the salt during mainte-
nance operations.

Heat is transferred from the primary salt to a
secondary fluid, sodium fluoroborate, having a compo-
sition of NaBF;-NaF (92-8 mole %) and a ligaidus
temperature of 725°F (658°K). Each of the four
secondary circuits has a 20,000-gpm centrifugal pump
with variablespeed drive. The secondary salt streams
are divided between the steam pgenerators and the
reheaters to obtain 1000°F steam temperatures from
each, Steam is supplied to a single 3500-psia,
1000°F/1000°F, 1035-MW(e) turbine-generator unit
exhausting at 1% in. Hg abs. Regenerative heating and
live steam mixing are used to heat the feedwater
entering the steam generator to 700°F (644°K) to
provide assurance that the coolant salt remains liquid.

The estimated plant capital costs for a fully developed
MSBR, although differing in breakdown, are about the
same as those for a light-water nuclear power station.
Fuel-cycle costs are expected to be quite low and
relatively insensitive to the prices of fissile and fertile
materials.

The major uncertainties in the conceptual design are
in the areas of tritium confinement, fuel-salt processing,
graphite and Hastelloy N behavior under irradiation,
suitability of the coolant salt, maintenance procedures, Principal design data for the reference MSBR power
and behavior of the fission product particulates. Al-  station are listed in Table S.1 both in English engi-
though more study is needed of these aspects, it is  neering units, as commonly used in the molten-salt
believed that they can be resolved with reasonable reactor literature, and in the International (metric)
difficulty. system of units.

ORNL-DWG 70-11SD6

FLOW DIVIDER
. .
BOTTLE = a4
STORAGE 18 17 16
CLEAN
PURGE
He E
15
14 3 4
12| *He
i He I
NS0°F
1
1300°F | Y 550°F  6x105 ib/hr
”I MIXER
2
"I sso°r |
1050°F 95 x 108 In/hr 71 %105 1b/he / \
it =
CHEMICAL

PROCESSING_}

Fig. S.1. Simplified flow diagram of MSBR system. (1) Reactor, (2) Primary heat exchanger, (3) Fuel-salt pump, (4) Coolant-salt
pump, (5) Steam generator, (6) Steam reheater, (7) Reheat steam preheater, (8) Steam turbine-generator, (9) Steam condenser, (10)
Feedwater booster pump, (11) Fuel-salt drain tank, (12) Bubble generator, (13) Gas separator, (14) Entrainment separator, (15)
Holdup tank, (16) 47-hr Xe holdup charcoal bed, (17) Long-delay charcoal bed, (18) Gas cleanup and compressor system.
xi

Table S.1. Summary of principal data for MSBR power station

Engineering units?

International system units?

General

Thermal capacity of reactor
Gross electrical generation
Net electrical output

Net overall thermal efficiency
Net plant heat rate

Structures

Reactor cell, diameter X height
Confinement building, diameter X height

Reactor

Vessel 1D

Vessel height at center (approx)

Vessel wall thickness

Vessel head thickness

Vessel design pressure (abs)

Core height

Number of core elements

Radial thickness of reflector

Volume fraction of salt in central core zone

Volume fraction of salt in outer core zone

Average overall core power density

Peak power density in core

Average thermal-neutron flux

Peak thermal-neutron flux

Maximum graphite damage flux (>50 keV)

Damage flux at maximum damage
region (approx)

Graphite temperature at maximum neutron
flux region

Graphite temperature at maximum graphite
damage region

Estimated useful life of graphite

Total weight of graphite in reactor

Maximum flow velocity of salt in core

Total fuel salt in reactor vessel

Total fuel-salt volume in primary system

Fissile-fuel inventory in reactor primary
system and fuel processing plant

Thorium inventory

Breeding ratio

Yield

Doubling time, compounded continuously,
at 80% power factor

Primury heat exchangers (for each of 4 units)

Thermal capacity, each
Tube-side conditions (fuel salt)
Tube OD
Tube length (approx)
Number of tubes
Inlet-outlet conditions
Mass flow rate
Total heat transfer surface

Shell-side conditions (coolant salt)

Shell ID
Inlet-outlet temperatures
Mass flow rate

Overall heat transfer coefficient (approx)

2250 MW(t)
1035 MW(e)
1000 MW(e)
44.4%

7690 Btu/kWhr

72X 421t
134 X 189 ft

0.37

22.2 kW/liter

70.4 kW/liter

2,6 X 10'® peutrons cm ™2 gec !
8.3 X 10'4 neutrons cm 2 sec ™!
3.5 X 10'? neutrons cm ™2 sec™!
3.3 X 10'* neutrons em 2 sec”!

1284°F
1307°F

4 years
669,000 1b
8.5 fps
1074 ft2
1720 13
3316 1b

150,000 1b
1.06

3.2 %/year
22 years

556.3 MW(1)

3/3 in.

22.2 ft

5896
1300-1050°F
23.45 x 10% ib/hr
13,000 f£t?

68.1 in.
850-1150°F
17.6 X 10% 1b/hr

850 Btu hr! f1~2 (°F) !

2250 MW(t)
1035 MW(e)
1000 MW(e)
44 49,

2252 J/kW-sec

220X 128 m
40.8 X 57.6 m

6.77 m

6.1m

5.08 cm

7.62 cm

5.2 X 105 N/m?

3.96 m

1412

0.762 m

0.13

0.37

22.2 kW/liter

70.4 kXW/liter

2.6 X 104 neutrons cm 2 sec "
8.3 X 10'? neutrons cm ™2 sec™}
3.5 X 10'? neutrons cm 2 sec™!
3.3 % 10'® neutrons cm 2 sec ™

969°K
982°K

4 years
304,000 kg
2,6 m/sec
30.4 m?
48.7 m?
1504 kg

68,100 kg
1.06

3.2 %/year
22 years

556.3 MW(t)

0.953 cm
6,8 m

5896
978—-839°K
2955 kg/sec
1208 m?

1.73 m

727-894°K

2218 kg/sec

4820 Wm™2 (°g)!
*®1i

Table S.1 (continued)

Engineering units?

International system units?

Primary pumps (for each of 4 units)
Pump capacity, nominal
Rated haad
Speed
Specific speed
Impeller input power
Design temperature

Secondary pumps (for each of 4 units)
Pump capacity, nominal
Rated head
Speed, principal
Specific speed
Impeller input power
Design temperature

Fuel-salt drain tank (1 unit)

Outside diameter
Overall height

Storage capacity
Design pressure
Number of coolant U-tubes
Size of tubes, OD
Number of separate coolant circuits
Coolant fluid

Under normal steady-state conditions:

Maximum heat load

Coolant circulation rate

Coolant temperatures, in/out
Maximum tank wall temperature

Maximum transient heat load

Fuel-salt storage tank (1 unit)
Storage capacity
Heat-removal capacity
Coolant fluid

Coolant-salt storage tanks (4 units)
Total volume of coolant salt in systems
Storage capacity of each tank
Heat-remaval capacity, first tank in series
Steam generators (for each of 16 units)
Thermal capacity
Tube-side conditions (steam at 3600-3800
psi)
Tube OD
Tube-sheet-to-tube-sheet length (approx)
Number of tubes
Inlet-outlet temperatures
Mass flow rate
Total heat transfer surface
Shell-side conditions (coolant salt)
Shell 1D
[nlet-outlet temperatures
Mass tlow rate

Apparent overall heat transfer coefficient
range

16,000 gpm

&N £a
v oA

890 rpm

2625 rpm(gpm)?*5 /()0 75
2350 hp

1300°F

20,000 gpm

300 ft

1190 rpm

2330 rpm(gpm)®*5 /(f)°-75
3100 hp

1300°F

14 ft

22 fi
2500 ft?
55 psi
1500

3/4 in.

40
TLiF-BeF,

18 MW(t)
830 gpm
900-1050°F
~1260°F
53 MW(t)

2500 ft3
1 MW(t)
Boiling water

8400 £t3
2100 £t
400 kW

120.7 MW(1)

1/2 in.

76.4 ft

393
700-1000°F
633,000 Ib/hr
3929 1t?

1.5 ft
1150-850°F
3.82 X 105 Ib/hr

490-530 Btu hr ™! ft™2 (°F) !

1.01 m®/sec

43.7m

93.2 radians/sec

5.321 radians/sec(m?/sec)®5 /(m)0*73
1752 kW

978°K

1.262 m?/sec

91.4m

124.6 radians/sec

4.73 radians/sec{m>/sec)?*5 /(m)®' 75
2310 kW

978°K

427 m

6.7l m

70.8 m?

3.79 X 10° N/m?
1500

1,91 ecm

40

TLiF-BeF;

18 MW(1)
0.0524 m?/sec
755-839°K
~955°K

53 MW(1)

70.8 m?
1 MW(t)
Boiling water

237.9 m?
59.5 m?
400 kW

120.7 MW(1)

1,27 cm
23.3m

393
644-811°K
79.76 kg/sec
365 m?

0.457 m
894-727°K
481.3 kg/sec

2780-3005 W m ™2 (°K) ™!
Table 8.1 (continued)

Engineering units? International system units?
Steam reheaters (for each of 8 units)
Thermal capacity 36.6 MW(t) 36,6 MW(1)
Tube-side conditions (steam at 550 psi)
Tube OD 3/4 in, 1.9 ecm
Tube length 30,3 ft 9.24 m
Number of tubes 400 400
Inlet-outlet temperatures 650-1000°F 616-811°K
Mass flow rate 641,000 1b/hr 80.77 kg/sec
Total heat transfer surface 2381 ft? 221.2 m?
Shell-side conditions (coolant salt)
Shell ID 21,2 in, 0.54 m
Inlet-outlet temperatures 1150-850°F 894-727°K
Mass flow rate 1,16 X 108 Ib/hr 146.2 kg/sec
Overall heat transfer coefficient 298 Btu hr ™! 72 (*F)~! 1690 W m ™2 (°K) ™
Turbine-generator plant (see “‘General” above)
Number of turbine-generator units 1 1
Turbine throttle conditions 3500 psia, 1000°F 24,1 X 10° N/m?, 811°K
Turbine throttle mass flow rate 7.15 % 10° Ib/hr 900.9 kg/sec
Reheat steam to IP turbine 540 psia, 1000°F 3.72 X 10% N/m?, 811°K
Condensing pressure (abs) 1.5 in. Hg 5,078 N/m?
Boiler feed pump work 19,700 hp 14,690 kW
(steam-turbine-driven), each of 2 units
Booster feed pump work (motor-driven), 6200 hp 4620 kW
each of 2 units
Fuel-salt inventory, primary system
Reactor
Core zone I 290 ft3 8.2 m?
Core zone 11 382 ft? 10.8 m®
Plenums, inlets, outlets 218 ft3 6.2 m*
2-in. annulus 135 f° 3.8 m?
Reflectors 49 ft3 1.4 m?
Primary heat exchangers
Tubes 269 ft3 7.6 m>
Inlets, outlets 27§63 0.8 m3
Pump bowls 185 ft3 5.2m?
Piping, including drain line 145 ft3 4.1 m®
Off-gas bypass loop 10 3 0.3m°
Tank heels and miscellaneous 10 i3 0.3 m3
Total enriched salt in primary system 1720 f3 48.7 m?
Fucl-processing system (Chemical Treatment
Plant)
Inventory of barren salt (Lif-BeF,-ThF ) 480 £ 13.6 m®
in plant
Processing rate 1 gpm 63,1 X 1078 m3/sec
Cycle time for salt inventory 10 days 10 days
Heat generation in salt to processing plant 56 kW/ft? 1980 kW/m?
Design properties of fuel salt
Components 7LiF-BeF,-ThF4-UF4 TLiF-BeF;-ThF4-UF 4
Composition 71.7-16-12-0.3 mole % 71.7-16-12-0.3 mole %
Molecular weight (approx) 64 64
Melting temperature (approx) 930°F 772°K
Vapor pressure at 1150°F (894.3°K) <0.1 mm Hg <13.3 N/m?
Density:€ p (g/cm?) = 3.752 — 6.68 X 10™%
CC); p (Ib/ft3) = 235.0 — 0.02317¢ (°F)
At 1300°F (978°K) 204.9 Ib/fe® 3283.9 keg/m?
At 1175°F (908°K) 207.8 Ib/it? 3330.4 kg/m?

At 1050°F (839°K) 210.7 1b/ft? 3376.9 kg/m?
Xiv

Table S.1 (continued)

Engineering units? International system units?

Viscosity:d M (centipoises) = 0.109

exp [4090/T (°K)); u (Ib ft=' hr™t)

=0.2637 exp [7362/T (°R)]

At 1300°F (978°K) 17.31bhr! f17! 0.007 N sec m™2

At 1175°F (908°K) 23.8 1b hr™? ft7? 0.010 N secm™2

At 1050°F (839°K) 34,5 b he ™! £t 0.015 N sec m™2
Heat capacity® (specific heat, ¢p) 0.324 Btulb™ C°F)"! £ 4% 1.3571g 1 CK) 7Y £ 49
Thermal conductivity (k)/

At 1300°F (978°K) 0.69Btuhr™! °H) ! fit7! LI9Wm™ CK)™!

At 1175°F (908°K) 0.71 Btu hr ™! °F) ! fr™? 1.23Wm™! CK)™!

At 1050°F (839°K) 0.69 Btu he ™! °F) "1 ! 1.19Wm™ (°K) !

Design properties of coolant salt

Components NaBF 4-NaF NaBF 4-NaF
Composition 92-8 mole % 92-8 mole %
Molecular weight (approx) 104 104
Melting temperature (approx) 7125°F AS8°K
Vapor pressure:® log # (mm Hg)

=9.024 — 5920/T (°K)

At 850°F (727°K) 8 mm Hg 1066 N/m?

At 1150°F (B94°K) 252 mm Hg 33,580 N/m*
Density:€ p (g/em®) = 2.252 - 7.11 X 10 %¢

(°C); p (Ib/1t3) = 141.4 — 0,0247t (°F)

At 1150°F (894°K) 113.0 Ib/f? 1811.1 kg/m?

At 1000°F (811°K) 116.7 b/ft? 1870.4 kg/m?

At 850°F (727°K) 120.4 lb/fe3 1929.7 kg/m?
Viscosity:? u (centipoises) = 0.0877

exp [2240/T (°K)]; & (Ib,, £t™! hr71)

=0.2121 exp [4032/T ("R)]

AL 1150°F (894°K) 2.6 1bft™ het 0.0011 N seem™2

At 1000°F (B11°K) 3.41bft™ het 0.0014 N sec m~2

At 850°F (727°K) 461 ft~! hr! 0.0019 N sec m ™2
Heat capacity” (specific heat, ¢ ) 0.360 Btu b~ (°F) ™! + 2% 1507 Tkg™ (°K) ™! £ 2%
Thermal conductivity (k)*

At 1150°F (894°K) 023Bwhr™! °FH~t it 0.398 Wm™ (°K)!

At 1000°F (811°K) 0.23 Btu hr™! (°F)~! 17! 0398 Wm™ (°K)!

At 850°F (727°K) 0.26 Btu hr ™! (°F)~! 1it™! 0450 Wm™ (°K)!

Design properties of graphitei

Density, at 70°F (294.3°K) 115 1b/ft3 1843 kg/m?
Bending strength 40006000 psi 28 X 105-41 X 10° N/m?
Modulus of elasticity coefficient 1.7 X 105 psi 11.7 ¥ 10 N/m?
Poisson’s ratio 0.27 0.27
Thermal expansion coefficient 23X 107%/°F 1.3 % 107%/°K
Thermal conductivity at 1200°F, 18 Btu hr ™! (°F) ! £t~} N2Wmt CK !

unirradiated (approx)
Electrical resistivity 89X 10%-9.9% 10™ Q-cm 89X 10-99x% 10 Q-cm
Specific heat

At 600°F (588.8°K) 0:33 B b~} (°F)7! 1380 J kg™ (°K) 1

At 1200°F (922,0°K) 0.42 Btu lb™ (°F)! 1760 J kg™ (°K) !
Helium permeability at STP with sealed 1% 1078 em?/sec 1% 1078 cm?/sec

surfaces
XV

Table S.1 (continued)

Engineering units?

International system units?

Design properties of Hastelloy NK
Density
At B0°F (300°K)
At 1300°F (978°K)
Thermal conductivity
At 80°F (300°K)
At 1300°F (978°K)
Specific heat
At 80°F (300°K)
At 1300°F (978°K)
Thermal expansion
At 80°F (300°K)
At 1300°F (978°K)
Modulus of elasticity coefficient
At 80°F (300°K)
At 1300°F (978°K)
Tensile strength (approx)
At 80°F (300°K)
At 1300°F (978°K)
Maximum allowable design stress
At 80°F (300°K)
At 1300°F (978°K)

Melting temperature

557 b/ft?
541 b/

6.0Btuhr™ CF)7! £t
12.6 Btuhr™! (°F) 7! 1t~?

0.098 Btu Ib ™! (°F)~}
0.136 Btu Ib ™! (°F)~!

5.7% 1078PF
9.5% 1078 °F

31 x 10 psi
25 % 105 psi

115,000 psi
75,000 psi

25,000 psi
3500 psi

2500°F

8927 kg/m?
8671 kg/m?

104Wm™ (°K)!
20 8Wm™! (°K) !

410 T kg™ ("K) !
569 1 kg™ (°K) !

3.2% 1078 °K
53% 107%°K

214 X 10? N/m?
172 % 10% N/m?

793 X 10% N/m?
517 x 105 N/m?

172 X 10® N/m?
24 % 10° N/m?

1644°K

?English engineering units as used in MSR literature.

Meter-kilogram-second system. Table closely follows International System (SI). See Appendix C for conversion factors from
engineering to SI units,

“See p. 147, Fig. 13.6, ORNL-4449 (ref. 1).
dSce p. 145, Table 13.2, ORNL-4449 (ref. 1).
“See p. 163, ORNL-4344 (ref. 2).

Sce p. 92, Fig. 9.13, ORNL-4449 (ref. 1), The value of k¥ shown is for salt with about 5% less LiF than the reference salt. Addition
of LiF would increase the average value, probably to 0.72-0.74. The established, and conservative, value of 0.71 was used in the
MSBR calculations.

2See p. 170, ORNL-4254 (ref. 3).

"See p. 168, ORNL-4254 (ref. 3).

See p. 92, Fig. 9.13, ORNL-4449 (ref. 1),
’Additional graphite properties are listed in Table 3.4,

Composition, wt %: Ni, balance; Mo, 12; Cr, 7; Fe, 0-5; Mn, 0.2-0.5; 8i, 0.1 max; B, 0.001 max; Ti, 0.5-2.0; Hf or Nb, 0-2; Cu,
Co, P, S, C, W, Al (total), 0.35.
Acknowledgment

This design study was carried out by the staff of the Molten-Salt Reactor Program (M. W. Rosenthal, Director;
R. B. Briggs and P. N. Haubenreich, Associate Directors). It was a team effort involving personnel from eight dif-
ferent divisions of the Oak Ridge National Laboratory. The conceptual design upon which the study is based was
produced by a group led by E, S. Bettis. The persons listed below all made substantial contributions to the study.

John L. Anderson
H. F. Bauman

C. E. Bettis

E. S. Bettis
Howard [. Bowers
Robert Blumberg
R. B. Briggs

J. H. Carswell
Lloyd Carter

C. W. Collins

W. H. Cook

W. K. Crowley

W. P. Eatherly

S

W. K. Furlong
W. R. Grimes
P.N.
J

Haubenreich

Peter P. Holz
P. R. Kasten
R. J. Kedl

H. T. Kerr

C. R. Kennedy
M. I. Lundin

xvi

THPEmEI SO E T

2EErprrPmemm

2222z

oy Z

sl
e
<

cCoy

cLain
cNeese

cWherter

oore
yers
elms
icholson

erry
1ckel

acPherson

Roy C. Robertson
M. W, Rosenthal
Dunlap Scott

W. H. Sides

M. Siman-Tov
A.N. Smith

Watts
1. Introduction

A major objective of the Molten-Salt Reactor Program
is to achieve a power reactor which will produce electric
energy at low cost and at the same time exiend the
nation’s low-cost fuel resources. A graphite-moderated
thermal breeder reactor making use of solutions of
fissile and fertile materials in fluoride carrier salts shows
considerable potential for meeting this objective. This
report summarizes present information on the design
characteristics of such a Molten-Salt Breeder Reactor
(MSBR).

Molten salts as reactor fuels and as coolants have been
under study and development for over 20 years, and
their chemical, physics, and irradiation properties are
excellent. The Molten-Salt Reactor Experiment (MSRE)
at ORNL, which was recently shut down after about
five years of very successful operation, contributed
significantly to molten-salt reactor technology. A sur-
vey report® was published in August 1966 which
summarized the potential of molten-salt thermal
breeder reactors and described preliminary designs and
fuel processing facilities for a 1000-MW(e) power
station, More detailed design studies followed,”” and a
comprehensive report® was written which covered the
status of the design studies as of January 1968. These
reports considered the two-fluid reactor concept, that
is, one in which the fissile atoms are carried in one
molten-salt solution, called the fuel salt,* and the fertile
material in another, called the blanket salt. In the fall of
1967, however, information was obtained that made a
single-fluid MSBR, in which fissile and fertile materials
are dissolved in the same salt, appear practical and
attractive. The two-fluid study was set aside and a
design study of the single-fluid system commenced.
Some of the factors involved were:

*The terms “primary salt” and *‘fuel salt”’ are used synony-
mously throughout the molten-salt reactor literature. In the
case of the single-fluid MSBR described in this report, the
primary salt contains both the fuel and fertile material. The
terms ‘‘secondary salt” and “coolant salt” are also used syn-
onymously.

1. Research in the processing of the molten-salt fuels
showed that protactinium and other fission products
could be separated from the salts containing both
uranium and thorium by reductive extraction into
liquid bismuth. A single salt containing both the
fissile and fertile materials could thus be processed,
although with more difficulty than if separate fuel
and fertile salts were used.

2. Nuclear calculations indicated that a conversion
ratio greater than 1.0 could be achieved in a one-fluid
reactor with an acceptably low inventory if the
graphite-to-fuel ratio were reduced in the outer
regions of the reactor core. While the fuel specific
power fell short of the performance of a two-fluid
type, yields of 3 to 4%)/year were indicated.

3. Reactor exposure limitations were found to exist
relative to use of a graphite moderator, making it
necessary to design for graphite replacement. In a
two-fluid reactor it appears more practical to replace
the entire reactor assembly, including the reactor
vessel, when replacing the graphite. The single-fluid
MSBR, however, permits easier access through the
top head, so that only the core graphite need be
replaced.

4. The two-fluid concept depends upon the integrity of
the graphite “plumbing” in the reactor vessel to
keep the fuel and fertile salt streams separated. The
single-fluid design eliminates this potential problem.

5. Radiation damage to graphite during reactor ex-
posure leads to dimensional changes in graphite
which are more easily accommodated in a single-
Tuid MSBR than in a two-fluid design.

The progress of the single-fluid design study is
covered in the MSRP semiannual reports,!+2:3:% and
the entire February 1970 issue of Nuclear Applications
and Technology'® was devoted to a review of molten-
salt reactor technology and to a description of a
conceptual design for an MSBR, Some of the general
criteria for the single-fluid MSBR design study are:
1. The design study is to establish concept feasibility,
to serve as a basis for preliminary estimates of cost
and performance features, to identify the research
and development needed to achieve a full-scale
MSBR, and to guide the design of an experimental

: £ ene w=f il o
prototype reactor that will test the features of the

larger plants to follow.

2. The conceptual design of the MSBR is to be based
on a technology which does not require major
inventions or technological breakthroughs. Reason-
able engineering development is considered permis-
sible, however.

3. The conceptual design is to be based on a plant
capacity of 1000 MW(e).

4. Cost estimates are to be based on existence of a
well-established MSBR power reactor industry.

The design of the MSBR plant is presented in terms of
various systems, or facilities, which are categorized as:

1. the reactor system, in which fission heat generated
in the fuel salt in its passage through the reactor
vessel is removed in primary heat exchangers;

2. an off-gas system for purging the fuel salt of fission
product gases and gas-borne particulates;

3. a chemical processing facility for continuously re-
moving fission products from the fuel salt, recover-
ing bred 222U, and replenishing fertile material;

#

a coolant-salt circulating system, steam generators,
and a turbine-generator plant for converting the
thermal energy into electric power;

5. general facilities and equipment, including controls
and instrumentation, maintenance tools, auxiliary

power equipment, waste disposal systems, condens-
ing water works, electrical switchyard, stacks, and
conventional buildings and services.

The above categories are not always separate and are
closely interdependent, but it is convenient to discuss
them separately, The reactor and its related structures
and maintenance system, the drain tank, the off-gas
system, and the chemical processing system are of
primary interest and are discussed in more detail in the
following sections. The steam turbine plant and the
general facilities are more or less conventional and are
discussed only to the extent necessary to complete the
overall picture as to feasibility and costs of an MSBR
station.

There are many alternatives open to the designer of
an MSBR station. These can be resolved by detailed
optimization work, but to initiate this preliminary
study it was necessary in many areas to make early
decisions largely on the basis of considered judgment,
Some examples are: selection of the number of coolant
loops and steam generators, use of 700°F feedwater, an
assumed useful graphite life of four years, etc. The
reference design described here, therefore, illustrates
that an MSBR power station is practical and feasible,
but it does not represent a design which has been
optimized for best performances and costs.

An effort has been made to revise and annotate the
report to indicate the status of the technology, particu-
larly with regard to the behavior of materials, up to the
late fall of 1970. As indicated above, however, major
features of the conceptual design were established much
earlier, generally on the basis of information available in
late 1969 and early 1970.
2. Overall Systems Descriptions and Features

E. S. Bettis

2.1 REACTOR PRIMARY SYSTEM

The MSBR primary system consists of the reactor,
four primary heat exchangers that transfer heat from
the fuel salt to the coolant salt, and four pumps that
circulate the molten fluoride fuel-salt mixture. All of
this equipment is contained within the reactor cell, as
shown in Sect. 13. The fuelsalt drain tank and
afterheat-removal equipment are considered to be a
scparate system and are described in Sect. 2.4.

The reactor primary system flowsheet is shown in Fig.
2.1. About 94.8 X 10° Ib/hr of fuel salt enters the
bottom of the reactor at 1050°F. Fission energy within
the graphite-moderated core raises the salt temperature
to an average value of 1300°F at the reactor exit at the
top. The salt then enters the bottom of the four
fuel-salt circulation pumps. (For simplicity, only one of
the four circuits is shown in Fig. 2.1.) These centrifugal
pumps force the salt through the tubes of the four
shell-and-tube primary heat exchangers, where the fuel
salt is cooled to about 1050°F before returning to the
bottom of the reactor.

Each of the fuel-salt circulation pumps has a bypass in
which about 10% of the total pump discharge flow is
circulated. This loop contains a gas bubble injection
section, where a sparging gas (principally helium) is
introduced as small bubbles. The bubble generator is a
venturi-like section in the pipe capable of generating
bubble diameters in the range of 15 to 20 mils. The
same bypass loop contains a gas separator, upstream of
the bubble generator, which removes the inert gas and
its burden of fission products with nearly 100%
stripping efficiency. Downstream vanes kill the swirl
imparted by the centrifugal gas separator. The removed
fission products consist principally of xenon, krypton,
tritium, and exceedingly small particles of noble metals.
Based on 10% bypass flow, after a bubble is introduced
it would make an average of ten passes through the
reactor before being removed by the separator.

The removed gases, along with a small amount of
entrained salt, are taken to a small tank, where the
off-gas is combined with that purged from the pump
bowls and from the exit annulus at the top of the
reactor. Since the off-gas leaving this tank is intensely
radioactive, the line is cooled by a jacket in which there
is a flow of 1050°F fuel salt taken from the reactor
drain line just upstream of the freeze valve. This
relatively small flow of fuel salt, which is subsequently
returned to the pump bowl, also assures an open line
between the drain valve and the reactor vessel.

Each fuel-salt pump bowl overflows about 150 gpm
through the small tank, and this fluid flows with the
off-gas to the drain tank. The overflow arrangement
simplifies liquid level control and helps cool the drain
tank head and walls. Salt-operated jet pumps at the
bottom of the drain tank continuously return the
molten salt to the circulation systems, as described in
Sect. 2.4. The drain tank is provided with ample
afterheat-removal capacity.

The fuel-salt drain tank is connected to the bottom of
the reactor vessel by a drain line having a freeze-plug
type of “valve.”” At the discretion of the plant operator,
the plug can be thawed in a few minutes to allow
gravity drain of salt from the system into the drain
tank. The freeze plug would also thaw in the event of a
major loss of electric power or failure of the plug
cooling system. The drain system is provided primarily
in the event a leak develops in the fuel salt circulating
loop and for safe storage of salt during maintenance
operations. Although drainage is a positive reactor
shutdown mechanism, it is not normally used as an
emergency procedure since the reactor control and
safety rods can quickly take the reactor subcritical
while fuel-salt circulation is continued to remove fission
product decay heat via the primary heat exchangers.

A catch basin is provided at the bottom of the heated
reactor cell in the unlikely event of a major spill of fuel
salt from the system. The basin pitches toward a drain
ORNL DWG 69-10494AR

y CHEMICAL PROCESS PURGE
o T He SUPPLY SECONDARY PUMP PURGE
% X % ACCUMULATOR REACTOR CELL SECONDARY EXHAUST
o1 | o FILTERS MAINTENANCE VENT LINE
DELAY t PRETREAT
g | . i eole
MAINTENANCE ¥ 3
HoLBop & CONTAINMENT FILTERS
WL HI- VENTILATION
BED STO! SURGE § CHARCOAL
; TANK T BED
, 85kr AND 1 C
TRITIUM
REMOVAL GAS 5 s
TANKS coocant—] [ —codthnr O%Tack”
He RETURN _ waTER VAPOR & ¥ PARTICLE TRAPS
ALARM AND TRAP
MAINTENANCE VENT LINE PRIMARY SECONDARY
4——(’0] SALT PUMP SURGE SALT PUMP
1A | He <] TANK 110" 1b/hr STEAM AT 1000 °F
REACTOR ! X VENT_  !\X107 ib/he FEEDWATER AT 700 °F
L \
A iy 4 U ST .
PPLY : 6 REHEAT
! o i | 196%10% ib/hr l' o
r LR | AT 150 °F AT 1000 °F
7
e 9.48X107 Ib/hr j . .
AT 1300°F GAS y 742107 1b/he T T 850°F
”’”””I EPARATOR AT 150 °F 6.16X107 (b/hr a8
REACTOR RSO
6108
I l\fiffi?h L b/hr | STEAM REMEATER
_ e GENERATOR
HEAT EXCHANGER 85Q°F
9.48x10" [b/hr AT 1050 *F g ) 850°F
'\| [~1.0XI0° fb/hr, PUMP COOLANT [ T o
CATCH BASIN s * 3 "RS'EFF—:KT_
AT 650°F
ANT
|~ VENT LINE rom—— s ~=
—IF} * PRESSURE
] | 24 ) OR VENT
cHemeaL 1= CONDENSATE 3 1 RO
CONDENSATE 7o yEAT J | x
PROCESSING 2 13 REJECT DENSATE STACK  SEonon VANK
STACK
TANK | " ; FROM .= o
| CHEMICAL
10 I o PROCESSING (¥ COOLANT (HEATING) GAS
CHEMIGAL = 3 ! g l CW COOLANT WATER
PROCESSING b--emeemesn | i ) { [ PRIMARY SALT [F] FREEZE VALVE
g O Ane The gy copg Py o FLOW RATES ARE TOTAL FOR

1000 Mw(e) PLANT

Fig. 2.1. Flow diagram for MSBR reactor plant. (A) Reactor core, (B) Fuel-salt circulating pumps, (C) Primary heat exchangers,
(D) Bubble generators, (E) Gas separators, (FF) Off-gas combiner tank, (G) Fuel-salt drain tank and gas holdup volume, (H) Particle
traps, (I) Charcoal beds for Xe and Kr holdup, (J) Off-gas cleanup and storage system, (K) Coolant-salt circulating pumps, (L) Steam

generators, (M) Steam reheaters.
which would allow the salt to be collected in the
fuel-salt drain tank.

A fuel-salt storage tank is provided in addition to the
drain tank in the event the latter requires maintenance.
The heat-removal system for the storage tank has less
stringent requirements and consists of simple U-tubes
immersed in the salt. Water is boiled in the tubes and
the steam condensed in a closed system by air-cooled
coils located in the base of the natural-draft stack. A jet
pump in this tank is used to return the fuel salt to the
circulation system or to the drain tank.

2.2 SECONDARY-SALT CIRCULATION SYSTEM

The secondary system in the MSBR consists of the 4
coolant-salt circulation pumps, 16 steam generators,
and 8 reheaters, all located in the steam generator cells,
as described in Sect, 13. Coolant-salt storage tanks are
located in cells directly beneath the steam generator
cells.

The molten sodium fluoroborate coolant salt is
circulated at a rate of about 71.2 X 10° Ib/hrt, as
indicated on flowsheet in Fig. 2.1. The coolant enters
the shell side of the primary heat exchangers at 850°F
and leaves at 1150°F. Each of the four coolant-salt

pumps circulates the coolant through four steam
generators and two steam reheaters, with the flow
proportioned so that outlet steam temperatures of
1000°F are obtained from each. The coolant-salt pumps
can be operated at variable speed to minimize tempera-
ture excursions during power transients, and the
steady-state temperature can be adjusted to match
station load.

2.3 STEAM-POWER SYSTEM FOR THE
TURBINE-GENERATFOR PLANT

The steam-power system consists of a single 1035-
MW(e) gross electrical capacity turbine-generator unit,
condensing system, condensate polishing and regener-
ative feedwater heating systems, steam-turbine-driven
main feedwater pumps, feedwater and reheat steam
preheating equipment, and associated controls, switch-
gear, station output transformers, etc. All the steam-
power system equipment, with the exception of the
feedwater and reheat steam preheating facilities, is
conventional in present-day power stations and will not
be described in detail.

A simplified steam system flowsheet is shown in Fig.
2.2, and some of the principal data are summarized in

ORNL-0OWG 70-1307

540 psio  1000°F 5. ¥ 408 /e
10 % 10% im/hr 3600psi0_ 1000°F 7.2 % 105 w/hv

1150"F,
e 1k 1150°F

} STEAM

PREMEATER l
“» |GENERATOR < HP GENERATOR
\, 1932 Mw(l) > 1D Mw(') 600 pslo  §52°F T
U I iy

= Ll Y HADP

3500 pslo
BBE°F 2.9x10°8 Ib/hr

BOOSTER
PUMPS
46 Mw(e)

b

REACTOR HEAT = 2250 Mw(1)

GROSS QUTPUT = 1035 Mw(e)

NETOUTPUT = 1000 Mw(e)
NET EFFICIENCY = 444 T

\72 P 7T06°F

GENERATOR
508 Mwle)
1.5in. Hg obs

CONDENSER CONDENSER

45‘ 3 STAGES |
FW HEATING ¥

£ 'EH;]
5

5 STAGES
|_FW HEATING

MIXING
CHAMBER
695°F 550°F 3STAGES L.
MIXING _FW HEATING |
CHAMBER

..... . 8
5 STAGES
FW HEATING |

Fig. 2.2. Simplified MSBR steam system flowsheet,
Table S.1. About 7.15 X 10° Ib/hr of steam at 3500
psia and 1000°F is delivered to the turbine throttle. The
high-pressure turbine exhaust steam is first preheated to
650°F and then reheated to 1000°F before readmission
to the intermediate-pressure turbine. The turbine ex-
hausts at 1'%, in. Hg abs to water-cooled candencere,
The turbine is indicated on the flowsheet as a cross-
compounded unit, but a tandem-compound machine
could be used.

Eight stages of feedwater heating are shown, with
extraction steam taken from the high- and low-pressure
turbines and from the two turbine-driven boiler feed
pumps. The 600-psia, 552°F steam from the high-
pressure turbine exhaust is preheated to about 650°F in
a steam-to-steam U-shell, U-tube type heat exchanger,
with steam (at about 3600 psia and 1000°F) from the
steam generator outlet entering the tube side and
leaving at about 866°F. This exit steam is directly
mixed with the high-pressure 551°F feedwater leaving
the top extraction heater to raise the water temperature
to about 695°F. Motor-driven canned-rotor centrifugal
pumps then boost the water from about 3500 to 3800
psia and 700°F before entering the steam generator.

A supercritical-pressure steam system was chosen for
the MSBR because the 700°F feedwater needed for the
steam generator because of the coolant-salt character-
istics can be conveniently and efficiently attained
through mixing of the supercritical-pressure steam with
high-pressure feedwater. Also, the supercritical-pressure
system affords a thermal efficiency of 44.4%, as
compared with 41.1% for a 2400-psia cycle using a
Loeffler boiler principle to attain the 700°F feedwater
temperature, Further, the capital cost of a
supercritical-pressure system for the MSBR is judged to
be about the same as, and possibly less than, the cost of
the 2400-psia system.

2.4 FUEL-SALT DRAIN SYSTEM

The MSBR drain system consists of the drain tank,
the drain line and freeze valve, a pump and jet system
to return salt to the circulation loop or to the fuel
processing plant, the off-gas heat disposal system, an
afterheat disposal system, and heater equipment which
maintains the salt above its liquidus temperature. The
drain system is housed in separate cells apart from the
reactor cell.

The drain tank serves several functions, the chief one
being a safe storage volume for the fuel salt when it is
drained from the circulation loop. A critical mass
cannot exist in the tank because of insufficient neutron

moderation, and the afterheat-removal system has
assured reliability in that it is independent of the need
for mechanical equipment, power supply, or initiating
action by the operating personnel. Cell heaters assure
that the tank and its contents remain above the salt
liquidus temperature of about 935°F.

The drain tank serves as a 2-hr holdup volume for the
highly radioactive fission product gases after they are
separated from the circulating fuel salt in the processing
system. Also, the drain tank acts as a sump for the
overflow streams from the bowls of the salt-circulation
pumps. The small stream of fuel salt which is sent to
the fuel-processing cell for removal of fission products,
protactinium, excess bred material, and impurities i8
taken from the drain tank and returned to it after
treatment and adjustment of the uranium concentra-
tion. An additional use of the drain tank is that its
storage volume, which is about 50% greater than the
fuel-salt inventory, permits accommodation of some of
the coolant salt in the unlikely event that a heat
exchanger tube failure and pressure differential reversal
permit coolant leakage into the primary system.

The fuel-salt drain tank contains a liner to absorb
gamma heat and to form an annular flow passage at the
tank wall for about 600 gpm of overflow salt from the
pump bowls. The salt stream passes along the bottom
surface of the top head and down the sides to maintain
metal temperatures within the design limits.

A well in the bottom head of the drain tank contains
five salt-actuated jet pumps. Four of the jets are
provided with salt from the primary pump discharges to
actuate the jets and return the overflow salt to the
respective circulation systems. Siphon breaks prevent
fuel salt from the pump bowl from draining back in the
event a jet stops operating. The fifth jet pump is
activated by about 100 gpm from a separate fuel-salt
pump and is used to transfer salt to the fuel-processing
cell or to fill the primary-salt circulation loop.

Afterheat released in the drain tank is removed by a
natural convection system employing an intermediate
heat transport fluid. As shown in Fig. 2.3, "LiF-BeF,
coolant salt circulates through U-tubes immersed in the
fuel salt to heat exchangers located at the base of a
natural-draft stack. There are 40 separate and indepen-
dent natural-convection circuits to afford a high degree
of reliability. The heat exchangers transfer heat from
tubes containing the transport salt to water-cooled
plates which make no physical contact with the salt
tubes. The steam generated in the plates is condensed in
finned air-cooled coils in the natural-draft stack.

An alternate drain tank cooling system using NaK as
the coolant is described in Sect. 6.4
ORNL-DWG BI-104A5AR

PRIMARY SALT PUMP

=
AR
REACTOR A 1 II |
VESSEL ("} f S
\ SR (= l% 2 TO HEAT EXCHANGER
p— f G O & |
| b
- \\ GA' ' H
HEAT X b 9
REJECT STACK o, [~pg N, e U( |
r- —T - T3 e |
’- ‘ Z%OAMY ’vaREEZE PLUG |  FrROM
, CHEMICAL —
> L EXCHANGER ‘ HEMICAL =4y
. [ He SUPPLY |
s § e '
- ® CATCH BASIN I
Lm0 T0
CHEMICAL
PROCESS
2 = |
3 41 TLIF ~BeF; '
3 3 2 |
: ' F |
] A
VENT ’*“’—‘_l,l ) TO OFF-GAS SYSTEM ! I
I i

"CHEMICAL PROCESSING
PRIMARY SALT ORAIN TANK
AND GAS HOLDUP

Fig. 2.3. Simplified flow diagram of primary drain tank and heat removal system using TLiF-BeF; salt as coolant. (A) Combiner
tank for separated gases and overflow salt, (B) Off-gas line with cooling jacket, (C) Fuel-salt drain line, (D) Drain line continuous
bleed flow, (E) Jet pumps for returning overflow fuel salt to primary system, (F) Ancillary fuel-salt transfer pump, (G) Jet pump for
filling primary system and sending salt to chemical processing, (H) Freeze-plug type valve, (1) 7LiF-BeF2-to-H20 heat exchanger, (J)

Natural draft stack, (K) Water-to-air heat exchanger.

2.5 OFF-GAS SYSTEM

The off-gases will be held in the fuel-salt drain tank
for about 2 hr, during which time a portion of the
noble metals will probably deposit on the internal
surfaces. Referring to Fig. 2.1, the gases vented from
the drain tank pass through particle traps, where
remaining particulates are removed before the gases
enter the charcoal beds for absorption and 47-hr holdup
of the xenon, permitting decay of 97% of the '*%Xe.
Most of the gas leaving the charcoal bed is compressed
for reintroduction into the salt-circulation system at the
bubble generators. A small portion of the gas leaving
the 47-hr charcoal bed enters the long-delay charcoal
bed (about 90-day xenon holdup), the outflow of
which passes through tritium and krypton traps before
entering a gas storage tank. The gas from this tank is
augmented by makeup helium if required and reintro-

duced into the circulation system as purge gas for the
circulation pumps and at other places where clean
helium is needed. The accumulated krypton and tritium
are stored in tanks in the waste cell facility.

2.6 FUEL-SALT PROCESSING SYSTEM
L. E. McNeese

Breeding with thermal neutrons is economically feasi-
ble with a molten-salt reactor because it is possible to
process the fluid fuel rapidly enough to keep the
neutron losses to protactinium and fission products to a
very low level. The equipment used to strip gaseous
fission products from the fuel salt was described in
Sects. 2.1 and 2.5. The concentrations of protactinium,
rare earths, and some other fission products are limited
by continuously processing a small stream of the fuel
salt in an on-site processing system, described below.
There are several basic processes which could be
incorporated in a molten-salt reactor “kidney.” The
effective cycle times for protactinium and fission
product removal assumed in the calculations of breed-
ing performance (Table 3.7) were based on the use of
the system described in ref. 1. Recent developments
have shown that it is possible to attain the same
breeding performance by using a somewhat different
processing plant having equipment that should be
considerably simpler to develop and operate.!' The
newer, more attractive concept is described here and in
Sect. 8.

The flowsheet for the continuous salt-processing
system is shown in Fig. 2.4. In essence, the process
consists of two parts: (1) removal of uranium and
protactinium from salt leaving the reactor and reintro-
duction of uranium into salt returning to the reactorand
(2) removal of rare-earth fission products from the salt.

A small (0.88-gpm) stream of fuel salt, taken from the
reactor drain tank, flows through a fluorinator, where

about 95% of the uranium is removed as gaseous UFg.
The salt then flows to a reductive extraction column,
where protactinium and the remaining uranium are
chemically reduced and extracted into liquid bismuth
flowing countercurrent to the salt. The reducing agent,
lithium and thorium dissolved in hismuth, is infroduced
at the top of the extraction column, The bismuth
stream leaving the column contains the extracted
uranium and protactinium as well as lithium, thorium,
and fission product zirconium. The extracted materials
are removed from the bismuth stream by contacting the
stream with an HF-H, mixture in the presence of a
waste salt which is circulated through the hydrofluorin-
ator from the protactinium decay tank. The salt stream
leaving the hydrofluorinator, which contains UF; and
PaF,, passes through a fluorinator, where about 95% of
the uranium is removed. The resulting salt stream then
flows through a tank having a volume of about 130 ft®,
where most of the protactinium is held and where most
of the protactinium decay heat is removed, Uranium

ORNL=DWG 70-2B1IRA

FROCESSED SALT

SALT UFg
PURIFICATION REDUCTION

r—
|
} He EXTRACTOR |
|
el
UFg SALT CONTAINING RARE EARTHS
| —— e S
| 1 —Lfi 9 1
l |
| |
EXTRACTOR |
e UFg |
l REACTOR I COLLECTION :
] EXTRACTOR —
i Ot p—=Bi=Li
! (0.5 MOLE FRACTION Li)
IUF6
: e EXTRACTOR

FLUORINATOR

i—{ ‘[F‘J
HYDROFLUORINATOR FLUOR!NATORH Po DECAY l l—-og. I;ILJALENT RARE

1 - ] ! [ s

HF F ~ |==——-Bi-Li

’ Pa DECAY (0.05 MOLE FRACTION
Fz—." b | F
: FLUORINATOR & ! Li)
Bll EXTRACTOR
Bi-Li

: SAL‘!’ TO b EAE?}\@LENT RARE
4 REDUCTANT | WASTE i (4 F s ]

Fig. 2.4. Flowsheet for continuous salt-processing system for a single-fluid MSBR by fluorination-reductive extraction and the

metal-transfer process.
produced in the tank by protactinium decay is removed
by circulation of the salt through a fluorinator.

Materials that do not form volatile fluorides during
fluorination will also accumulate in the decay tank;
these include fission product zirconium and corrosion
product nickel. These materials are subsequently re-
moved from the tank by periodic discard of salt at a
rate equivalent to about 0.1 ft®/day.

In summary, in the protactinium isolation system, all
the uranium that leaves the reactor, plus that produced
by decay of the protactinium, appears as UF¢ , whereas
the effluent salt from the extraction column carries
fission products but no uranium or protactinium.

The rare earths are removed from the salt stream
leaving the top of the protactinium extractor by
contacting it with a stream of bismuth that is practi-
cally saturated with thorium metal. This bismuth
stream, with the extracted rare earths, is contacted with
an ‘“‘acceptor salt,” lithium chloride. Because the
distribution coefficient (metal/salt) is several orders of
magnitude higher for thorium than for the rare earths, a
large fraction of the rare earths transfer to the LiCl in

this contactor, while the thorium remains with the
bismuth. Finally, the rare earths are removed from the
recirculating LiCl by contacting it with bismuth streams
containing high concentrations of lithium (5 and 50
mole %). These materials, containing the rare earths, are
removed from the process.

The fully processed salt, on its way back to the
reactor, has uranium added at the rate required to
maintain or adjust the uranium concentration in the
reactor (and hence the reactivity) as desired. This is
done by contacting the salt with UF4 and hydrogen to
produce UF, in the salt and HF gas.

2.7 AUXILIARY AND OTHER SUPPORT SYSTEMS

In addition to the principal systems previously de-
scribed, the molten-salt reactor complex requires an
emergency power system, cell heating systems, coolant-
salt storage tanks, and a maintenance and graphite-
replacement facility. The steam-power system will
require an oil- or gas-fired boiler for preheating the
feedwater and the turbine equipment during startup.
3. Reactor Primary System

3.1 GENERAL DESCRIPTION

3.1.1 Design Objectives

The MSBR conceptual design study was concerned
with exploring and delineating design problems and
with evolving a design which would establish the
feasibility of the concept.

The basic objective was to provide the fissile concen-
tration and geometry of graphite and fuel salt to obtain
a nuclear heat release of about 2250 MW(t) at condi-
tions affording the best utilization of the nation’s fuel
resources at lowest power cost. A good indicator of the
performance of a breeder reactor is the total quantity
of uranium ore that must be mined to fuel the industry
before it becomes self-sustaining. An index of good
performance in a growing reactor industry is GP?,
where G is the breeding gain and P is the specific power
in megawatts of thermal power per kilogram of fission-
able material. This term, the so-called conservation
coefficient, was used in nuclear physics optimization
studies to determine the dimensions of the reactor core
and reflector and the salt-to-graphite ratios, as discussed
in more detail in Sect. 3.3.2. (In general, the conditions
for the highest value of the fuel conservation coefficient
also corresponded with the lowest fuel-cycle cost and
lowest overall cost to produce power.)

Neutron fluences and maximum graphite tempera-
tures were kept low enough to provide an estimated
core graphite life of about four years. The salt flow
through the core passages was designed for each stream
to have about the same 250°F temperature rise, with
the pressure drop due to flow being kept within the
head capabilities of a singlestage circulation pump.
Cooling was provided for the reactor vessel and other
metal parts to keep the temperatures within the
tolerances imposed by stress considerations. The design
aspects, that the coefficient of thermal expansion of
Hastelloy N is about three times that of the core
graphite, that the graphite experiences dimensional
changes with irradiation, and that the graphite has
considerable buoyancy in the fuel salt, were all

10

accommodated in such a manner as to maintain the
core interpals in a compact array without significant
changes in the fuel-tographite ratios and salt velocities,
and to prevent vibrations, The salt will be maintained
well above its liquidus temperature of 930°F, and the
salt flow is upward through the core lo promote natural
circulation. The reactor is capable of being drained
essentially free of salt, and afterheat following shut-
down can be safely dissipated. The reactor vessel and
the reflector graphite are expected to last the life of the
plant, but the core was designed to facilitate periodic
replacement of the entire assembly.

There are, of course, many possible arrangements for
a molten=salt breeder reactor and power station. The
concept described here represents one design that
appeared feasible; more detailed study and optimization
would probably produce a better arrangement.

3.1.2 General Description and Design Considerations
E.S. Bettis

The principal design data are summarized in Table
S.1, and more detailed reactor data are given in Table
3.1. The detailed nuclear physics data are listed in Table
3.7. Overall plan and elevation views of the reactor are
shown in Figs. 3.1 and 3.2. The Hastelloy N vessel
material and the moderator and reflector graphite are
described in Sects. 3.2.3 and 3.2.4.

The reactor vessel is about 22 ft in diameter and 20 ft
high and is designed for 75 psig. It has 2-in.-thick walls
and 3-in.-thick dished heads at the top and bottom. Salt
at about 1050°F enters the central manifold at the
bottom through four 16-in.-diam nozzles and flows
through the lower plenum and upward through the
passages in the graphite to exit at the top at about
1300°F through four equally spaced nozzles which
connect to the 20-in,diam salt-suction lines leading to
the circulation pumps. The 6-in.-diam fuelsalt drain
line connects to the bottom of the reactor vessel inlet
manifold.

Since graphite experiences dimensional changes with
neutron irradiation, the reactor core musi be designed
Table 3,1, Principal reactor design data

Reactor vessel ingide diameter, ft 22.2
Vessel height at center” ft 20
Vessel wall thickness, in. 2
Vessel head thickness, in. 3
Vessel design pressure, psig 75
Number of core elements 1412
Length of zone I portion of core elements, ft 13
Overall length of core elements (approx), ft 15
Distance across flats, zone {,2 ft 14
Outside diameter of undermoderated region, 16.8
zone 11, ft
Overall height of zone I plus zone II .b ft 18
Radial distance between reflector and core, 2
zone 11,7 in,

Radial thickness of reflector, in, 30
Average thickness of axial reflectors (approx), in. 22
Volume fraction salt in zone 12 0,13
Volume fraction salt in zone 11?7 0.37
Core power density, kW/liter

Average 22,2

Peak 70.4
Core fuel-salt power density, kW/liter

Average 74

Peak 492
Core graphite power density, kW/liter

Average 2.3

Peak 6.3
Core thermal neutron flux, neutrons cm ™2 sec™!

Average 2.6 X 1014

Peak 8.3x 10'4
Maximum graphite damage flux (>50 keV), 3.5 % 10'4

neutrons cm 2 sec ™}
Graphite temperature at maximum graphite damage
flux region, °F

1307

Estimated useful life of graphite, years 4
Total weight of graphite in reactor,© Ib 669,000
Weight of removable core assembly @ Ib 600,000
Maximum flow velocity in core, fps 8.5
Pressure drop due to salt flow in core, psi 18
Volume of fuel salt, ft3
Total in core (seec Table 3,2) 1074
Total in primary system 1720
Fissile-fuel inventory of reactor plant and fuel 1470
processing plant, kg
Thorium inventory, kg 68,000
Breeding ratio 1.06
Yield,® %/year 3.3
Doubling time,® compounded continuously, years 21

“Does not include upper extension cylinder.

BSee Table 3.3 for definition.

:Doc.s not include 60,000 Ib in alternate head assembly.
Hoist load to be lifted into transport cask.

€At 80% plant factor.

for periodic replacement. The design chosen for the
reference MSBR has an average core power density of
22.2 W/cc, which, based on the irradiation behavior of
materials presently available, indicates a useful core

11

graphite life of about four years. [t was decided to re-
move and install the core graphite as an assembly rather
than by individual pieces, since it appeared that this
method could be performed quickly and with less likeli-
hood of escape of radioactivity. Handling the core as an
assembly also permits the replacement core to be care-
fully preassembled and tested under shop conditions.
(Maintenance procedures are described in Sect, 12.)

The reflector graphite will normally last the 30-year
life of the plant, The radial reflector pieces are installed
inside the vessel with no special provisions made for
replacement, The bottom axial reflector will be re-
placed each time a new core is installed, since this is a
more convenient design arrangement. The top axial
reflector is attached to the removable top head, but
since two heads are provided, which will be alternated
each time the core is replaced, this graphite should last
the life of the plant without replacement.

The reactor has a central zone in which 13% of the
volume is fuel salt, an outer, undermoderated region
having 37% salt, and a reflector region containing about
| % salt. There is a 2-in.-wide annulus which is 100% salt
between the removable core and the reflector blocks to
provide clearance when removing and inserting a core
assembly. The volumes and weights of salt and graphite
in the various portions of the reactor are summarized in
Table 3.2. For convenience, a terminology for reactor
zones and regions was established, as shown in Table
3.3, and these designations will be used in the descrip-
tions to follow.

The central portion, zones I-A and I-B, is made up of
4-in. X 4-in. X 13-ft-long graphite elements, as indicated
in Figs. 3.1 and 3.2 and shown in more detail in Figs.
3.3, 3.4, and 3.27. The elements will be manufactured
by an extrusion process and will require only relatively
minor machining. After fabrication, the pieces may be
treated with a sealing process to increase the resistance
to gas permeation, as discussed in Sect. 3.2.3. Holes
through the centers and ridges on the sides of the
graphite elemenlts sepurate the pieces, furnish flow
passages, and provide the requisite salt-to-graphite
ratios. The interstitial flow passages have hydraulic
diameters approximately equal to the central hole. A
more detailed discussion of the thermal and hydraulic
considerations in design of the elements is given in Sect.
34.

The fission energy release in the reactor is highest at
the center of the core, with the power density (in
kilowatts per liter) falling off approximately as a cosine
function of the core radius. By varying the salt velocity
from 8 fps at the center to about 2 fps near the
periphery, a uniform temperature rise of 250°F is
GRAPHITE
REFLECTOR~__

REACTOR
VESSEL~—

———22ft-6% In.OD ————————— ~

r‘_-— - — . —
SALT TO PUMP
(4 TOTAL) R
N i
*V b .~"- -~
' P
\\ r“?‘ — ' L 24

ORNL- DWG 69-6803R

—

(LlFTlNG ROD

PORTS (8) O

-CONTROL RODS

O

Fig. 3.1. Plan view of MSBR vessel.

obtained. The salt velocities are determined by the hole
size, by the flow passage dimensions between the
graphite elements, and by orificing at the ends of the
flow channels. An element hole size of 0.6 in. ID is used
in the most active portion of the core, and a
1.347-in.-ID hole is used in the outer portion. In the
latter case the size of the interstitial passages is reduced
to maintain the desired 13% salt volume. The 0.6-in.
hole size was selected for the inner region, zone I-A,
primarily on the basis that a smaller opening would

present significantly more difficulty in sealing the
graphite during manufacture of the elements. The ends
of the graphite elements are machined to a cylindrical
shape for about 10 in. on each end to provide the
undermoderated 37% salt region at the top and bottom
of the reactor. The top of each element is also
machined, as shown in Fig. 3.4, to provide a 3-in.-deep
outlet plenum at the top of the core to direct the salt
flow to the four exit nozzles of the vessel. Under the
effects of buoyancy and drag forces, the 1%-in.-OD
13

ORNL-DWG 69-6003R

8ft Oinn——————o

REACTOR
COVER

NN
L

REACTOR
VESSEL—"~

I —————— -

CONTROL RODS —

GRAPHITE
REFLECTOR——_

SA:.T TO PU)MP s

4 TOTAL)— R
NN

/ i\.'\*‘

— J RN
-

5

GRAPHITE -
REFLECTOR ———=___

MOBERATOR 32 ft 10in.
ELEMENTS ———

REFLECTOR
RETAINING RINGS

N\
N

) "

\

AR
o o
NN
b

N

\

N
SN

N
A AN
AN N

QBN

CLEARANCE
SPACE

\

~

] L \ N \
N NNNNN

ek
| —
| i |
N0
N
\\ f)
.‘Ir, ~
NN
%‘\\\
A \
L
f
J
|
—

ZONE 1

SO 6 ft Gin,

GRAPHITE _ >
REFLECTOR-— =D

SALT FROM HEAT
EXCHANGER (4 TOTAL) —

\fi

DRAIN e

Fig. 3.2. Sectional elevation of MSBR vessel.
Table 3.2. Volumes and weights in the
MSBR core and reflector

Pero Salt  Graphite
ercent :
A volume weight
(ft%) (Ib)
Core, zone I (14 ft actagon, 13 288 221,400
13 in. high)
Lower and upper axial, zone [I, 37 94.5 18,500
9 in. thick, top and bottom
Upper plenum, 3 in. thick, 85 36.2 700
top only
Lower plenum, 2‘/2 in. thick, 100 354
bottom
Radial, zone [I, 16.8 ft. 37 282 55,000
diam X 14.5 ft high
Annulus, 17.2 ft diam X 15 ft 100 132
high (2-in. gap)
Salt inlet (lower section), 98 11
3.5 ft diam X 1.2 ft high.
Salt inlet (upper section), 50 9 900
4 ft diam X 1.2 ft high
Lower vessel coolant 100 8.2
passage, 75 in,
Radial vessel coolant plenum 62.5 465 3,400
Radial vessel coolant, l/.;-in. gap 100 21.2
Radial reflector, 17.2 ft 1.2 26.9 254,400
high X 22.2 ft OD
Axial reflector, bottom 3 14.7 54,800
Axial reflector, top 4 14.7 54,800
Control rod entrance thimble 2.9
Outlet passage 42.1 5400
Annulus between upper head 100 8.7
flange extension and
vessel, 1/2 -in. gap
Total 1074 669,000

neck of each prism is pressed firmly against the top
reflector blocks. When the reactor is empty of salt the
graphite rests on the Hastelloy N support plate at the
bottom of the vessel.

Four 6- by 6-in. graphite elements with a 4-in.-diam
hole are shown installed axially at the center line of the
reactor in Figs. 3.1 and 3.2. More rods may be required,
however, as discussed in Sect. 10.2, Two or more of the
holes receive relatively simple graphite control rods
which, on insertion, increase the reactivity by displacing
some of the fuel salt. Since these rods have a pro-
nounced tendency to float in the salt, they are
self-ejecting with respect to decreasing the reactivity,
even if the graphite should fracture. The other two
holes are for neutron-absorbing rods used only for
reactor shutdown. These 6- by 6-in. elements are
retained at the bottom by fitting them into a Hastelloy
N enclosure in the bottom of the bottom-head salt-
distribution assembly. Since the elements are restrained

14

in position, they serve as a base around which the core
elements can be stacked when the core is assembled. A
jig is used to hold the elements until the entire core is
assembled and the restraining rings are in place.

The undermoderated zone with 37% salt, or radial
“blanket,” surrounding the miore active puriion serves
to reduce neutron leakage from the core. This zone is
made up of two kinds of elements: 4-in. X 4-in. X 13-
ft-long elements like those in the core except fora larger
hole size (2.581 in. ID) (Fig. 3.5), and 2-in.-thick X
13-ft-long slats arranged radially around the core, as
shown in Fig. 3.1. The slats average about 10.5 in. in
width, the dimension varying to transform the generally
octagonal cross-sectional shape of the core element
array into a circular one, The slats also provide stiffness
to hold the inner core elements in a compact array as
dimensional changes occur in the graphite, Dowel pins
separate the slats to provide flow passages, and vertical
elliptical graphite sealing pins at the outer periphery of
the array isolate, to a large extent, the salt flowing
through the core from that flowing through the
reflector region. The slabs are separated from each
other by graphite buttons located at approximately
18-in. intervals along the length. Each slab has a groove
running axially about 1% in. from the outside edge to
accommodate the long ellipticalshaped graphite dowels
which are inserted between adjacent slabs to isolate the
slab salt flow from the flow in the previously men-
tioned 2-in. annulus. There are similar elliptical-shaped
dowels running axially between the prisms of the outer
row of the core to perform the same function, in that
they isolate the flow in zone I from that in zone II.

There are eight graphite slabs with a width of 6 in. in
zone II, one of which is illustrated in Fig. 3.3. The holes
running through the centers are for the core lifting rods
used during the core replacement operations mentioned
above. These holes also allow a portion of the fuel salt
at essentially the reactor inlet temperature of 1050°F
to flow to the top of the vessel for cooling the top head
and axial reflector.

Figure 3.3 also shows the previously mentioned
2-in.-wide annular space between the removable core
graphite in zone II-B and the permanently mounted
reflector graphite. This annulus, which is 100% fuel salt,
provides clearances for moving the core assembly, helps
absorb the out-of-roundness dimensions of the reactor
vessel, and serves to reduce the damage flux arriving at
the surface of the graphite reflector blocks.

Since the reflector graphite is in a position of lower
neutron flux, it does not have to be sealed to reduce
xenon penetration. Also, because of the lower neutron
dose level, it does not have to be designed for
15

Table 3,3 Terminology used to designate regions and zones of reactor

Term used

Region or zone?

Core

Zone 12 (zone I-A, zone [-B, etc.)
Zone 1I€ (zone 11-A, zone 11-B, etc.)
Annulus

Lower plenum
Upper plenum
Radial reflector

Upper axial reflector
Lower axial reflector
Radial vessel coolant passage
Upper vessel coolant passage

Lower vessel coolant passage
Inlet

Outlet passage

This includes the 13, 37, 85, and 100% salt
regions out to the inner face of the
reflector but does not include the reflector

~13% salt region of core
~37% salt region of core

~100% salt annular region of core between
zone Il and radial reflector

~100% salt region of core between zone [I
and lower axial reflector

~B5% salt region of core between zone II
and upper axial reflector

Graphite region surrounding core in radial
direction

Graphite region above core
Graphite region below core
Gap between radial reflector and vessel wall

Gap between upper axial reflector and
upper vessel head

Gap between lower axial reflector and
lower vessel head

Salt inlet passage in lower vessel head and
lower axial reflector

Salt outlet passage in upper axial reflector

9See Figs. 3.1, 3.3, and 3.26.

I-A, [-B, etc., are used to designate different element shapes in a zone.
“The terms “radial zone 11, *“‘upper axial zone 11,” and “lower axial zone II"" should

be used as necessary,

replacement during the reactor lifetime. The reflector is
comprised of molded graphite blocks which require
only minor machining operations to fabricate. The
radial reflector graphite is made up of slightly wedge-
shaped blocks to provide a reflector about 2% ft thick.
The blocks are about 10 in. wide at the vessel wall,
about 9 in. wide at the inner end, and about 43 in. high
and are assembled in four layers. Hastelloy N axial ribs,
indicated in Fig. 3.3, provide a Y -in. standoff space
from the vessel wall and also align the reflector blocks
together in the vertical direction. Fuel salt from the
reactor inlet plenum flows upward through this vertical
space to cool the vessel wall and the outer portion of
the reflector graphite,

In addition to the axial flow of salt for cooling the
radial reflector graphite, an inward flow of fuel salt is
maintained by 1-in.-OD graphite pins, or dowels, which
are inserted in the reflector pieces to hold them apart.
The salt flow passages are about 0.05 in. wide in the
cold condition and widen to about 0.l in. at operating
temperature. Slotted Hastelloy N orifice plates are set

into the reflector graphite at the outer wall to distribute
the radial flow of salt between the top and bottom
passages to provide more uniform cooling in the
reflector. About 1% of the reflector volume is fuel salt.
All the radial flow channels slope downward toward the
vessel wall to allow the salt to drain when the system is
emptied.

Since graphite has about one-third the thermal coef-
ficient of expansion of Hastelloy N, the clearances
between blocks will tend to increase as the system is
heated to operating temperature. Even distribution of
these clearances is maintained by restraining lateral
shifting of the graphite. Each reflector block in the
bottom layer of graphite has a shallow radial groove
milled for about 18 in. in the bottom center. These
grooves fit over radial webs welded to the bed plate on
which the reflector blocks are stacked. The webs
maintain each block at a given position relative to the
metal bed plate as the plate expands. The upper layers
of radial reflector blocks are forced to maintain registry
with the bottom keyed block by the previously
AXIAL RIBS

S ~ _—-—"_‘u

s Y

2. ¢ 4

RADIAL
REFLECTOR

7%,

72
%%,

77

ANNULUS

7772%

ZONE 11-B~

72
77

7%
%
A \?/
QA

(7,

<<
/f .
//2’//,//) /

G 4,

.
$

» .

ZONE II-A -~

ZONES (-4
AND =@

SEE

{7
o .
%

ORNL—-DWG 69—-5955A

N

07
2%

7

7

W
X

RS

N

2

&
¥ 7
p

7 I/’

Ge%45 27

N

"

-

Fig. 3.3. Detailed plan view of graphite reflector and moderator elements.

mentioned Hastelloy N axial ribs, which also provide
the cooling gap between the blocks and the vessel wall.
The radial reflector blocks are pushed outward against
the spacer ribs as the vessel expands by Hastelloy N
hoops inserted in circumferential slots at each layer of
blocks in the reflector as indicated in Fig. 3.2, The rings
or hoops expand at the same rate as the vessel and keep
the reflector blocks pushed outward to follow the vessel
wall.

Since the radial blocks on the top layer are wedge
shaped and there is not room to lower the last block
into place from above, two of the top-ayer blocks are
split wedges which form a rectangular space into which
a block can be moved laterally to complete the
assembly. After all reflector pieces have been put in
place, a segmented metal retainer plate is put on top of
the top layer and bolted to gussets which are attached
to the overhanging vessel wall. This retainer plate
prevents the reflector from floating when the reactor is
filled with salt.

The axial reflectors at top and bottom are made up of
wedge-shaped pieces of graphite, the inner end being
about 2 in. wide and the end at the outer circumference
being about 16 in. wide, In addition, because of the
dished heads on the vessel, the wedge-shaped pieces are
about 30 in. thick at the center and about 15 in. thick at
the outer edge. The top head of the vessel (and its
alternate) contains a permanently installed axial reflec-
tor assembly supported in the manner indicated in Fig.
3.2. The lower axial reflector graphite is renewed with
each core, since it forms the base upon which a new
core is assembled. A support structure around the
bottom inlet supports the bottom graphite, and the
axial reflector assembly is prevented from floating in
the fuel salt by the weight of the 3-in.-thick Hastelloy N
inner head (core support plate) to which it is attached.

A flow of fuel salt is provided for cooling the axial
reflectors in much the same manner as for the radial
reflector graphite. Salt for the lower reflector taken
from the reactor inlet flow is used to cool both the
17

s j P, I

Fig. 3.4. Graphite moderator elements for zone 1.
18

13/4in.

e

m
. A
M —_—_
=
LI I IRy T umt

U6 Wb

Ulp Wb

2

A -

Fig. 3.5, Graphite moderator elements for zone II.
lower head of the vessel, the inner head (core support
plate), and the axial reflector graphite, The inner head
is provided with standoffs to permit salt to flow
between it and the bottom head of the vessel. Holes
through the inner head allow some salt to flow upward
through passages between the bottom axial reflector
pieces. The lower passage between the bottom heads
also supplies the salt which flows upward at the wall to
cool the vessel and the radial reflector graphite. Fuel
salt for cooling the top head and upper axial reflector
flows upward through the control rod region at the
center of the core and through the core lifting rod holes
in zone 1I, as shown in Fig. 3.1 and described below.
This salt is initially at near the inlet salt temperature of
about 1050°F, and, after absorbing the heat in the
upper head and graphite, it leaves the reactor with the

exit salt flow,

The top head of the reactor vessel is flanged to
facilitate access to the core. The flange is located several
feet above the top dished head for better accessibility
and a lower radiation and temperature environment.
For the core removal and replacement operations, the
core is temporarily attached to the top head and axial
reflector and the entire assembly moved as a unit. To
accomplish this, eight seal-welded flanged openings in
the top head of the reactor vessel give access to vertical
holes in the graphite core structure for insertion of
2% -in.-diam molybdenum lifting rods which attach by
a ball latch to the forged support ring at the bottom of
the reactor core, as shown in Figs. 3.6 and 3.7.
Molybdenum was selected for the rod material because
of its strength at elevated temperatures, it being
anticipated that the core temperature would increase
above its average 1100°F operating temperature during
the transfer operation. The ball latch mechanism is
activated from above by a push rod running inside the
length of the lifting rods. An enlarged section at the top
of each rod engages the top head to clamp the core and
head together. The rods are used to lift the entire core
assembly into the transport cask, in which it is then
moved to the storage cell for eventual core disassembly
and discard into the waste cell. The core assembly is
about 16 ft in diameter and weighs about 240 tons.

The reactor vessel is supported from the top by an
extension of the outer wall which carries a large flange
at the top (see Figs. 3.2 and 12.2) that rests on the
reinforced concrete roof structure. This cylindrical
piece extends about 15 ft above the top of the reactor
vessel and has walls 2 in. thick. The top head of the
reactor vessel also carries a cylindrical extension with a
flange at the top to mate with the vessel flange. The
flanged joint is thus located outside the high-

ORNL~DWG 70-11908

N :
' 4
— o
. e
R e N
NN o
"N\ RS
0
<
ataatatat
»
— TYPICAL-B PLACES
2in, r‘
GRAPHITE RADIAL
REFLECTOR
) —— GRAPHITE AXIAL
s REFLECTOR
i
o
. 'E % et |
& - o
e |
l'fl
f & S
. ‘l

Fig. 3.6, Core-lifting rod access holes.

temperature region of the reactor cell, is elevated above
the maximum salt level, and is not subjected to high
temperature gradients or strong irradiation levels. Dou-
ble metal gaskets with a leak detection system are used
in the joint. The flanges are held together by clamps,
with the bolting readily accessible from the operating
floor level. It may be noted that with this arrangement
the weight of the roof plugs augments the bolting in
clamping the flanges together.
OURNL-DWG 70-11509

LpEmmm——— kb b
|

|
l‘\ TOP HEAD

ACCESS PORT

LIFTING ROD

ACTUATING ROD

s
N N ’
o]
N N
N
O NN | S,
e o
N“ \ w .‘\ - ..<
E: N \ e .n’:‘
) L .\ o
N
< e
BALL LATCH
: o
: i

e~ CORE SUPPORT
FLANGE

Fig. 3.7. Reactor core-lifting rod assembly.

As previously mentioned, a second reactor vessel top
head and its cylindrical extension piece will be required
in order to assemble a new reactor core prior to the
core replacement. After each use and a suitable decay
time, the top head will be reclaimed for the next core
replacement operation. The new core will be assembled
under shop conditions in a “clean room” located
outside the MSBR containment. The core will be

20

erected on a new Hastelloy N support plate which has
been provided with new graphite lower axial reflector
blocks. When all the elements are in place in the
octagonal array, a segmented graphite band is installed
around the top head and bottom to hold them in place,
a indicated n Fig. 3.Z. After assembiy of the core 1s
complete, it is moved through the gas lock into the
containment. The reactor top head and top axial
reflector assembly, which has been cleaned and in-
spected after previous use, is now attached to the new
core with the previously mentioned tie rods. After the
spent reactor core is removed from the vessel, as
described above, the replacement assembly can be
lowered into place, the tie rods removed, and the rod
access port and top head flanges sealed. Maintenance
operations are described in more detail in Sect. 13.

3.2 SPECIAL MATERIALS

The fuel and coolant salts, the reactor graphite, and
the modified Hastelloy N are special MSBR materials
which have been studied and developed at ORNL in a
program that started over 15 years ago. The background
information and documentation supporting this area of
the MSBR design study are far too extensive to be
reviewed here. In general, each of the materials has been
investigated sufficiently to give confidence that their
use, within the limits prescribed, is feasible and prac-
tical in the MSBR. Selected physical properties of the
four materials are listed in Table S.1, and some general
characteristics, as specifically related to the MSBR
design study, are briefly discussed below,

3.2.1 Fuel Salt

The fuel salt selected for use in the MSBR is
TLiF-BeF,-ThF,-UF; (71.7-16-120.3 mole %). The
lithium is enriched to 99.995% "Li. In brief, the fuel
salt melts at about 930°F and has a low vapor pressure
at operating temperatures. It has low thermal-neutron
capture cross sections and is stable throughout the
proposed range of application."® With a viscosity about
twice that of kerosene, a volumetric heat capacity
about the same as that of water, and a thermal
conductivity more than twice that of water, it has
adequate heat transfer characteristics” and a reasonable
pressure drop due to flow.'? It is compatible with the
materials in the system,’?

In selecting a fuel salt for the MSBR it was recognized
that the fuel salt must consist of elements having low
capture cross sections for neutrons typical of the
chosen energy spectrum. The fuel must dissolve more
than the critical concentrations of fissionable material
(#?5U0, 233U, or 2*?Pu¥) and high concentrations of
fertile material (*?*?Th) at temperatures well below
1050°F. The mixture must be thermally stable, and its
vapor pressure needs to be low in the operating
temperature range of 1050 to 1300°F. It must possess
heat transfer and hydrodynamic properties adequate for
service as a heat-exchange fluid. It must be nonaggres-
sive toward the materials of construction, notably the
Hastelloy N and the graphite. The fuel salt must be
stable toward reactor irradiation, must be able to
survive fissioning of the uranium or plutonium, and
must tolerate fission product accumulation without
serious deterioration of its useful properties. It must be

*Plutonium, as 239PuF,, could be used instead of 233U or
235U for the initial fissile loading, and there may be economic
advantages to doing so for the nuclear startup and shakedown
runs on an MSBR station. (The molten-salt reactor could not
breed on the 238U.239Py cycle, however, because of pluto-
nium’s low value of n for thermal neutrons.)

21

capable of being processed for turnaround of unburned
fissile material, effective recovery of bred fissile ma-
terial, and removal of fission product poisons, all with
sufficient economy to assure a low fuel-cycle cost.

As discussed by Grimes,'? fluorides are the only salts
with acceptable absorption cross sections and the
requisite stability and melting temperatures. Both ura-
nium tetrafluoride (UF,;) and thorium tetrafluoride
(ThF,;) are sufficiently stable, and, fortunately, their
relatively high melting temperatures are markedly de-
pressed by use of diluent fluorides. The preferred
diluents are BeF, and 7LiF. The phase behavior of
systems based on these diluents has been examined in
detail,'* and the system LiF-BeF,-ThF4 is shown in
Fig. 3.8.

Successful operation of the MSRE lent confidence
that oxide contamination of the fuel system can be
kept to adequately low levels and that ZrF; (5 mole %),
used as a constituent of the fuel in the experimental

ORNL-LR-DWG 37420AR7

Thi, 1414
Wad TEMPERATURE IN °C
COMPOSITION IN mole %o
\0
LiF-4ThE,
LiF-Thi, LiF-2ThF,
3LiF-ThF, ss s 4 1000
LiF-2ThE,
P 897 a5
2LiF -Ber
LiF'ThF; 300
P 762 850
P 597
£ 568
BLIF-THE s
4
£ 565 s
700,
o) 850
6
f«b%"o o e

S, o 50 526

(o) \
LiF V BeF.
. 4 2
848 2LiF- BeF; 500?450 400 400 450 500 555

P 458 £ 360

Fig. 3.8. The system LiF-BeF,-ThF;.
reactor to preclude inadvertent precipitation of UQ,,
would not be needed in the MSBR.

The single-fluid MSBR requires a concentration of
ThF,; near 12 mole %, and criticality studies indicate
that the 2**UF, concentration should be about <0.3
mole % The ratio of T1LiR to BeF, should b¢ high to
keep the viscosity low. To maintain the liquidus
temperature below about 932°F (for a melt with 12%
ThF,), the BeF, concentration must be in the range of
16 to 25%. The most likely choice for the MSBR fuel
salt composition was thus "LiF-BeF,-ThF,-UF,
(71.7-16-12-0.3 mole %). This salt is undamaged by
radiation and is completely stable at operating condi-
tions.

As indicated in Table D.2 of Appendix D, the
estimated cost of the primary salt for the MSBR
reference design is about $13 per pound for the
"LiF-BeF, carrier salt; about $9 per pound for the
"LiF-BeF,-ThF, barren salt, and about $57 per pound
for the enriched fuel salt, based on a fissile material cost
of $13 per gram. The total cost of the primary salt
inventory in the MSBR reactor and chemical treatment
plant systems is thus about $23 million,

3.2.2 Coolant Salt

The MSBR uses a circulating secondary fluid to
transport heat from the fuel salt to the steam generators
and reheaters. This coolant must be stable at all
temperatures up to 1300°F, must not be damaged by
radiation (including the delayed-neutron emissions in
the primary heat exchangers), must be compatible with
other materials, must have acceptable heat transfer and
hydraulic properties, and, because of the relatively large
volume required, must be reasonable in cost. The
coolant selected for the reference design is a eutectic
sodium fluoroborate salt having the composition
NaBF;-NaF (92-8 mole %). Pertinent physical prop-
erties are listed in Table S.1.

The NaBF,-NaF system is shown in Fig. 3.9. The
eutectic has a vapor pressure at 1150°F of about 252
mm Hg and could operate with a dilute mixture of BF;
in helium as the cover gas. It has a melting temperature
of about 725°F and has a viscosity, volumetric heat
capacity, and thermal conductivity properties close to
those of water. The salt mixture is stable in the system
environment, If the sodium fluoroborate is free of
contaminants and water, test loop experience indicates
that the corrosion rate of Hastelloy N at the reactor
system conditions will probably be less than 0.2
mil/year, Commercial grades may have acceptable
purity and would have a modest cost of less than 50
cents/lb.

22

ORNL WG GT 9423AR

1000 ¢ W =y : o §
] [ [ [ oo
995°C T ¢ SOLIDUS
300 : J\T\L— “{é CRYSTAL NVERSION 4
' ‘
-
>
800 l l i r— «\L L LIouID
§ \ NaF + LlOlJlI’) “L ; o l
o ({8 ] ‘ — —\ —
'
= | l
Iz 600 }— «p—oud —-4 4 \— :
[\ ] N BF, .
& (HIGH-TEMPERATURE FORM) \"08 C\,
= 500 g +LIQUID =t}
av0 |38avc l 1 - E .
| ‘ T ];
A 1 NnF +NaBF (HIGH « 1EMP[RMURE FORM) |
2a3¢__ | e S QN
200 1 t‘oF_*_Nqfl_Fa (LOW-TEMPERATURE FORMI | =t
NaF 20 40 80 20 NoBF 4
NaBF, (mole %)

Fig. 3.9. MSBR coolant salt — The system NaF-NaBF,.

The choice of sodium fluoroborate was based on a
survey of possible molten-salt reactor coolants by
McDuffie et. al.'® Consideration of a number of
coolants has been previously reported'**'? and sum-
marized by Grimes.'® The remaining uncertainties and
problems in the use of sodium fluoroborate are de-
scribed in Chap. 16, along with a discussion of
alternative coolants and the effect their use would have
on the MSBR design.

3.2.3 Reactor Graphite
W.P. Eatherly C. R. Kennedy

3.2.3.1 Introduction. Graphite is the principal ma-
terial other than salt in the core of a molten-salt
reactor. As such, its behavior under radiation damage is
of considerable significance. Prior to 1966, data on
graphite behavior at elevated temperatures and high
fluences were scattered, and there was good reason to
believe the effects of radiation damage in graphite were
self-limiting and would saturate at exposures of the
order 2 X 10%? neutrons/em? at 700°C.

During 1966—1967, British data,'® quickly con-
firmed in this country,’ 7 demonstrated that the dimen-
sional changes induced by radiation did not saturate but
eventually resulted in gross expansion of the graphite
accompanied with structural deterioration. Under the
fluences and temperatures existing within proposed
high-performance molten-salt reactors, this meant that
the graphite in the core would not last the life of the
reactor and would have to be replaced at rather
frequent intervals.
In view of this situation, two studies were imme-
diately initiated: (1) to ascertain the effect of the
graphite on reactor performance'® and (2) to estimate
the probability of improving existing graphites.'” The
general results of these studies were as follows:

1. The behavior of existing graphites can be tolerated
from the standpoint of both design and economics.

2. The cost and design penalties are significant enough
to justify search for an improved material.

3. There is a reasonable probability that better graph-
ites can be developed.

Subsequent events have justified these conclusions.

3.2.3.2 Structural and dimensional stability. There
are two overriding requirements on the graphite,
namely, that both molten salt and xenon be excluded
from the open pore volume. Any significant penetration
of the graphite by the fuel-bearing salt would generate a
local hot spot, leading to enhanced radiation damage to
the graphite and perhaps local boiling of the salt. It
would obviously also lead to uncertainties in the reactor
fuel inventory and dynamic reactor behavior. Since the
salt is nonwetting to the graphite, this requires only
that the graphite be free of gross structural defects and
that the pore structure be largely confined to diameters
less than 1u. Both requirements can be met by
currently available commercial graphites.

Xenon-135 is a serious poison to the reactor and
could cost several percent in breeding ratio if not
stripped from the salt or excluded from the graphite.
Calculations? indicate that with graphite having a gas
permeability of the order of 10™® ¢m? /sec STP helium,
a reasonably effective gas stripping system can reduce
the poisoning to a negligible level. The best com-
mercially available graphites have gas permeabilities in
the 107 to 107 em? /sec range,although experimental
materials have achieved levels of 10™° to 107% cm? /sec.
These values seem to be the achievable limit relative to
closure of pores by repeated carbonaceous impregna-
tion and graphitization of bulk graphite,

It is obvious that the structural deterioration of
graphite under radiation damage will lead to eventual
loss of impermeability and hence to a definable lifetime
of graphite in the core. In addition, the dimensional
changes will lead to changes in the core configuration
and behavior. Data available by 1968 on graphite
behavior were analyzed, and a set of curves was
established representing the expected behavior of the
graphite obtainable at that time. The resultant curves?®
for isotropic graphite are shown in Fig. 3.10.* From
these curves and the presumed temperature distribu-

ORNL -DWG 69-5261
+2 l

BS0°c  |80O0°  J780°

"
+1 )

//

—

LINEAR DISTORTION G (%)
& o

o
n
W

FLUENCE & X 1022 (E >50 kev)

Fig. 3.10, Dimensional behavior of graphite as a function of
fluence at various temperatures.

tions in the core, the changes in core configuration have
been calculated.?®+?' It is concluded that changes in
reactor performance due to strictly geometrical changes
are not significant.'®

For lack of a better definition, it has been assumed
that permeability will improve or remain unaffected
during the period of time the graphite is in a contracted
phase, and hence the point at which the graphite
returns to its original density defines its useful life. This
leads to the conclusion®? that the graphite can absorb a
fluence of 3 X 10?2 neutrons/cm? (E > 50 keV) before
deteriorating significantly or, equivalently, that it will
have to be replaced in the core about every four years
in the present design. The associated operating cost
penalty for replacing graphite is estimated to be
between 0.1 and 0.2 mill/kWhr.

*Subsequent data indicate that the temperature effects may
be less than those shown in Fig. 3.10. Graphite now under
development may also have better dimensional stability,
Recent results obtained from irradiations in the High
Flux Isotope Reactor (HFIR) at ORNL indicate that
the definition of lifetime based on return to original
density is indeed conservative for almost all graphites
examined. The observed changes in microstructure
represent an increased generation of extremely small
pores at the expense of reduction in the size of the
larger pores during the contraction phase. The excep-
tions are graphite containing a high portion of low-
density phases, a type of material which can be avoided
for MSBR applications.

Other than the common commercial technique of
repeated liquid impregnation of graphite, several other
alternatives exist: (1) metallic or carbide coating, (2)
pyrolytic carbon coating, (3) gaseous impregnation and
decomposition (pyrolytic impregnation), and (4) liguid
or solid salt impregnation. The use of metal-containing
coatings has been investigated, and successful coatings
were demonstrated. The useful metals and coating
thicknesses required lead to a significant loss of
neutrons, however, and this approach has been aband-
oned, at least temporarily. Of the pyrolytic techniques,
the impregnation approach was initially preferred over
coating because of the less fragile nature of the
impregnated surface. An apparatus has been designed??
which permits gaseous impregnation of graphite accom-
panied by pyrolytic decomposition. This leads to filling
of the pores near the surface with pyrolytic carbon and
graphite, and permeabilities of 107'% cm?/sec have
been easily achieved. Various samples of such impreg-
nated materials have been irradiated, however, and they
have withstood fluences only to about 1.5 X 10%?
neutrons/em? (£ > 50 keV).2' Such results are to be
anticipated, because both the base graphite and pyro-
lytic material undergo dimensional changes under irra-
diation. A variety of behavior of pyrolytic materials can
be obtained by altering the hydrocarbon gas used as the

24

source of carbon, the temperature of decomposition,
and partial pressures of hydrocarbon and other inert or
catalytic gases. Considerable work may be required to
define a process for a given base graphite, and such a
process may be unique to each base stock. The program
ccks both techni

Nsase

anlls.

caity

------

feasible and economically attractive,

Pyrolytic coating, on the other hand, is a much more
tractable process and requires less process control.
Coatings only 3 to S mils thick readily yield perme-
abilities in the range 107 to 107 cm? /sec. These have
survived irradiations to 2 X10?? neutrons/cm? with
negligible loss in permeability and hence look very
attractive. However, the samples must be protected
against chipping due to external mechanical stressing. It
is probable that a combination of impregnation and
coating will turn out to be the preferred technique.
Pending more results on these experiments, work has
been curtailed on studying the feasibility of liquid or
solid salt impregnation.

3.2.3.3 Thermal and mechanical properties, The ther-
mal conductivity of the graphite becomes important
only as it affects the internal temperature of the
material due to gamma and neutron heating. For the
reference design of the MSBR, this heat is quite
significant, up to 8.3 W/cm® . The temperature gradients
thus developed lead not only to thermal stresses but
also to radiation-induced stresses generated by the
temperature dependence of the damage, Values of the
relevant properties of a fine-grained isotropic graphite
have been estimated from properties of various grades
of graphite given in the literature. The estimates are
given in Table 3.4. Although some of these values, such
as the thermal conductivity, will change during irradia-
tion,.the changes will probably not seriously affect the
calculated stresses.

Table 3.4. Estimated design graphite properties of base graphite for an MSBR”

Thermal conductivity,? W em™' °C)~!
Thermal expansion, (°C) ™}

Young's modulus, psi

Ultimate tensile strength, psi
Poisson's ratio

Creep constant, psi~! neutron™ ¢m?
Anisotropy

Density, g/cm3

Permeability, cm? (STP He)/sec
Accessible void volume, %

37.63(T) %7, where T="K

552x 1078 +1.0x 107°T

1.9x% 10°

5000

0.27

(53— 145% 10727+ 14 % 10757%) x 10727
<0.05%

~1.9

<IX 1072

<10

®All temperatures expressed as degrees centigrade, except as noted,
Unirradiated; radiation may decrease conductivity.
Constitutive equations for the graphite prisms have
been set up and solved??+2! to obtain the thermal and
radiation-induced stresses. The equations include the elas-
tic response as well as primary and secondary creep,??
the important contribution being that of secondary
creep. Despite the fact that the radiation-induced
strains far exceed the maximum tensile strain of
graphite at failure, the relaxation due to creep largely
keeps pace with these strains and reduces the induced
stresses 10 relatively low levels. A curve of stress vs time
for MSBR graphite is shown in Fig. 3.11, The initial
thermal stresses anneal out in a few weeks’ time, and
there is a gradual buildup of the radiation-induced
stresses. In no case do the stresses exceed 600 psi,
which is quite acceptable in view of the anticipated
5000-psi ultimate tensile strength.

It is concluded, therefore, that the induced stresses in
the graphite do not constrain the reference MSBR
design or performance,

3.2.3.4 Improved graphites. Before considering the
probability of improvement of the graphite, it is
advantageous to review briefly the mechanism of
damage. On the average, each fission neutron will
produce 500 to 1000 interstitial-vacancy pairs in the

ORNL—DWG 69— 5538A

——

500 f——r

400 |t

SURFACE STRESS o OR a; (psi)

-60
0 { 2 3 4
TIME AT 80% PLANT FACTOR (yr)

Fig. 3,11, Maximum stress produced in MSBR reactor core
graphite as a function of time.

graphite. At the temperatures under consideration
in the MSBR, the interstitials are highly mobile,
and the vacancies are slightly mobile. Although some
direct recombination of vacancies and interstitials
does occur, enough survive to generate both inter-
stitial and wvacancy aggregates. The vacancy aggre-
gates collapse to lead 1o a shrinkage of the crystallites in
the g-axis direction, whereas the interstitial aggregates
grow into new planes leading to crystallite growth in
the c-axis direction. Rather fortuitously, the net growth
leads to virtually no change in crystallite volume,
although the shape change is marked.

There is general agreement on the above qualitative
explanation, but detailed attempts to quantify the
model have not led to satisfactory results. The British
work?? demonstrated (at low temperatures) a relation-
ship between radiation damage and thermal expansion.
Later work at General Atomic®>® on pyrolytics has
demonstrated an effect of crystallite size and density.
More recent work at ORNL has shown even more
complex effects, presumably due to intracrystallite
plastic deformation and micropore structure. It is
apparent that existing models of radiation damage are
still incomplete and do not imply that the behavior of
existing graphite represents an ultimate behavior of the
graphite in general.

Subsequent to the analysis of dimensional behavior of
materials represented by Fig. 3.10, type AXQ graphite*®
has been studied in the HFIR at ORNL." This graphite
demonstrates some improvement in behavior, as shown
in Fig. 3.12, and satisfies many of the desired require-
ments, but it is a molded material and cannot be
fabricated to the required shapes. At present, both
ORNL and the vendor are studying methods to form
the material.

3.2.3.5 Conclusions. On the basis of the survey of the
capabilities of the graphite industry, coupled with
current programs on radiation damage and fabrication,
the following conclusions have been made.

1. Current state-of-the-art materials are adequate to
produce base graphites meeting the technical require-
ments for an MSBR. These graphites will have a core
lifetime at the reference MSBR flux levels of the order
of four years, which introduces a cost penalty of 0.1 to
0.2 mill/kWhr.

2. Early studies of gaseous impregnation have demon-
strated the capability of meeting the permeability of
<107® cefsec that would be desirable to help minimize
the '3%Xe neutron absorption. It remains 1o be

*Supplied by Poco Graphite, Inc., a subsidiary of Union Oil
Company of California, Decatur, Tex.
ORNL~DWG 7014910

e Y | S i — oo

S /' w GRAPHITE /

=2

(o)

g ;‘

7]

a

&

W REFERENCE

5 .l | e
2

1 2
FLUENCE [10%2 vt (£> 50 keV) |

Fig. 3.12. Behavior of type AXQ graphite at 715°C con-
trasted to the presumed behavior of the reference graphite nsed
in MSBR design calculations.

demonstrated that such impregnated materials will
satisfactorily withstand radiation damage.

3. Geometrical restrictions introduced by require-
ments of fabricability do not restrict reactor per-
formance.

4. Sufficient data now exist to imply that improved
graphites for MSBR usage can be developed. However,
these improvements will most probably be incremental
relative to the best graphites tested to date.

3.2.4 Hastelloy N
H. E. McCoy

In this reference design of the MSBR, the material
that is specified for nearly all of the metal surfaces
contacting the fuel and coolant salts is an alloy which is
a slight modification of the present commercial Hastel-
loy N. (The only exceptions are parts of the chemical
processing system, which are made of molybdenum,
and the infrequently used fuel storage tank, which is of
stainless steel.) As described below, the modified
Hastelloy N anticipated in the MSBR design is currently
in an advanced stage of development. It is very similar
in composition and most physical properties to stan-
dard Hastelloy N, which has been fully developed and
approved for ASME Code construction and was used

26

successfully in the MSRE. The modified alloy is
superior to standard Hastelloy N, however, in that it
suffers much less loss of ductility under neutron
irradiation. The design of the MSBR reactor vessel
counts on this improvement, and throughout the
description of the design in this report “Hastelloy N
means the modified alloy unless otherwise stated. The
consequences of failure to commercially produce an
approved alloy with the desired properties are discussed
in Sect. 16.2.3.

3.24.1 Primary system. The metal in the reactor
vessel and in the primary piping will be exposed to
molten fuel salt at temperatures up to 1300°F on one
side and to the cell atmosphere (95% N,—5% O,) at
1000°F on the other. The anticipated service life is 30
years, during which time the most highly irradiated
portions of the reactor vessel will be exposed to a
fast-neutron (£ > 0.1 MeV) fluence of less than 1 X
102" neutrons/cm? and a thermal-neutron fluence of
about 5§ X 10%? neutrons/cm?.

Hastelloy N is an alloy developed specifically for use
in molten fluoride systems,?® with the composition
shown in Table 3.5. Among the major constituents,
chromium is the least resistant to attack by the
fluorides. The chromium content of Hastelloy N is low
enough for the alloy to have excellent corrosion
resistance toward the salts. (The leaching of chromium
is limited by the rate at which it can diffuse to the
surface.) The chromium is high enough, on the other

Table 3.5, Chemical composition of modified Hastelloy N

Concentration (wt %)%

Element
Standard alloy Modified alloy

Nickel Balance Balance
Molybdenum 15.0-18.0 11.0-13.0
Chromium 6.0-8.0 6.0-8.0
fron 5.0 5,0
Carbon 0.04—-0.08 0.04-0.08
Manganese 1.0 0.2
Silicon 1.0 0.1
Tungsten 0.5 0.1
Aluminum 0.1
Titanium 0.5 2.0
Copper 0.35 0.1
Cobalt 0,20 0.2
Phosphorus 0.015 0.015
Sulfur 0.020 0.015
Boron 0.010 0.0010
Others, total 0,50
Hafnium 1.0
Niobium 2.0

aSingle values are maximum concentrations.
hand, to impart good oxidation resistance toward the
cell atmosphere, The molybdenum content was ad-
justed to give good strength without an embrittling
second-phase formation. The resulting alloy has very
good physical and mechanical properties.?72?

Standard Hastelloy N was approved by the ASME
Boiler and Unfired Pressure Vessel Code Committee
under Case 1315-3 (ref. 30) and Case 1345-1 (ref, 31)
for nuclear vessel construction and was the primary
structural material in the MSRE. In the fuel system of
this reactor, Hastelloy N was exposed to salt at about
1200°F for 22,000 hr. Corrosion was very moderalte,
with chromium leaching equivalent to complete re-
moval from a layer only 0.2 mil deep. (Surveillance
specimens showed a chromium gradient to a depth of 2
mils.) Oxidation on surfaces exposed to the cell
atmosphere amounted to only 2 mils. However, surveil-
lance specimens exposed just outside the reactor vessel
and at the center of the core showed marked reduction
in fracture strain and stress-rupture life due fo neutron
irradiation.? 274

In the MSBR reference design the metal in the vessel
walls is protected by a thick graphite reflector and sees
a fast-neutron fluence only on the order of 1 X 10?%'
neutrons/cm? (actually less than was received by core
specimens in the MSRE). This fast-neutron fluence is
too low to produce the swelling or void formation that
is associated with the metal used for cladding the fuel in
fast reactors.?® The major concern in developing an
improved alloy for use in the MSBR was therefore not
fast-neutron damage but the production of helium in
the metal, primarily due to the thermal-neutron trans-
mutation of '°B to *He and 7Li. Boron is an impurity
of Hastelloy N that comes from the refractories used in
melting the alloy. Careful commercial practice makes it
possible to produce alloys containing 1 to S ppm boron
(18.2% of natural boron is '°B). Irradiation tests,
however, show that the amount of helium (and thus
boron) required to cause embrittlement is so low that
even alloys containing 0.1 ppm of boron are badly
damaged in this respect.?® The strong influence of such
a small quantity of boron is due to the segregation of
boron at the grain boundaries, where helium production
can have a profound effect on the fracture behavior. It
was thus concluded that the problem of irradiation-
induced embrittlement could not be solved by reducing
the boron level.

The embrittlement problem was approached by
adding alloying metals, such as titanium, niobium,
zirconium, and hafnium, so as to form borides that
would be dispersed as precipitates and not particularly
segregated at the grain boundaries. This approach

27

proved successful, with a fine dispersion of MC-type
carbides giving the most desirable properties.®” The
postirradiation fracture strains of several promising
alloys are shown in Fig. 3.13. (Although the fluence
received by these specimens is low compared with that
expected in the MSBR, over one-half of the boron will
have been transmuted at the 5 X 10*°-neutron/cni®
fluence level, and there is relatively little change in
ductility beyond this point,)

To obtain the desired structure and welding prop-
erties of the modified alloy, close control is required of
the concentrations of titanium, niobium, and hafnium.
Successful highly restrained test welds have been made
in ' -n.-thick plate using alloys containing 1.2% tita-
nium, 0.5% hafnium, combined 0.75% hafnium and
0.75% titanium, and combined 0.5% titanium and 2%
niobium. (Zirconium induced severe weld metal crack-
ing and is no longer considered as a constituent.) The
composition of the Hastelloy N for the MSBR has not
been optimized, but the anticipated values are given in
Table 3.5.

ORNL-DWG 69-5877BR2

15.4
AND M 23.6,4Ti 16.7, 04 Ti+0.6Cb+0.5Hf
22.7]|22.0,1Ti

19 9196, Ti ¢
i "
L) 3L | /
13 fre —-»~-tcm -t 7
LI | //
2 THERMAL FLUENCE: 5x10°? neutrons/cm® ////
IRRADIATION TEMPERATURE ! 760°C ,{/;
1" A HE 1 ‘Z///A
0.8 Hf + ’ {,//;j
'0 =3 i I A /a
_ 9 {HE+1Y %V%fl
- f l Y
= B | 0.6 Hf-1,2 cu///;/ :
2 Vit 1 11V
& ke llso.BTi+t3Cb+ | |/7 / AL
w. ¥ A 0.6Hf+0.3551 | J/ /,/
S 0.9 Cb+0.8Hf 7
6 1 e 7
: sl A
S o.sn~o.ecu*r S
LI 1| //////fl
a || LIS o5Ti vo.ant 4] 74
1 7
%' 2Cb+0.5Ti
3 12 H A1 +
2
{ ~STANDARD HASTELLOY-N ]|
i rieses ([ 1 ]11
0.00 0.04 Y { 10 100

STRAIN RATE (%/nr)

Fig. 3.13, Postirradiation ductility of Hastelloy N at 650°C.
The corrosion resistance of the modified material has
been tested, and specimens have been exposed in the
MSRE core, The melts used to date have <0.1% iron and
have even lower corrosion rates than observed for the
standard alloy with 4 to 5% iron. Iron does not serve a
critical rele in the alloy and could be 1emwuved o give a
lower corrosion rate in sodium fluoroborate should this
prove to be necessary. The presence of titanium and the
other reactive metals will not contribute appreciably to
the corrosion rate at the anticipated concentra-
tions.?#+%? The molybdenum was dropped from 16% in
the standard material to about 12% in the modified
alloy to obtain the desired carbide.

The mechanical properties of the modified alloys are
generally better than those of standard Hastelloy N and
are considerably better than those of the early heat
used in establishing the allowable design stresses under
the ASME Code. For the purposes of this reference
MSBR design, however, the approved stresses, listed in
Table S.1, were used.

In summary, the reference MSBR design assumes that
material having strength and corrosion resistance equal
to standard Hastelloy N will be available. The reactor
vessel requires, in addition, that the postirradiation
ductility be much better than that of the standard
alloy. Many experimental heats of modified Hastelloy N
meet these requirements. There appears to be no reason
why a selected alloy cannot be produced commercially
and be approved for code construction,

3.2.4.2 Secondary system. The coolant salt in the
MSBR is sodium fluoroborate, This does not present a
basically different corrosion situation from that for
other fluoride salts, since the elements present as
fluorides are more stable than are the fluorides of the
metals present in the Hastelloy N. Impurities in the salt,
however, may present mechanisms for corrosion.

Static corrosion tests showed insignificant attack of
Hastelloy N by NaBF,;-NaF mixtures (4 to 8 mole %
NaF) on Hastelloy N with low amounts of oxygen and
water present,*? Increased amounts of oxygen and
water may accelerate the corrosion rate,

Dynamic corrosion test experience with Hastelloy N
in sodium fluoroborate includes several thermal convec-
tion loops and a single forced-circulation system.
Results indicate that metal will be removed from the
hotter portions of the loop and deposited on the cooler
sections. For the thermal convection loops the mini-
mum rate of metal removal was about 0.2 mil/year over
about 10,000 hr of operation. Accelerated corrosion is
associated with high levels of H,O and O, . Purging the
system with a gaseous mixture of hydrogen fluoride,
BF;, and helium appears to be an effective method of

28

purifying the coolant salt of moisture and oxygen,
however. In general, the compatibility of Hastelloy N
with sodium fluoroborate appears acceptable. If on-ine
methods for removing corrosion products and moisture
are included in the system, the corrosion rate is likely
10 be less than about U.2 mil/year.

The compatibility of Hastelloy N with supercritical-
pressure steam has been tested by exposing specimens
in the TVA Bull Run steam station. In over 10,000 hr
the corrosion rate has been less than % mil/year,*' a
rate that by industry standards would certainly be
acceptable in the steam generator tubing. (There is no
significant difference between the standard and modi-
fied Hastelloy N in this respect.) Results of continued
testing, but with stressed specimens, are not yet
available.

3.3 NUCLEAR CHARACTERISTICS

3.3.1 Selection of MSBR Core Design

H. F. Bauman

The core of the single-fluid MSBR consists of two
zones: a well-moderated inner zone (I) (see Table 3.3
for definition of zones) surrounded by an undermod-
erated outer zone (1I). The same fuel salt, containing
both fissile and fertile material, is used in both zones
and throughout the reactor, The neutron spectrum in
each zone is controlled by adjusting the proportion of
salt to graphite, from a salt fraction of about 13% in
zone | to about 37% in zone II. The overall spectrum is
adjusted for the best “performance” associated with a
high breeding ratio and a low fissile inventory (optimi-
zation of the core is discussed in following sections).
The spectrum in zone 11 is made harder, to enhance the
rate of thorium resonance capture relative to the fission
rate, thus depressing the flux in the outer core zone and
reducing the neutron leakage.

Earlier MSBR designs achieved excellent performance
(good breeding ratio and low fissile inventory) by
incorporating fissile and fertile materials in two separate
fluids.* Both fluids (separated by graphite walls) were
present in the core, which was surrounded by a blanket
of fertile salt. The advantages of this two-fluid design
were low fissile inventory (because the fissile material
was confined to a relatively small volume of fuel salt)
and ease of processing (because the fuel salt was free of
fertile material and the fertile salt was practically free
of fission products). The main disadvantage was the
complex graphite structure required to separate the two
fluids, a structure that would have to be replaced at
intervals because of neutron damage to the graphite.
A design intermediate between the two-fluid and the
reference single-fluid designs is the single-fluid core with
separate blanket.! The core is a well-moderated region
like zone I of the single-fluid design, surrounded by a
blanket of thorium-bearing salt separated from the core
by a thin wall of Hastelloy N (or possibly graphite). The
core salt contains both fissile and fertile materials and
thus offers no processing advantage over the single-fluid
design, but the presence of the blanket controls neutron
leakage without involving a large fissile inventory in the
blanket region and results in a low total reactor fissile
inventory. Exploratory calculations have shown that
the performance of this design approaches that of the
two-fluid reactor mentioned above. Its major disad-
vantage is the necessity for a dividing wall between the
core and the blanket, a wall that would have to be
replaced periodically (along with the core graphite)
because of fast-neutron damage.

Another possibility is a single-fluid design with a
power density low enough for the allowable damage
flux to the core to not be exceeded in the lifetime of
the reactor. Preliminary calculations show that such a
reactor should have a large core (on the order of 30 ft
in diameter) and that an undermoderated zone Il is not
needed because leakage is inherently low from such a
large core. The advantage of this design is simplicity of
construction and the elimination of core replacement.
Its disadvantages are the relatively high fissile inventory
and the large size of the reactor vessel.

The performance of typical examples of these four
reactor designs is summarized in Table 3.6. The second
one listed, the single-fluid two-zone replaceable core
design, was selected for detailed analysis in this design
study because it offers moderately good breeding
performance in a design that can be built with only a
modest extension of today’s technology.

29

3.3.2 Optimization of Core Design

H. F. Bauman

The ROD (Reactor Optimum Design) code, used to
optimize the core design for the single-fluid MSBR,
consists of three major sections:

1. A multigroup, one-dimensional neutron diffusion
calculation based on the code MODRIC with a routine
added to synthesize a two-dimensional calculation in
cylindrical geometry.

2. An equilibrium reactor calculation based on the
code ERC. The equilibrium concentrations of up to 250
nuclides including fission products may be calculated
for considering continuous fuel processing with up to
ten removal modes, each with its individual processing
time. The breeding ratio, fuel yield, material inven-
tories, and fuel-cycle costs are calculated in this section.

3. An optimization procedure, based on the gradient
projection method or “method of steepest ascent,” for
locating the maximum of a specified figure of merit
when given reactor parameters are allowed to vary. The
figure of merit may be any desired function of the
breeding ratio, the specific fuel inventory, the fuel-cycle
costs, or similar factors, while such parameters as core
salt fraction, the core zone dimensions, reflector thick-
ness, and processirig cycle times may be variables.
Parameter surveys at specified levels of the variables
(without optimization) may also be performed.

3.3.2.1 Cross sections. Cross-section sets for use in
MSBR calculations were developed using XSDRN,*? a
discrete ordinates spectral code for the generation of
nuclear multigroup constants in the fast, resonance, and
thermalization energy regions. Cross-section sets were
made for each of the four major regions of the reactor:
the 13.2% salt zone, the 37% salt zone, the 100% salt
gap, and the reflector. In each case a “cell structure”

Table 3.6. Calculated nuclear performance of 1000-MW(e) MSBR design concepts

Typical performance value

Design concept Conservation Fissile Annual fuel
(in increasing order of complexity) coefficient? Breeding ratio inventory yield
[MW(t)/kg) ? (kg) (%]year)
Single-fluid, nonreplaceable core S 1.06 2300 2.0
Single-fluid, two-zone replaceable core 15 1.06 1500 3.2
(reference MSBR)
Single-fluid core with separate blanket 50 1.07 900 7.0
and replaceable core
Two-fluid core plus blanket and 75 1.07 700 8.0

replaceable core

@An index of merit, the definition and significance of which are discussed in sect. 3.3.2.2.
was set up to describe a part of the particular regions.
The cross sections were then flux weighted over the
cell. The input data for XSDRN were taken from the
123-group XSDRN master library tape. This 123-group
structure was reduced to a 9-group structure in the
XSDRN calculations; this hraad group etrunture son
sists of 5 fast groups and 4 thermal groups. Nuclide
concentrations for these calculations were obtained
from a ROD calculation. All the nuclides appearing in
the reactor plus four 1/v “nuclides’ were considered in
each region so that four sets of cross sections were used
to describe the entire reactor.

3.3.2.2 Conservation coefficient. The figure of merit
selected for optimization of the single-fluid MSBR has
been named the ““conservation coefficient,” defined as
the breeding gain times the square of the specific power
in thermal megawatts per kilogram of fissile material
(which is proportional to the inverse of the product of
the doubling time and the fuel specific inventory). The
conservation coefficient is related to the capability of a
breeder reactor system to conserve fissile material in a
nuclear power economy expanding linearly with time.
For this power growth condition, maximizing the
conservation coefficient results in a minimum in the
total amount of uranium that must be mined up to the
point when the breeder system becomes self-sustaining
(i.e., independent of any external supply of fissionable
material).

3.3.2.3 Optimization. The optimization of the reac-
tor design, while based on maximizing the conservation
coefficient, was subject to several economic constraints,
including limits on the power density (and hence the
graphite life) and the overall reactor vessel dimensions.
In addition, the rare-earth and *?3Pa processing rates
were fixed at rates found reasonable for the reductive-
extraction processing method considered here. Fuel-
cycle costs were computed as part of the core calcula-
tions, as shown in Table 3.7. Although not used as the
basis for the optimization, it turned out that fuel-cycle
costs were near minimum at the selected optimum
configurations.

3.3.2.4 Reference design. The results of the optimi-
zation study led to the selection of a reference design
with the characteristics given in Table 3.7. Additional
data on the flux spectrum and the neutron absorption
by individual fission product nuclides in the reference
design are given in Appendix B. The data given are from
the calculation of the reference design and include
details of the processing and buildup of higher isotopes.
However, another calculation, which differs only in

detail from the reference design calculation, was used
for several subsidiary calculations, such as the neutron
and gamma heating in the core and the power distri-
bution in the core, and for several of the parameter
surveys given in the following section.

Table 3.7. Characteristics of the single-fluid
MSBR reference design

A. Description

Identification CC120
Power
MW(e) 1000
MW(t) 2250
Plant factor 0.8
Dimensions, ft
Core zone 1
Height 13,0
Diameter 14.4
Region thicknesses
Axial:
Core zone 2 0.75
Plenum 0.25
Reflector 2.0
Radial:
Core zone 2 1.25
Annulus 0.167
Reflector 2.5
Salt fractions
Core zone 1 0.132
Core zone 2 0.37
Plenums 0.85
Annulus 1.0
Reflector 0.01
Salt composition, mole %
UF, 0.232
PuF, 0.0006
ThF, 12
Bng 16
LiF 72
B. Processing
Processing group Nuclides Rl tme

(at full power)

Rare earths Y, La, Ce, Pr, Nd, Pm, 50 days

Sm, Gd

Eu S00 days

Noble metals Se, Nb, Mo, Te, Ru, Rh, 20 sec

Pd, Ag, Sb, Te
Seminoble metals  Zr, Cd, In, Sn 200 days
Gases Kr, Xe 20 sec
Volatile flnorides  Br, | 60 days
Discard Rb, Sr, Cs, Ba 3435 days
Salt discard Th, Li, Be, F 3435 days
Protactinium 2351’3 3 days
Higher nuclides 7‘37Np, 242py 16 years

Table 3.7 (continued)

C. Performance

31

Conservation coefficient, [MW(1)/kg)| 14,1
Breeding ratio 1.063
Yield,? % per annum 3.20
Inventory, fissile, kg 1504
Specific power, MW(t)/kg 1.50
Doubling time, system,? years 22
Peak damage flux, £ > 50 keV,
neutrons cm 2 gec ™!
Core zone 1 3.3 X lO14
Reflector 3.7 % 10"?
Vessel 4.3 % 10!
Thermal-neutron flux. neutrons em ™2 sec ™
Average, core 2.6 X 10'?
Peak 8.3 x 10"*
Fraction of fissions from thermal neutrons 0.84
Power density, W/cm3
Core
Average 22.2
Peak 70.4
Core, fuel salt
Average 74
Peak 492
Core, graphite (gamma and neutron heating)
Average 2.3
Peak 6.3
Fission power fractions by zone
Core zone 1 0.790
Core zone 2 0.150
Annulus and plenums 0.049
Reflector 0.012
Ratio, C/Th/U
Core zone 1 8660/52/1
Core zone 2 2240/52/1
D. Neutron balance
Constituent Concentration? Absorptions  Fissions
2321y 375 x 1072 0.9779 0.0030
233py 3.88 x 1077 0.0016
23317 6.64 x 1073 0.9152 0.8163
$34p 2.31x 1078 0.0804 0.0004
g ¥ 6.01 % 107° 0.0747 0.0609
236y 6.21 X 107° 0.0085
237Np 8.59 % 1077 0.0074
238py 6.10 x 107° 0.0074
239p, 1.29 x 1077 0.0073 0,0045
240py 6.83x 107" 0.0027
241 py 6.21 x 107 0.0027 0.0020
2p, 1.23 x 1077 0.0006
6Li 1.95% 1077 0.0035
TLi 2.24x 1072 0.0157
9Be 5.00 x 1073 0.0070 0.0045¢
19F 4.77 x 1072 0.0201
Graphite 0.0513
Fission products 0.0202
Leakage 0.0244
ne 2.2285

Table 3.7 (continued)

E. Fuel-cycle costs?

[tem Cost (mills/kWhr)
Inventory
Fissile 0.364
Salt 0.077
Replacement salt 0.040
Processing 0.360
Fissile production credit —-0.088
Total 0.753
At 0.80 plant factor,

bNuclide concentration in fuel salt (atoms b~ cm ™).
€(n,2n) reaction.

dBases for the fuel-cycle cost estimate are summarized in
Table D.2.

3.3.3 Effect of Changes in the Fuel-Cycle and
Core Design Parameters

H. F. Bauman

3.3.3.1 Power density and core life. The power
density of the core affects both the reactor perform-
ance and the core graphite life. As the first step in
selecting the core power density, the core dimensions
(and the salt fraction of zone [) were optimized to
maximize the conservation coefficient. Then several
cases were run in which the maximum permissible
fast-neutron fluence was limited to low values, which
had the effect of increasing the core size, limiting the
peak power density, and increasing the core graphite
life. The results of this study are shown in Fig. 3.14, in
which the performance parameters are plotted as a
function of graphite life. Both the breeding gain and the
fissile inventory increase as the ‘core is made larger, but
the increase in breeding gain flattens out for larger
cores, so that a maximum conservation coefficient is
obtained at a core life of about three years, which
corresponds to about a 15-ft-diam core with a peak
power density of about 100 W/cm?®. However, there is
little change in the conservation coefficient as the core
is enlarged to increase the graphite life to about four
years, which corresponds to the reference design core
diameter of about 17 ft and peak power density of
about 70 W/cm?.

3.3.3.2 Salt volume fraction and thorium concentra-
tion. The function of thorium as the fertile material in
the reactor is to absorb neutrons and thereby produce
fissile 2?3U. Thorium competes for the available
neutrons with fissile material on the one hand and
parasitic absorbers such as fission products and the
ORNL—DWG 69— 7397A
20

@

>

CONSERVATION COEFFICIENT; SPECIFIC

YIELD (%/yr); GAIN
>

INVENTORY [kg/Mw (e)]

-~
n

REFERENCE DESIGN

cC

4 5 6
GRAPHITE CORE LIFE (yr)

Fig. 3.14. Performance of 1000-MW(e) MSBR as a function
of core life (at 0.8 plant factor).

material of the carrier salt and the moderator on the
other. As a result of this competition there is an
optimum concentration of thorium in the core. If the
thorium concentration is high, the breeding ratio will be
high, but a large amount of fissile material (to compete
with the thorium for neutrons) will be required to make
the reactor critical. If the thorium concentration is low,
the fissile inventory required will be low, but the
breeding ratio will also be low because more neutrons
will be lost to the parasitic absorbers (because of a lack
of competition from thorium and uranium). The
thorium concentration also affects the neutron energy
spectrum, which becomes harder as the thorium is
increased. Hardening the spectrum tends to increase the
resonance absorptions in thorium while decreasing the
relative absorptions in fissile and parasitic materials,
thus reinforcing the competitive effect of thorium
already described.

In the MSBR the core thorium concentration is
determined by the core salt fraction and the concen-
tration of thorium in the salt. The thorium concen-
tration in the salt determines the ratio of thorium to
most parasitic absorbers, while the concentration and
salt fraction together determine the thorium-to-uranium
and carbon-to-uranium ratios.

The effect of thorium concentration on performance
of the MSBR is shown in Fig. 3.15. The cases
represented in this figure were calculated before the
reference design was selected and were based on a
slightly smaller external salt inventory. Details of these
cases are given in Table 3.8 and ref. 9. The core

32

ORNL-DWG 69-7398A

8 e 18 o
__———\ —
-\‘ g:—w
i
26| e — T
= G (x100) —
o o
>4 |— = 40 &
o
2 ; ge
a | ot ) 32
od ‘ Bae=
REFERENCE DESIGN §§
l-(;,-lu
b
0 ! 103
10 T 12 13 14 15

THORIUM CONCENTRATION (mole %)

Fig. 3.15, Influence of thorium concentration on the per-
formance of a single-fluid MSBR.

dimensions and zone I volume fraction were allowed to
optimize. As each case approached an optimum the
cross sections were reweighted to allow for spectrum
changes. The broad maximum in the conservation
coefficient occurs in the vicinity of the 12 mole %
thorium concentration, and this concentration was
selected for the reference design.

One of the principal conclusions reached in the study
of the MSBR was that the performance of the reactor is
not sensitive to small changes in the thorium concen-
tration in the salt, provided that the salt fraction is
freely adjusted to maintain about the optimum carbon-
to-thorium ratio. The optimum thorium concentration
tends to increase as the core power density is decreased,
but this effect is small over the range of power densities
that give graphite lifetimes in the range of two to four
years.

The effect of allowing the core zone I volume fraction
to change, with all other parameters held fixed as in the
reference design, is shown in Fig. 3.16. There is a broad
optimum in the conservation coefficient at 13 vol % salt
and a very broad optimum in the fuel yield at 14 vol %.
The reference design value of 13.2 vol % sall is the
result of a ROD optimization calculation.

The effect of the core zone II volume fraction was
also studied. With the total yolume of fuel salt in zone
I held fixed at its optimum value, a very broad
optimum in the conservation coefficient was found to
lie between 35 and 60 vol % salt. The salt fraction of
37% in the reference design was chosen to permit the
use of a random-packed ball bed (of 37% void volume)
for zone 11 if desired.
33

Table 3.8, Influence of thorium concentration on the
performance of a single-fluid MSBR

Fuel salt, mole % LiF-BeF,-ThF,
Core height,? ft

Core diameter,? ft

Radial blanket thickness,? ft
Axial blanket thickness,? ft
Radial reflector thickness, ft
Axial reflector thickness, ft

Core salt fraction?

Radial blanket salt fraction

Axial blanket salt fraction
Reactor power, MW(t)

Average power density, W/cm3
Maximum power density, W/cm®
Graphite replacement life, years
Specific fuel inventory, kg/MW(e)
Breeding ratio

Annual [uel yield, %/year
Conservation coefficient

74-16-10 72-16-12 70-16-14
9.75 9.8 11.7
11.2 I1.1 11.5
2.52 2.20 1.89
1,45 1.19 0.91
3.0 3.0 3.0
2.0 2,0 2.0
0.137 0.121 0.114
0.37 0.37 0.37
0.37 0.37 0.37
2250 2250 2250
29.4 33.14 33.5
97.3 106.3 101.6
2.1 1.9 2.0
1.23 1.26 1.33
1.051 1.055 1.060
3.18 3.35 342
17.19 17.65 17.05

@Variables allowed to optimize,

ORNL-DWG 70-14314

\
\

1(%40)

o

CC

YIELD (%/yr), GAIN (%)

CONSERVATION COEFFICIENT
SPECIFIC FUEL INVENTORY (kg/Mwle))

~
)

" 12 13 14 15 16

CORE VOLUME FRACTION SALT (%)

Fig. 3.16. Effect of core zone I volume fraction of salt on
MSBR performance. (Other parameters are held fixed at
reference design values.)

3.3.3.3 Reflector. Both the thickness and the salt
fraction of the reflector are important to the MSBR
design. Increasing the reflector thickness over the range
from 1 to 4 ft was shown to increase the conservation
coefficient of a typical MSBR design.” Much of the
benefit of the reflector stems from its effect in
increasing the neutron flux in the outer region of the
core, thus giving a more even core power density
distribution and improving the specific power without
increasing the peak damage flux in the core. However,
the improvement in performance was slight beyond a

3-ft thickness. On this basis, a 2-ft axial and 2.5-ft radial
reflector thickness were selected for the reference
design.

The salt fraction in the reflector is also important.
Calculations have shown that if all the fuel salt were
eliminated from the reflector region, the conservation
coefficient of the reference design could be improved
by 20% over the reference design, mainly due to a
significant reduction in the neutron leakage from the
reactor, However, the reflector salt fraction of 1%
selected for the reference design was determined by
engineering considerations and is about as low as could
be achieved in a practical design.

3.3.3.4 Processing. The ROD code was set up to
model in detail the reductive extraction processes
described in ref. 1, The various parasitic absorber
groups and the processing cycle times assumed in the
calculation of the reference design are given in Table
3.7. The treatment of the processing appears compli-
cated, but only two of the steps, the protactinium
removal and the rare-earth removal, control the eco-
nomics and performance of the MSBR. The effect on
the conversion ratio of varying the processing rate of
these two main steps, along with proportionate rate
changes for subsidiary steps (e.g., seminoble metals with
protactinium removal), is given in Fig. 3.17.

The most obvious conclusion from this study is that
rapid processing is essential to good breeding perform-
ance. Another conclusion is that somewhat less strin-
gent processing times than were assumed for the
reference design, say a 10-day instead of a 3-day
ORNL-DWG 69-12698A

1.08 - .
3doys Pa CYCLE TIME
“@
“e
.05 [— 1O i
30 \\JL
L Sl
= \
- 3000
N
2 ool B
g .00 ‘t ——]
N S; A 233,
o AN
l;' \.Q\ | Pu
= A P \
o \\ \ 235U
O &
\ \ 233,
0.95 \A\ —
\AZ”U
0.90
0 100 200 300 400 500 600

RARE EARTH CYCLE TIME (days)

Fig. 3,17. Conversion ratio of a single-fluid MSBR as a function of processing cycle times and feed materials. (The plutonium
feed composition was assumed to be 239Pu = 60%, 24°Pu = 24%, 24! Pu = 12%, and 2%2Py = 4%, Other parameters were held fixed

at reference design values.)

protactinium cycle and a 100-day instead of a 50-day
rare-earth cycle, can still give fairly good breeding
performance. Further, increasing the protactinium
processing cycle time can be *“‘traded™ for a decreased
cycle time for the rare earths. Thus, use of a 10-day
protactinium removal cycle time and a 25-day rare-
earth removal cycle time would give about the same
breeding ratio as would the processing times assumed
for the reference design, that is, about 3 days for
protactinium removal and 50 days for rare-earth re-
moval. (The processing plant described in Sects. 2.6 and
8 gives a 10-day protactinium cycle time.)

Rapid and inexpensive processing is the potential
advantage of fluid-fueled reactors. However, very long
processing times have been considered in order to
examine the performance of the MSR at processing
rates more typical of solid-fueled reactors. For long
cycle times, where the conversion ratio drops below
1.00, three makeup feed fuels were investigated: *?3U,
a plutonium mixture typical of that from water
reactors, and 93% enriched 233 U. The results are shown
in Fig. 3.17. The calculations show, for example, that
with no protactinium processing and a 500-day rare-
earth cycle (which would correspond to about a
three-year batch-processing interval), the conversion
ratio is well over 0,90, which is very good compared
with solid-fueled converters. The study also shows that

plutonium would be an attractive fuel for converter
operation.

An important parasitic absorber that was not con-
sidered to be removed in the reductive-extraction
processes is 2> 7Np. There are now indications that it
can be successfully eliminated. If **7Np were removed
on a 200-day cycle, a ROD calculation indicates thal
the breeding ratio of the reference MSBR would
increase from 1.063 to 1.070 and the conservation
coefficient from 14.1 to 16.2.

3.33.5 Plant size. Neutron leakage is important in
the single-fluid MSBR due to the absence of a blanket.
Furthermore, the undermoderated core zone 11, which
substitutes for a blanket, although reasonably effective
in reducing leakage, contains a large volume of fuel salt
and therefore adds heavily to the fissile inyentory. The
performance of the reactor, then, is strongly affected
by factors which affect the leakage; the most important
of these is the size of the reactor.

The 1000-MW(e) plant size selected for the reference
MSBR was chosen because this has become a standard
size for comparative studies of reactor plants. No
attempt was made to revise the plant design for larger
or smaller sizes, but a simple scaling study was made to
indicate the performance that could be expected from
other size plants, particularly larger ones.
The scaling study was started by taking the external
fissile inventory and the volume of core zone 1
proportional to plant power and holding fixed the
thicknesses of core zone II, annulus, plenums, and
reflector, The results of this study for reactor plants of
500 to 4000 MW(e) are shown as the dashed curves in
Fig. 3.18. There was considerable spread in the peak
power densities, and therefore the core graphite life, in
this set of cases, and a second set was run in which the
core zone I volumes were adjusted to give about the
same peak power density in each case. The results of
this set are shown as the solid curves in Fig. 3.18 and
are given in Table 3.9. The performance, as measured
by both the conservation coefficient and the fuel yield,
increases sharply with increase in plant size. The
single-fluid MSBR, then, is well suited to large plants.
For small plants, reactor designs less sensitive to
neutron leakage, such as the single fluid MSBR with
fertile blanket, should be considered.

3.3.4 Reactivity Coefficients
O.L.Smith  J.H, Carswell

A number of isothermal reactivity coefficients were
calculated using the reference reactor geometry. These
coefficients are summarized in Table 3.10. The Doppler
coefficient is primarily that of thorium. The salt and
graphite thermal base coefficients are positive because
of the competition between thermal captures in fuel,
which decrease less rapidly than 1/», and thermal
captures in thorium, which decrease nearly as 1/v, with
increasing temperature. The salt density component
represents all effects of salt expansion, including the
decreasing self-shielding of thorium with decreasing salt
density.

35

ORNL~-DWG 65-7396A

‘0 v T 1 '
CONSTANT CORE LIFE
g | == == =me CONSTANT RATIO OF REACTOR ———— 40
POWER TO CORE VOLUME

w
w 100 —
2 ==
w ’-D" E
g 7 f—l ! g 30 w
S e 16T 3]
~ W
o 6 - r W
d 7 3
> 5| 29 %/yr— 20 2
= =
3 4 - S
o N x
= / — o
G o
g -2 3|

0

0 \ 2 3 4 5

REACTOR POWER (10> Mw(e))

Fig. 3.18. Effect of plant size on MSBR performance.

The graphite density component includes both
changing graphite density and displacement of graphite
surfaces. In calculating the displacements, it was
assumed that the graphite-vessel interface did not move,
that is, the vessel temperature did not change. For
short-term reactivity effects, this is the most reasonable
assumption, since inlet salt bathes the vessel's inner
face. In any case, it should be noted that the graphite
density coefficient is a small and essentially negligible
component.

From Table 3.10 it is seen that the total core
coefficient is negative. But more important, the total
salt coefficient, which is prompt and largely controls

Table 3.9. Performance of single-fluid MSBR’s as a function of plant size?

Reactor power [MW(e)] 500 1000 2000 4000
Core height, ft 9.44 11.0 17.44 23.0
Core diameter, ft 10.42 14.4 19.36 5.5
Salt specific volume, ft3/MW(e) 1.75 1.68 1.62 1.55
Fuel specific inventory, kg/MW(e) 1.65 147 1.36 1,28
Peak power density, W/cm? 62.2 65.2 66.1 65.9
Peak flux (£ > 50 keV), 10'% neutrons cm ™ sec™! 3.04 3.20 3.25 3.24
Core life, years at 0.8 plant factor 3.9 3.7 3.7 330
Leakage, neutrons per fissile absorption X 100 3.89 2.44 1.53 0.96
Breeding ratio 1.043 1.065 1.076 1.083
Annual fuel yield,? %/year 1.99 3.34 4.28 4.95
Conservation coefficient 8.0 15.1 21.0 25.9

AThe thickness of core zone II, annulus, plenums, reflectors, and other parameters not otherwise indicated were held fixed at the

reference design values indicated in Table 3.1.
PThe plant factor is assumed to be 0.80.
Table 3,10, Isothermal reactivity coefficients
of the reference reactor

Reactivity coefficient,
Component 1 2k

o
o (per C)

% 107°
Doppler —4,37
Salt thermal base +0.27
Salt density +0.82
Total salt -3.22
Graphite thermal base +2.47
Graphite density -0.12
Total graphite +2.35
Total core —0.87

the fast transient response of the system, is a relatively
large negative coefficient and affords adequate reactor
stability and controllability.

The salt density coefficient is particularly important
with regard to bubbles in the core salt. It is expected
that the salt will contain a few tenths of a percent of
xenon bubbles. Under certain circumstances the bub-
bles might expand or decrease in volume without
change in core temperature and hence without invoking
the total salt temperature coefficient. Since the salt
density component is positive, without decreasing
density, bubble expansion would produce a positive
reactivity effect. Using a salt expansion coefficient of
SV/V = 2.1 X 107%/°C, an increase in core bubble
fraction from, say, 0.01 to 0.02 would yield a reactivity
change of 8k/k = +0.00039. This is approximately
one-fourth the worth of the delayed neutrons in the
core. Analogously complete instantaneous collapse of a
0.01 bubble fraction would yield a reactivity change of
8k/k =—0.00039.

Finally, the equilibrium fuel concentration coeffi-
cient, (8&/k)/(6n/n), where n is atomic density, was
calculated to be 0.42 for **3U and 0.027 for 35U,
and 0.39 for total fissile uranium. (The coefficient for
2350 is much smaller because the *?*%U inventory in
the MSBR is very low relative to 2*?U.)

3.3.5 Gamma and Neutron Heating in the MSBR
O.L.Smith  J.H. Carswell

Gamma and neutron heat sources in the one-fluid
reactor, vessel, and thermal and biological shields were
calculated using gamma and neutron transport tech-
niques based on the ANISN transport code.

Results are given here for one axial and two radial
traverses of the reactor and shields. The region thick-

nesses and composition are shown with the results in
Figs. 3.19-3.24. For the radial traverses, two one-
dimensional infinite-cylinder calculations were per-
formed — the first at the core midplane and the second
in a plane two-thirds of the distance from the midplane
10 the top of the core. In each case the neutron (and
gamma) flux was normalized to the value of the actual
center-line core flux at that elevation. No allowance was
made for axial buckling. Thus, particularly in the
shields, the calculated heat sources should be con-
sidered as upper limits to the actual heat sources. It is
estimated that the calculated sources inside the reactor
vessel are only a few percent high. But because of the
large air gap between the vessel and shields, the
calculated heat sources in the thermal shield and
concrete should be reduced by about 50% to account
for the actual finite height of the reactor.

In the axial center-line calculation, the system was
represented in slab geometry, infinite in the radial
dimension. Again, transverse buckling effects inside the
vessel are small. The results for the thermal and
biological shields are upper limits, but the overesti-
mation is lower in the axial direction since the air gap is
only a few feet.

The calculations were performed in several linked
stages starting with a one-dimensional ANISN transport
calculation of the neutron space and energy distribution
in the reactor and shields. From neutron fluxes and
scattering cross sections, the neutron heat distribution
was determined. The neutron heating in the reactor is
shown in Figs. 3.19, 3.21, and 3.23 for the two radial
traverses and one axial traverse. In each figure, curve A
shows the heat source per unit volume of homogenized
core, blanket, reflector, or plenum. Curves B and C
show, respectively, the heat source per unit volume of
graphite and salt separately in those regions. Curve D
shows the heating in the INOR vessel.

Figures 3.20, 3.22, and 3.24 show the gamma and
neutron heating in the thermal and biological shields.
The thermal shield is treated as pure iron, The concrete
is a standard grade.

The gamma heat distribution is similarly presented in
the figures. Three sources of gammas were calculated
from the neutron flux distribution: prompt fission,
delayed (fission product), and capture gammas. The
first and last of these had the spatial distribution of the
neutron flux. The delayed source was assumed uniform
in the circulating salt. Since the salt spends approxi-
mately half its time in the reactor, approximately half
of the delayed gammas are emitted inside the vessel.

These three sources of gammas were combined in a
fixed-source ANISN gamma transport calculation using
37

g ORNL-DWG €9~12607
= AA
g) 2
o \
o
:§ | 8
T3
3 N
o
c
o
= . \\\ 4 AND 8 \‘/
w v v
z 100 % SALT ANNULUS 100 % SALT
(5.09 cm) \ (0,64 cm) ?SEQ%TES)VESSE'-
L CORE com—: REFLECTOR %
0.132 VOLUME FRACTION SALT o 37 0.04 | VvoID
R | | Jio =%
€ . CURVE 4: SALT AND GRAPHITE
3 CURVE &: GRAPHITE ONLY X
= CURVE C: SALT ONLY
= '\\ CURVE 0: INOR
=
b T —
Q
D 4 A
8 \
a x
[,
o \
T2 <3 A ¢
a g A\
: = \
© N\l 4AND & W2
0 AN—
0 40 80 120 160 200 240 280 320 360 1000

RADIUS (cm from core aoxis)

Fig. 3.19. Gamma and neufron heating in the core midplane of a 1000-MW(e) MSBR (R = 0 to 1000 cm).

seven gamma energy groups. From the gamma fluxes
the gamma heat sources were then calculated.

From the results it should be particularly noted that
neutron thermalization is a major heat source in the
graphite.

3.3.6 Fission Product Heating in the MSBR
R. B. Briggs  J. R. Tallackson

One of the principal design considerations for an
MSBR is the safe disposal of reactor afterheat. The five
major sources of heat which remain in the primary
system after shutdown are:

1. fission heat due to decay of flux at shutdown,
including the effect of delayed-neutron precursor
transport by the salt;

2. decay of fission products (and daughters) dispersed
in the primary salt;

. decay of noble-metal fission products (and daugh-
ters) deposited on the graphite and Hastelloy N
surfaces;

decay of gaseous krypton and xenon (and daughters)
diffused into the graphite;

5. heat stored in moderator and reflector graphite.

The heat loads imposed by fission products must be
recognized and evaluated in order to design cooling
systems for the chemical processing equipment, the
off-gas system, the drain tanks, and the primary salt
circuit. The distribution of heat producers within the
system depends on chemical behavior, half-life and
complexity of decay chains, graphite characteristics,
and the effectiveness of the chemical and off-gas
removal systems. The available evidence indicates that
the noble metals (Nb, Mo, Tc, Ru, Rh, and Te) plate
out on metal and graphite surfaces almost as soon as
they are formed, collect at liquid-gas interfaces in the
ORNL—-DWG 69-12608

2
o
e
O
o)
o
8 \
@ \
- v
al \
W
=5
& o P
&
~
=
w
2
THERMAL SHIELD (7.62-cm Fe)
daid .
ly-——a-{ |--—-—CONCRETE ———*{
30

n
o
- —

]
—

GAMMA HEAT PRODUCTION (mw/em3)

GAMMA DOSE = 183 r/hr \
\\ v

1040 1080 1120
RADIUS (cm from core oxis)

0
1000 Heo

Fig. 3.20. Gamma and neufron heating in the core midplane
of a 1000-MW(e) MSBR (R = 1000 to 1160 cm).

salt system, or are removed with the off-gas. The
krypion and xenon either diffuse into the graphite or
are removed with the off-gas. The iodine daughters of
the telluriums are assumed to remain with their parents,
and the iodine produced directly by fission remains
dissolved in the salt. The remaining heat producers are
either dissolved in the salt or retained in the chemical
processing plant.

Kedl” has calculated the rates of diffusion of krypton
and xenon from the salt to the graphite and to the gas
bubbles in the salt. The theory and calculations are
outlined in Appendix A of this report. Briggs, using
MSRE data as a guide, estimated the distribution of
fission products in a typical MSBR design, as sum-
marized in Table 3.11. The estimate indicated that 10%
of the noble-metal production would deposit on sur-
faces of the graphite in the core, 40% would deposit on
metal surfaces in the circulation system, and 50% would
enter the gas bubbles and be transported to the off-gas
system.

38

Fig. 3.25, prepared by Tallackson, shows the distri-
bution of afterheat in the reference MSBR based on the
estimates of distribution by Briggs and Ked! and using
afterheat rates computed with the FOULBALL and
CALDRON programs by Carter. Although further
expenimental evidence supporting the choice of dif-
fusion coefficients and sticking coefficients is needed
and the throughput to the chemical processing plant is
subject to revision, the data of Fig. 3.25 probably
would produce a conservative design.

Some of the factors associated with afterheat have
been studied by Furlong,”? including various combi-
nations of magnitude and rate of reactivity insertion,
salt flow rate changes, and delay prior to the reactivity
insertion. In an example cited by Furlong,” the case of
flow coastdown, with 1% negative reactivity inserted at
0.1%/sec after a 1-sec delay (with 235U fuel), there
would be 3.75 MWhr of energy production in the salt.
Using only the heat capacity of the salt, this would
result in a 1I3°F rise in the salt temperature after
shutdown. The core graphite heat capacity, which is
twice that of the salt, would become available as a heat
sink after the salt reached an average temperature of
about 1200°F, with the net effect that the salt
temperature could be raised to about 1250°F in 5 min
after shutdown due to the effect of fission heat
production alone (assuming adiabatic conditions).

Most of the heat generated after normal reactor
shutdown will be dispersed by continued circulation of
the fuel and coolant salts and condensation of steam in
the turbine condenser. In event of a fuel-salt drain, the
heat generated in the salt would be dissipated through
the primary drain tank cooling system, as described in
Sect. 6.

3.3.7 Tritium Production and Distribution
P. N. Haubenreich

3.3.7.1 Introduction. Tritium is produced in all
reactors as a fission product and in some as a result of
neutron absorptions in deuterium, lithium, or boron in
the reactor. Because of the abundant lithium in the
MSBR, the tritium production rate is relatively high:
comparable with that in heavy-water reactors, or
roughly 20 to 50 times that in light-water reactors of
equal electrical output. Even though the tritium consti-
tutes only an extremely small fraction of the total
radioactivity that is produced, it stands out as a special
problem because at high temperatures it readily diffuses
through most metals and is difficult to contain.

Tritium in the primary salt, in its off-gas, or in the
secondary salt does not add significantly to the bio-
39

ORNL-DWG 69-12609

lvv_

NEUTRON HEAT PRODUCTION
(w/cm3)

—
&
\ c o
\NA c
=—~___ |4 AND & X
{00 % SALT
100 % SALT ANNULUS (0.64 em) REACTOR VESSEL
(5.09 cm - (5.08 cm)

CORE |

cORE 2\

REFLECTOR

- =S E—fy T
&> 0.132 VOLUME FRACTION SALT HELING Y Bl 0,04 = VOID =
P I [ Y
) CURVE 4: SALT AND GRAPHITE
E6 —— CURVE &: GRAPHITE ONLY =i
> CURVE C: SALT ONLY
= CURVE &: INOR
i
o
G 4
=
8
& Q 4
3 \
é A '—*\\ :
c
g \\ A ~N -
: \
g o % o .\D
0 40 80 120 160 200 240 280 320 360 1000

RADIUS (ecm from core axis)

Fig. 3.21. Gamma and neutron heating in a radial plane two-thirds of the distance from the midplane to the top of the core of a

1000-MW(e) MSBR (R = 0 to 1000 ¢m).

logical hazards of these fluids. Neither does diffusion of
tritium from the salt systems into the containment cell
atmosphere present a serious problem, since it should
be simple to extract the tritium from the atmosphere in
a concentrated form. It is very important, however, that
the fraction of the tritium production that reaches the
steam system be limited to a few percent. Higher
concentrations could require special precautions in
dealing with steam leaks or in handling the condensate,
and, most importantly, unacceptable amounts of
tritium must not be released into the environment in
the normal, unavoidable discharges from the steam
system,

In the reference MSBR design described in this report,
it was assumed that the barriers presented by the tubes
in the primary and secondary heat exchangers were
enough to limit the tritium reaching the steam system
to a rate that required no special precautions. Recent
developments, however, cast doubt on the validity of

this assumption. One aspect is the experience with the
MSRE, where a significant fraction of the tritium was
observed to diffuse through the secondary heat ex-
changer tubes into the coolant air. Another aspect is the
new emphasis on reducing releases of radioactivity from
any source to minimum practicable levels. Some modi-
fications in the MSBR reference design to deal with
tritium are to be anticipated, but what they will be
depends on the outcome of investigations currently
under way. The discussion which follows presents some
considerations that will be involved in specifying the
modifications.

3.3.7.2 Tritium in the MSRE. Disposal of tritium
produced in the MSRE was never a serious problem,
and for the first several years of operation the only
measurements were those necessary for health physics
monitoring of liquid wastes. Then, in 1969, with the
increasing awareness of the importance of tritium in
future molten-salt reactors, a campaign was launched to
Table 3.11. Distribution of heat produced by decay of fission products in a 1000-MW(e) MSBR

Noble gases are assumed to be stripped and noble metals to plate out on a 30-sec cycle;
other fission products are removed on a 50-day cycle

Heat Production (kW)
At shutdown 10°-sec (17-min) decay 10%-sec (2.8-hr) decay 10°-sec (28-hr) decay

Element Graphite Pr . Graphite Graphite Graphite

Salt Off-gas and °";n‘st‘“3 Salt Off-gas and Salt Off-gas  and Salt Offgas  and

metal P metal metal metal

Zn 0.00020 0.00020 0.00020 0.00013
Ga 0.26 0.033 0.0093 0.0047
Ge 1.8 13 0.56 0.083
As 45 5.5 2.8 0.14
Se 206 37 1.7 0.00023
Br 4,220 242 18 0.014
Kr 1,130 2,370 250 15 1,270 142 5.6 560 60 0.016 2.8 0
Rb 5,560 2,930 290 434 1,620 164 2.0 450 51 0.0058 1.0 0
Sr 4,270 374 40 1.5 1,660 376 38 1010 340 34 131 220 23
Y 4,750 140 17 170 2,640 140 16 1530 128 14 267 79 7
Zr 648 350 648 600 349
Nb 314 1,790 1790 318 592 760 760 950 212 212 406 97 97
Mo 69 835 835 12 460 460 0.043 188 188 0.026 145 145
Tc 25 1,240 1240 27 670 670 0.042 38 38 0.0051 30 30
Ru 2.5 160 160 0.24 144 144 0.097 126 126 0.0032 88 88
Rh 4.1 52 52 0.17 47 47 0.015 40 40 0.0038 33 33
Pd 2.0 0.4 0.4 1.3 0.4 0.4 0.4 0.3 0.3 0.12 0.1 0.1
Ag 14 0.1 0,1 0.2 9.1 0.1 0.1 6.8 0.1 0.1 1.8
cd 53 3.5 1.5 0.38
In 14 43 1.5 0.28
Sn 60 50 15 0.20
Sb 5,450 970 272 14
Te 510 1,970 1970 1,190 1,290 1290 153 233 233 3.2 78 78
I 4,510 2,120 2120 610 2,110 2110 356 1470 1470 48 745 745
Xe 1,080 2,770 414 42 730 125 52 340 56 28 180 22
Cs 4,000 2,600 383 200 1,700 265 5.2 98 20 2.5 8.1 0
Ba 4,030 490 58 1,450 480 52 342 210 24 230 96 10
La 4,620 470 50 68 3,030 470 50 1980 460 49 1380 450 46
Ce 1,260 3 0.5 154 650 3 0.5 558 3 05 375 3 0.5
Pr 1,740 492 1,150 413 230
Nd 213 25 173 11 80
Pm 150 12 116 108 72
Sm 10 0.3 7.9 5.7 3.8
Eu 38 0.04 3.6 34 2.9
Gd 0.055 0.047 0.043 0.017
Tb 0.0024 0.0018 0.0018 0.0016
Dy 0 0 0 0
Pa 500 5000 500 500 485

Total 49,400 20,300 9670 6590 16,500 12,300 6330 9010 4900 2610 4110 2260 (330

ORNL-DWG €9-42640

2
o
5
a 10
St
o "’E
=<
w 2 \\
T o2
s 0 —~—
@
[
>
|
2
/THERMAL SHIELD (7.62-cm Fe)
j@e— CONCRE TE ————————={
VOID
2 20
=
’—
5}
=
fal
8 —
n."’g \
= 3 AQ
«3 2
52 \
<
4;- GAMMA DOSE =124 !/hf\
{000 {040 1080 {120 4160

RADIUS (cm from core axis)

Fig. 3.22, Gamma and neutron heating in a radial plane
two-thirds of the distance from the midplane to the top of the
core of a 1000-MW(e) MSBR (R = 1000 to 1160 cm).

determine the distribution of tritium in the MSRE and
to compare it with calculated production rates.*?

The calculated production of tritium in the MSRE
fuel salt when the reactor was operating at 7.25 MW
with 223U fuel* amounted to 40 Ci/day. Of this, 35
Ci/day was from thermal-neutron absorptions in ®Li,
which comprised 0.0048% of the lithium, and 5 Ci/day
from fast-neutron reactions with 7Li. There was also
some production of tritium in lithium in the thermal
insulation around the reactor vessel. Because of the
large uncertainty in the lithium content of the par-
ticular batch of insulation that had been used in the
MSRE, the calculated production from this source
could be anywhere from 0.1 to 6 Ci/day.

Moisture condensed from the containment cell atmos-
phere had, since the beginning of power operation,
carried with it tritium which had been routinely
measured before disposal. Measured rates, which were
averages over collection periods of several months,

*With 2%5U fuel the fissile concentration was higher, the
thermal-neutron flux lower, and the tritium production rate 24
Ci/day,

41

ranged from 4 to 6 Ci/day. When the change to ?*3*U
was made, the change in tritium collection, if any, was
within the scatter of the measurements.

Tritium in the MSRE fuel off-gas at the exit from the
fission product absorbers was measured in November
and December 1969 at intervals through a 23-day
shutdown, a startup, and the final 16-day run at full
power. The tritium was collected by flowing the gas
through hot copper oxide and then trapping out water.
Experiments with the copper oxide at different tem-
peratures indicated that roughly half of the tritium was
present as hydrocarbons (presumably as a result of
exchange with hydrogen in oil vapors coming from the
fuel pump). Just before the shutdown, after more than
a month of operation at full power, the tritium effluent
in the off-gas was measured to be 23 Ci/day. Nineteen
days after the fuel was drained, the effluent rate was
still half as high, indicating tritium holdup somewhere
in the fuel or off-gas systems. During the final run,
several analyses showed tritium gradually building up in
the fuel off-gas over a two-week period, extrapolating
to between 25 and 30 Ci/day.

It had been recognized that tritium could diffuse in
atomic form through metal walls, and samples of the
off-gas from the MSRE coolant salt showed 0.6 Ci/day,
clearly more than the 0.0001 Ci/day calculated to be
produced in the coolant-salt system. Much more tritium
was found to be leaving the reactor in the air that had
passed over the coolant radiator. The concentration was
extremely low (<0.1 uCi/m?), and divergent results
were obtained by various methods of sampling and
analysis. The values thought to be most reliable fell at
around S Ci/day.

It thus appeared from the measurements that in the
MSRE about 60 to 70% of the calculated production in
the fuel salt eventually found its way out through the
fuel off-gas system. About 12 to 15% of the production
in the fuel diffused through the heat exchanger tubes,
and about nine-tenths of this went on out through the
radiator tubes into the cooling air. The uncertainty in
the production in the thermal insulation clouded the
interpretation of the tritium observed in the reactor
cell. The rate was 10 to 15% of the production in the
fuel, but the lack of measurable change when the
substitution of **3U nearly doubled the production in
the fuel strongly suggested that a large fraction prob-
ably originated in the insulation, The sum of the most
probable values of the measured effluent rates
amounted to only about 85% of the calculated total
production in the reactor. Although the probable errors
in the calculations and measurements amount to at least
this much, the comparison suggested the retention of
tritium somewhere in the system.
42

ORNL-DWG €9-1264¢

AN—

\
1] ~

NEUTRON HEAT PRODUCTION
(w/cm3)

~
4 AND & .
O S eiliin.
100 % SALT M—
PLENA 0.85 (0.64 om) REACTOR VESSEL
(7.62 cm) ) (7.62 cm)
CORE | on REFLECTO
A 0./32 VOLUME FRACTION SALT 0.37 = 0.01 vOoID W"'
'3; CURVE 4: SALT AND GRAPHITE
s 6 CURVE &: GRAPHITE ONLY e
CURVE €: SALT ONLY
E3 —
> '\ CURVE £: INOR
S \
-
E . F\\ Sed.
(o]
(vl
2 X b
b 4 X
114:1 2 ; \__,_ o
L= 4 C\
s A AND & ,\o
=
(L]
0 b
0 40 80 120 160 200 240 280 320 360 ' '435.96

AXIAL POSITION (ecm from core midplane)

Fig. 3.23. Neutron and gamma heating near the core axis of a 1000-MW(e) MSBR (R = 0 to 436 cm).

An attempt was made to determine whether one
could, with existing data, calculate a distribution of
tritium in the MSRE that agreed with the observed
distribution.** The calculations were based on con-
ventional mass transfer and diffusion equations and
made use of constants obtained from the technical
literature or calculated by conventional methods. They
indicated that of the tritium produced in the MSRE
fuel salt, up to 15% should come out of the radiator
tubes, more than 50% should leave in the fuel off-gas,
and up to 40% should appear in the reactor cell
atmosphere. This distribution was in reasonable agree-
ment with that observed, except for the much larger
fraction which would be expected to escape into the
cell atmosphere. The calculations further indicated that
in addition to the hydrocarbons deposited in the off-gas
system from fuel pump oil leakage, graphite in the core
and metal in the salt containers could have been
reservoirs for the tritium that was seen to persist after
shutdown.

3.3.7.3 Production and distribution in the MSBR.
Kerr and Perry®® estimated that a 1000-MW(e) MSBR
would produce a total of about 2420 Ci per full-power
day from the various sources shown in Table 3.12.

Using the same basic tritium behavior information
applied to the MSRE analysis, Briggs and Korsmeyer' '
calculated the tritium distribution in the reference
MSBR design, as shown in Table 3.13. These calcula-
tions assumed that shortly after birth the tritium would
form either *H, or tritium fluoride, > HF, The sparging
action of the helium bubbles used to strip xenon would
remove virtually all of the *HF but only a fraction of
the *H,. The cause of the different behavior is that
3H, which reaches a metal wall would readily dissociate
to form *H atoms, which can diffuse into the walls,
while *HF molecules would not dissociate. (There
would be some reaction of *HF with the metal to
release *H, but this was assumed to be negligible.) The
ratio of *H, to *HF would depend on the UF;/UF,
ratio in the fuel salt, assumed to be 0.001 in the
ORNL-DWG B89-12642

=
=) M'
5 !
a
O 4
& m
s \
33
£
z W
&
5 A'k
=
/THERMAL SHIELD (1045 ¢m Fe)
YO g f e CONCRETE——— ]
Al
1A
120 -
o
=]
—
o
==
B
' RO
<
=
E
u E
v \
3 a0
=
<t
s GAMMA DOSE = 71 mr/hr —_|
% N
AB
400 440 480 520 560 720

AXIAL POSITION (ecm from core midplane)

Fig 3.24, Neutron and gamma heating near the core axis of a
1000-MW(e) MSBR (R = 400 to 720 cm).

calculations reported in Table 3.13. A fraction of the
tritium from the fuel salt would pass through the pipe
and vessel walls to the reactor cell atmosphere, but a
major part would diffuse through the relatively large
area and thin walls of the tubes in the primary heat
exchanger into the secondary-salt system. Some of this
tritium would diffuse out through the walls into the
steam cell, a very small fraction would be carried out of
the coolant-salt loop with the cover gas, but the larger
proportion would dissociate and diffuse through the
steam generator tube walls to form tritiated water in
the heat-power system. In the calculations for Table
3.13 no account was taken of the resistance of the
oxide film on the water side of the heat exchanger
tubes. Some data indicate that this resistance should
appreciably reduce the transfer to the steam system,
which tends to make the rate in Table 3.13 a
conservatively high estimate.

3.3.7.4 Concentrations and release rates. The steady-
state tritium concentration that is reached in the steam
system is the ratio of the tritium infusion rate to the

43

ORNL— DWG 63~-6006

317 yeurs™-

o2l
~0

o* w0
ELAPSED TIME (sec)
Fig. 3.25. Afterheat distribution with saturation concentra-

tion of fission products in a 1000-MW(e) single-fluid MSBR
fueled with 235U,

Curve A. Afterheat in core region produced by Kr and Xe
diffused into the graphite plus heating by 10% of the total
noble metal fission products assumed to be plated on surfaces.

o ud o)

Cutve B. Afterheat in the four heat exchangers produced by
40% of total noble metal fission products plated on metal
surfaces.

Curve C, Afterheat in the chemical processing system produced
by protactinium and long-lived fission products.

Curve D. Afterheat in the off-gas system produced by Kr and
Xe, plus heating by 50% of the total noble metal fission
products.

Curve E. Afterheat produced by fission products which remain
dispersed in the primary salt.

Curve F. The sum of all curves, A through E.

In curve A the concentration of Kr + Xe is that which
produces a poison fraction of 0.0056 §k/k and is obtained by
gas sparging on a 30-sec removal cycle, Curves A, B, and D are
based on the assumption that the noble metals are either
deposited immediately on metal and graphite surfaces or enter
the off-gas system immediately. In curves A, B, and D the
afterheat includes that from decay of the daughter products of
the noble metals and gases.

rate of water discharge from the system (leaks, blow-
down, and sampling streams). A reasonable estimate for
44

Table 3,12, Rates of tritium production in the

MSBR at 2250 MW(1)
Production
(Ci/day)
Ternary fission 31
bLi(n,)H 1210
"Lin,an)*H 1170
'9F(n, 1 70)3H 9
2420

Source: ref, 45,

the water discharge rate is 1% of the 2.1 X 10° Ib of
water in the system per hour. Assuming that 1670
Ci/day does enter the system, the tritium concentration
would level off in about two weeks of full-power
operation at 7 uCi of *H per gram of water.

In the current Standards for Protection against
Radiation,*® the maximum permissible concentration
of tritium in water for 40 hr/week occupational
exposure is 0.1 wCi/ml. Thus, if the tritium in the
MSBR steam is anywhere near as high as the 7 uCi/g
calculated, means would have to be taken to limit
exposure of plant operators. These measures would not
have to be nearly as elaborate as those required around
some heavy-water reactors, where tritium concen-
trations are more than 10 times that predicted for the
MSBR steam,* 7 but the precautions in the MSBR steam
plant would certainly include tritium monitors, good
ventilation of work areas, restrictions on handling
discharged water, and possibly use of masks in working
on steam leaks. (Air saturated at 100°F with vapor
from the steam system would contain 3 X 107 uCi of
3H per cubic centimeter, or 70 times the MPC for air
for 40 hr/week exposure.)*®

It would be convenient if the water bled from the
MSBR steam system could be released by simply mixing
it with the ~440,000 gpm of condenser cooling water
effluent. If 1670 Ci/day were being discharged, the
concentration in this stream would be 0.7 X 1073
uCi/ml. This is less than the 3 X 1072 uCi/ml currently
specified as the MPC for water discharged to an
unrestricted area.*®

It thus appears that even if the conservatively high
estimate of tritium transfer to the steam system were
correct, the concentration in the MSBR steam would
not seriously hamper plant operation and maintenance,
and the plant effluent would meet the current standards
for release to unrestricted areas. Expert reviews of the
biological effects of tritium lead to the conclusion that

Table 3.13. Calculated distribution of the tritium
produced in the reference MSBR design

Rate

Percent of Curies/day
total >°H at
production 2250 MW(1)

Removed from primary system with

sparge gas
As °H, 5.8 140
As THF 7.0 170
Entering secondary system cover gas 0.1 2
Entering reactor cell atmosphere 8.7 211
Entering steam cell atmosphere 94 227
Entering steam-power system 69.0 1670

100.0 2420

Source: ref, 11.

the currently specified maximum permissible concen-
trations are conservative and limit increased dose to the
population to a negligible fraction of background.*®
Nevertheless, it would be quite unrealistic to assume
that the reference design of the MSBR is satisfactory
with regard to tritium control. Release of a curie of
tritium per megawatt-day of electricity from an MSBR
plant will not be tolerated, especially since other
reactors and fuel-reprocessing plants release far less.
Fortunately, there appear to be several practical ways
to ensure that the tritium release from an MSBR is far
below the values listed in Table 3.13. These are
discussed briefly in Sect. 16.4 of this report.

3.4 THERMAL AND HYDRAULIC DESIGN OF CORE
AND REFLECTOR

W. K. Furlong  H. A. McLain
3.4.1 Core

A basic objective for the thermal and hydraulic design
of the core is to regulate the salt to achieve a uniform
temperature rise of the salt flowing through each of the
channels. From plenum to plenum, this rise is set at
250°F. There are other important factors, however,
which must be minimized or kept within allowable
limits, such as the fuel-salt inventory, the pressure drop
due to flow, the graphite temperatures, and the vessel
wall temperatures.

Neutron-induced volume changes in the graphite are
sensitive to temperature, as discussed in Sect, 3.2.3;
thus the temperatures should be minimized in the
regions of high damage-neutron flux (£ > 50 keV) if
the design goal of a four-year graphite life is to be
achieved. Figure 3.10 gives a graphical representation of
the graphite volume changes as a function of fluence,
with temperature as a parameter. The minimum graph-
ite temperature is set by the salt temperature at its
boundary. However, the graphite is heated internally by
neutron scattering and absorption of gamma radiation,
raising its temperature above the salt datum and making
it dependent upon the film heat transfer coefficient as
well.

The gamma and neutron heating has been calculated
from transport theory, as reported in Sect. 3.3.5. The
radial variation of fission power density, which governs
the radial flow distribution, is shown in Fig. 3.26. The
discontinuity in the curve is between zone I, having
13.2% salt by volume, and zone II, having 37 vol %
(see Table 3.3 for definition of zones). For the purpose
of temperature calculations, the axial power density
variation in zone [ was approximated by a cosine
function of the form

Q= Qpay cos (m2/H) ,

where z is the distance from the midplane and H is an
extrapolated height of 16.2 ft. (The actual design
height, excluding reflectors, is 15 ft.)

45

The choice of prismatic moderator elements with a
central hole was based on a combination of neutronic
and heat transfer considerations. Two alternatives con-
sidered were tangent solid cylinders and spheres. The
cylinders have a less-than-optimum salt fraction of
about 9%. An objection to this geometry is the cusp
formed near the region of contact; the relatively poor
heat transfer in this area could be a problem at the
power densities used in the present design. Also, the
cylinders have only line contact, and the possibility
exists for misalignment or bridging, particularly after
dimensional changes. Spheres which are randomly
packed have a 37% void space, This would give a salt
fraction far too great for the major portion of the core.
Use of two different sphere sizes would reduce the void
fraction closer to the value needed in zone I for
optimum breeding performance, but pressure drop
considerations made this approach questionable. The
37% void space in the spheres would, however, be about
optimum for the undermoderated portion, or “blanket™
region. The graphite balls would require some sort of
barrier to contain them, however, and the spheres did
not appear to offer any particular advantages over the
graphite element design selected for the under-
moderated region.

ORNL-DWG 62-6010

40
35 l ANNULUS o g O
{00 vol 7 SALT
o
S 30 S, (R———
"
5
S e
= — 1 37wl % LT~ -y
> ZONE 1 o | N
= 3.2 vol % SALT
2 ) - & e
w 20 [— —1 3
a
x
w
3 15 - - \ j
a
z
o
9 1o A VS e
s
5 — . = L e
0
0 50 {00 150 200 250
RADIUS (cm)

Fig. 3.26. Radial distribution of fission power density averaged over length of MSBR core and both axial reflectors,
As a result of the above considerations, the selected
moderator element consists of a long prism with a 4-in.
square cross section containing a central hole. Ribs on
the faces separate adjacent elements and form inter-
stitial salt flow channels. The geometry of the cross
section is a2 compromise hetween the neutronic, heat
transfer, and fabrication considerations. In zone I it is
desirable from a nuclear viewpoint to have a more
heterogeneous cell (larger dimensions), but the con-
trolling consideration is heat conduction out of the
graphite. In zone Il the neutronics favor a smaller
element, but buckling and vibration impose a lower
limit. Although not an optimum dimension, the 4-in.
square appeared to be the best compromise.

The optimized physics calculations indicated that the
volume fraction of salt in zones | and II should be
0.132 and 0.37 respectively. These fractions are ob-
tained by adjusting either the diameter of the center
hole or the rib size (which alters the interstitial channel
size). Minimum dimensions on both the hole and the
ribs are influenced by fabrication considerations. Spe-
cifically, to achieve relatively low costs of fabrication
by the extrusion method will require that the element
geometry contain no radii of less than about 0.25 in.
Also, it is believed that the center hole diameter should

————‘4.302———“—1

[, 000

s —
I‘—t.305——]‘—1.565

I ‘
~

o

o)

\g

o

—w= = 0.302
£
r___\J
/ —— | BAY e
/\ \/\
ALL DIMENSIONS
IN INCHES

(a)

46

not be less than about 0.6 in. to assure successful
deposition of the pyrolytic graphite coating on the
graphite surfaces.

The graphite moderator elements are shown in Figs.
3.4 and 3.5. The central part of the core, zone I-A, will
be comprised of elemenis of e lype shown in Fig.
3.27, while those at a larger radius (lower power
density) will be of the type shown in part b of the
figure and are designated as zone I-B, The salt fraction
is 0.132 in both zones I-A and I-B, but the interstitial
channels have been made smaller and the central hole
larger in zone [-B. The purpose of this arrangement is to
achieve flow control by orificing only the central hole
rather than by complicating the design with orifices for
the interstitial channels as well. The calculations in-
dicate that in the present design the average tempera-
ture rise through each flow channel approximates
250°F. For a given moderator element near the reactor
center line the temperature rise for the salt flowing
through the hole is essentially the same as that flowing
through an interstitial passage; away from the center
line the temperature rise through the hole is greater
than 250°F and that in the interstitial channel is less
than 250°F. The orificing for the central holes will be
designed so that the salt streams discharging from all

ORNL-DWG 689-601

4.100

| ——

3.900

(428 —o=fee——1.,728 —---—‘ l

|
xL

-
L~

(&)

Fig. 3.27. Graphite moderator element for (¢) zone I-A and (b) zone I-B of reactor core.
flow channels associated with a given element will
combine to give a bulk temperature of 250°F above the
inlet value.

The elements for core zone II-A are prismatic and are
shown in Fig. 3.5. They are identical to the elements
used in zone I-B except that the central hole diameter is
2.581 in. to obtain the 0,37 salt fraction needed in the
undermoderated region. The elements for zone 11-B are
in the form of rectangular slats spaced far enough apart
to provide the 0.37 salt fraction. As shown in Fig. 3.3,
the slats are separated by pins and elliptical rods. The
latter are intended to minimize the cross flow which
would otherwise occur from zones I and IT into the
annulus due to the annulus being orificed at the bottom
and operating at a lower pressure than the core and
reflector regions. (The annulus was orificed in this
manner so that the salt flow will be predominantly
radially inward through the radial reflector, as will be
described subsequently.)

Since the center of zone [ is the region of highest
power and greatest flow requirements, if all the flow
channels at that location could have equal hydraulic
diameters, the pressure drop through the core could be
designed to be a minimum value. Unfortunately, the
restriction on the minimum hole size through the
elements, mentioned above, dictates that the hole have
a larger hydraulic diameter than the interstitial channels
and that orifices be used for the holes. The penalty is

47

not a great one, however, since the total pressure drop
across the core at rated flow is estimated at only 18 psi.
In this connection it may be noted that experiments
have been reported*® in which the flow through
channels formed by closely packed rods on a triangular
pitch is greater than that predicted by the equivalent
hydraulic diameter theory. Further study, and probably
model testing, will be required to verify the calcula-
tions, particularly with regard to the passages formed
by the corners of four adjacent core elements.

The flow divisions and various flow paths through the
reactor are shown schematically in Fig. 3.28. The salt
volumes and approximate power generation for each
region are also shown. The dashed lines in the figure
indicate lines of minimal flow, that is, paths for which
flow is purposely minimized by orificing or for which it
is unavoidable due to clearances. From Fig. 3.28 it may
be noted that there are three major flow paths: (1)
through zones I and II, where the bulk of the power is
generated, (2) between the vessel and reflectors and
through the radial reflector pieces to the annulus, and
(3) through the control rod region and lifting-rod holes,
The flow and temperature aspects will receive further
discussion in the sections that follow.

Peak and average steady-state temperatures in the
central moderator elements were investigated using the
HEATING code.’® This is based on the relaxation
method and employs constant thermal conductivity.

ORNL-DWG 70-11913

LOWER AXIAL ZONE I UPPER AXIAL
ZONE 11 ONE ZONE I
Q=51 Mw Q= 167& N 0= 51 Mw
V=467 3 V=290t v=467 1>
RADIAL ZONE
n UPPER PLENUM
INLET LOWER PLENUM O0=283 Mw 0=~30.5Mw [
— Q=~0 . |—= 0=30.5Mw V=288 i3 V=3571t3
V=831t V=35 ft3
[ UPPER VESSEL COOLANT PASSAGE J
AND
ng‘g,‘g‘i,L UPPER AXIAL REFLECTOR
Q= 4.9 Mw,
LOWER AXIAL \ V=15t
REFLECTOR ==
Q=49 Mw
=™ V=g ft3 LIFTING ROD —T
HOLES
OUTLET
et e i s i o A= g - WP g e e ] PASSAGE
Q=~0_
LOWER VESSEL RADIAL VESSEL RADIAL v=424 ft
COOL. PASS. COOL. PASS, REFLECTOR ANNULUS
Q=~0 Q=~0 Q=1(9.7 Mw Q=765 My
V=82 ft3 V=677 1" V=264113 V=135t

Fig. 3.28. Salt flow paths for MSBR core and reflectors.
The center of the core is the region of maximum
damage flux, but the maximum element centerline
temperature occurs at an axial position a few feet above
the midplane, as determined from a heat balance with
appropriate integration of the axial power density
variation. The worst combination of damage flux and
temperature, which will result in minimum graphite life,
is found to occur about 1 ft above the midplane and
along the center line of the core.®" Figure 3.29 shows
the results of the temperature calculations at the
midplane and at a plane 1 ft higher. The significant
input parameters used in the calculations are listed in
Table 3.14. The heat transfer coefficients were based on
the Dittus-Boelter correlation. Recent investigations at
ORNLS? indicate that in the range of Reynolds
numbers of interest, heat transfer coefficients for the
fuel salt are slightly lower (about 20%) than those
predicted by the correlation used in the MSBR con-
ceptual study, Even if the lower values are used,
however, it should not make any significant change in
the temperatures reported here, since the graphite itself
is the major resistance to heat transfer. The effects of
vertical flow and entrained gas on the heat transfer
coefficient remain to be investigated. It was assumed in
the calculations that the effect of volumetric heat
sources on heat transfer between graphite and salt was

48

Table 3.14, Input parameters for calculating MSBR
moderator element temperatures using the

HEATING code”
At At 1 ft
3 above
midplane . 3
midpiane
Salt temperature, °F 1175 1200
Heat generation rate, Btu hr ' in. 290.8 286.1
Graphite thermal conductivity, 1.415 1.415
Btuhr ' in.”! CF)?
Heat transfer coefficient for center 12,26 12.63
hole, Btu hr ' in. ™2 °F) !
Heat transfer coefficient for outer 12.85 13.22

surface, Btuhr ¥ in. 2 CF) !

AAt near the reactor center line, where the temperature rise
through the holes and through the interstitial passages is
essentially the same. Further out from the center line the rise is
not equal.

negligible and that there was no heat transfer between
graphite and salt for a distance of 0.1 in. on either side
of the apex of the ribs on the outer edge of the
moderator elements, The latter assumption is a first
approximation to account for the restricted flow in that
area.

ORNL-DWG €9-6042

(a)

(£)

Fig. 3.29. Temperature distribution in graphite moderator element at (¢) midplane of core and (b) 1 ft above midplane.
Temperatures have not yet been investigated in the
moderator elements at radial positions other than at the
center of the core, nor have they been examined in
zone II. In these areas of lower radial power density and
consequently lower salt flow rates, the heat transfer
coefficients will be less. However, the heat sources
within the graphite are also reduced, as is the damage
flux. Although a more detailed analysis may indicate
higher peak graphite temperatures at locations other
than those investigated, the reduction in damage flux is
expected to be more than compensating. On the basis
of the data presented in Sect. 3.2.3 on damage flux and
graphite life, the MSBR graphite will achieve the design
objective of a four-year life at the temperatures which
would exist in the reference design.

Preliminary calculations indicate that vibration of the
moderator elements should not be a problem. The
magnitude of the vibrations was determined by extrapo-
lating known information about the amplitude of rod
vibrations associated with parallel flow*? and adding to
this the rod deflection due to cross flow of salt between
the channels. Assuming the velocity of the salt between
adjacent channels to be % fps and extrapolating
information on vibration due to cross-flow vortex
shedding,’* the sum of the two effects gives a total
calculated amplitude of vibration at the center line of
less than 0.002 in. Model tests will be required for
substantiation, but on this basis it is believed that core
vibrations will not limit the design parameters.

It may be noted that a 12- by 12-in. area has been
assigned for control rods in the center of the reactor.
The salt flow in this region will be in excess of that
needed to cool the rods in order to bring sufficient cool
salt to the top axial reflector. Orificing of the flow in
this central region will also be required to limit
variations in the flow as a function of control rod
position.

3.4.2 Radial Reflector

Determination of reflector temperatures is important
because of their relationship to graphite life, amount
and temperature of coolant required, and stored energy
during afterheat removal, The relationship between life,
damage flux, and temperature is shown in Fig. 3.10.
For a pgiven nuclear design there is a maximum
allowable temperature for any reflector section which is
intended to remain fixed in position for the design life
of the reactor. Conversely, a temperature distribution
calculated for given reflector geometry and coolant
conditions may dictate a reduction in the incident
damage flux, even though this entails a departure from
optimum nuclear conditions. The amount and tempera-

49

ture of coolant are interdependent. The major part of
the coolant temperature rise is due to its own internal
fission heating, and it is desirable to have each unit
volume of salt experience the same plenum-to-plenum
temperature rise. On the other hand, the need for
improved heat transfer coefficients or lower sink
temperatures may dictate a higher flow rate than that
required to attain this rise,

A reflector design using graphite blocks averaging
about 1 ft* was rejected when analysis indicated
excessive temperatures. The principal cause was fission
heat from trapped interstitial salt. This heat had to be
transferred to a cooled surface by conduction, which
required large temperature gradients. A conclusion was
that regions of static salt must be avoided everywhere
within the reactor vessel. Without the presence of
internal fission heat, the sources in the reflectors consist
primarily of photons leaking from adjacent blanket
regions and from neutron slowing down. These sources
are shown in Fig. 3.30.

The present radial reflector design, shown in Figs.
3.1-3.3, has been analyzed using the HEATING
code.”® Boundary temperatures were based on the fluid
temperature required at a given location for an overall
250°F rise and also considered surface temperatures
due to the volumetric heat source in the fluid. The
volume fraction of salt in the reference design reflector
is about 1%, but as long as the salt is flowing this
quantity is not important to the temperature distri-
bution estimates in that heat generation within the salt
is carried away by the salt and the fission heating in the
salt far exceeds the heat transferred into it from the
graphite. Hence the conduction problems have been
treated with fixed boundary conditions rather than
having to couple the salt and graphite by an energy
balance. Heat transfer coefficients were based on
laminar flow of fluid between graphite segments and
between reflector and vessel and on turbulent flow of
the fluid at the reflector-blanket boundary in the
2-in,-wide annular space between the reflector and the
removable core assembly. Resulting temperatures at the
axial midplane are shown in Fig. 3.31. This is about the
location of the peak damage flux, which has been
constrained to about 4 X 10'? (£ > 50 keV) to achieve
the 30-year design life at the calculated 1250°F surface
temperature. The decrease of damage flux with distance
into the reflector overrides the effect on graphite life of
increasing temperature near the edge of the reflector.

In order to meet the heat-removal requirements and
the other objectives mentioned above, the flow of salt
through the reflector graphite must be in the radial
direction rather than vertically upward, as it is in the
ORNL=-DWG 869-6013

X (cm) =10

(7] (w/t:m3 of graphite)

“@=0MT712 exp (-0030323 x)
\T=X =38
\.

\\
L hd
- —— )
04
Q [e] 20 30 40 50 60

DISTANCE FROM INNER FACE OF REFLECTOR (cm)

Fig. 3.30. Heat sources in graphite radial reflector at mid-
plane.

core. In large part this is due to the fact that the
thermal coefficient of expansion of Hastelloy N is
greater than that of the graphite, The reflector graphite
could be restrained into essentially the room-tempera-
ture geometry with little change in the flow channel
geometry, but the expanding vessel would draw away
from the reflector and increase the salt volume in the
annulus between the vessel wall and the graphite. This
would result in an undesirable increase in the primary-
salt inventory. It was therefore decided to restrain the
reflector graphite to maintain its position relative to the
wall and let the flow passages in the graphite open up as
the system is brought up to temperature. With an
increase in the width of the flow channels in the
reflector graphite, axial flow passages for the reflector
are not fixed. Connecting the reflector flow passages,
the annular space at the vessel wall, and the annular
space between the reflector and the removable core to

50

ORNL -DWG 69-6014

1400 - ~
¥ yaso|— - e -
w \\
x N
F {4300 —{f - N -
- |
§ ( | \\ T5iNKk =, 1200 * R
3 t2s0|—t1—— | .
TNk = ’20? "F ~
1200—2= 1 - L
0 05 10 {5 20 25 30

RADIAL DISTANCE THROUGH REFLECTOR (f{t)

Fig. 3.31. Temperature distribution in graphite radial reflec-
tor at midplane based on heat sources shown in Fig. 3.30.

common plenums located at the upper and lower ends
of the reflector is not satisfactory with axial flow.
There would be inadequate axial flow through the
reflector if the pressure difference was limited to the
amount necessary to get the desired temperature rise
for the salt flow through the annular space. On the
other hand, there would be excessive salt flow through
the annular spaces if this pressure difference was
increased to get the necessary flow through the re-
flector region. However, the use of radial flow circum-
vents these design difficulties.

The salt in the reflector flows inward toward the core
in order to minimize the vessel wall temperature and
because of orificing considerations. The annulus be-
tween the core and reflector is orificed at the bottom
because of mechanical assembly considerations and
because this annulus serves as the collection plenum for
the radial flow through the reflector. Salt flow from the
undermoderated region of the core into the annulus is
restricted by graphite rib seals located between the
graphite slabs in the undermoderated region, zone Il.
Axial distribution in the radial flow through the
reflector is controlled by orifices located at the inlets of
the radial flow passages.

3.4.3 Axial Reflectors

The axial reflectors are subjected to a 66% higher
peak damage flux than the radial reflector. However,
the lower one is replaced with the moderator, and the
upper one must last only half of design life due to the
alternate use of the two heads. Hence, temperature and
damage flux considerations are not as stringent as in the
radial reflector. The heating rate in the upper axial
reflector was analyzed using the HEATING code.?°
The axial behavior of the source is shown in Fig. 3.32.
The radial variation was described by a cosine. The
ORNL-+ DWG 69~ 60156

20
fo
\(_,.-o=|.375529 exp(—0.0812817X) 0= X(em) < 12
1.0 \ =1 ¥ =
\\ | s
K \ i =
= —l e
[» 9
e
505 —\
o
" L el
5 N\ LO=A+8X+CX° 125 X<860
£
BF ¢ Blcod gl i _'<
.\
02 = ¥.<_~
'\._
(sX
0 0 20 30 40 50 60

DISTANCE THROUGH AXIAL REFLECTOR (cm)

Fig. 3.32. Gamma and neutron heating in graphite axial
reflectors.

inner face was subjected to 1300°F salt, while the other
faces were in contact with somewhat cooler salt, which
is transported from the reactor inlet via the control
region and lifting-rod holes, to provide a low-tempera-
ture fluid coolant sink for the vessel head. On the above
basis, the peak temperature was found to be 1363°F,
and the surface temperature in the region of peak
damage flux was 1265°F.

3.5 REACTOR VESSEL DESIGN

3.5.1 Reactor Vessel Description
E. S. Bettis

The basic features of the reactor vessel are shown in
Figs. 3.1 and 3.2. The vessel has an inside diameter of
22.2 ft, an overall height at the center line of about 20
ft, a wall thickness of 2 in., and a head thickness of 3
in. Major considerations in the design of the vessel
were:

. The core must be replaceable without undue dif-
ficulty.

2. The holdup of fuel salt in nozzles, plenums; and
other volumes exterior to the core must be a
minimum.

3. The vessel walls and heads must be protected from
excessive temperatures and radiation damage.

51

4. The vessel must be designed for 75 psig and a wall
temperature of 1300°F and must meet ASME code
requirements for nuclear vessels.>®

5. The vessel must be constructed entirely of modified
Hastelloy N.

The reactor vessel is constructed of the following major

pieces:

1. A cylindrical section, 22.5 ft OD X ~I13 ft high,
with a wall thickness of 2 in.

A transition section, about 4 ft high, with one end
having a diameter of about L8 ft and the other 22.5
ft. This section has four symmetrically spaced salt
outlet nozzles and radial gusset plates attached to if.
The wall thickness is 2 in.

Two cylindrical sections about 13% ft high with
2-in.-thick walls. One has an inside diameter of 18 ft
and the other an outside diameter of slightly less
than 18 ft, so that one fits inside the other, as shown
in Fig. 3.2. Forged flanges at the top provide the
vessel closure.

One upper and one lower dished head, each 3 in.
thick. The upper head is about 18 ft in diameter and
the lower about 22Y% ft.

With the exception of the flanged closure at the top,
the vessel is of all-welded construction, fabricated of
modified Hastelloy N having the physical properties
listed in Table S.1 and discussed in Sect. 3.2 4.

The design requirement for core replaceability led to
adoption of the cylindrical extension on the vessel and
top head which permits the closure flange to be located
in a relatively lower temperature region and one with
greatly reduced radiation intensity. The flange face is
about 6 in. wide and is machined for two metal ring
gaskets. The space between the two rings will be
continuously evacuated and monitored for fission gases.
The flanges are joined by a clamp which encircles the
outside of the flange and extends upward to the
operating floor level. Thirty-four 1-in. bolts in this
clamp are easily accessible and supply the force which is
transmitted to the flange faces for making the closure.
It may be noted that the weight of the upper layer of
roof plugs rests on the upper flange and reduces the
bolt tension required to maintain the gasket loading.

The transition section was adopted to conserve
fuel-salt inventory in the region of the outlet salt
nozzles and to minimize the diameter of the top head
assembly to be handled during core replacements. The
necking in of the vessel at the top prevents top loading
of the last row of reflector graphite and requires a
special shape for two of the blocks, as discussed in Sect.
3.1.2. The transition section also serves as a collection
header for the fuel salt leaving the top of the reactor,
diverting it into the four exit nozzles. These nozzles are
of a special shape, elliptical in cross section at the vessel
end and cylindrical in cross section where joined to the
fucisalt pipiig i€ading v e pump iniet. Reinforcing
webs are used in the construction of the outlet nozzle
to provide needed strength.

The cylindrical portion of the vessel is fabricated of
rolled plate, rough machined after heat treatment. The
roundness tolerance is probably about +', in. The
dished top head has a forged ring welded around its
circumference for joining it to the upper cylindrical
extension. The maximum thickness of the ring is about
4 in.

The fuel-salt inlet is at the center of the bottom head.
The inlet plenum is a well about 3 ft in diameter and 4
ft high at the center line of the vessel. The four
16-in.-diam fuelsalt pipes enter symmetrically around
this well, as indicated in Fig. 3.1. The 6-in. drain
connection is to a nozzle in the bottom head of the
well. Hastelloy N flow diverters, or turning vanes, are
provided in the plenum to direct the salt flow upward
and to reduce the turbulence in the reactor vessel inlet
nozzle.

The top head of the vessel has an 18-in.-diam nozzle
at the center line for the pipe containing the control
rod assembly. The cylindrical extension of the top head
is provided with lifting lugs into which the spider
carried by the hoisting machine engages to lift the
reactor core assembly from the vessel, as described in
Sects. 3.1.2 and 12.3.

3.5.2 Reactor Vessel Temperatures

W. K. Furlong

The reactor vessel will be heated above the 1000°F
ambient cell temperature by the hot molten salt flowing
on the inside and by neutron and gamma absorptions.
The maximum metal temperature and the temperature
distribution are important because they affect the
calculated and design stress intensities in the walls,
heads, and nozzles.

An analysis of the 2-in.-thick cylindrical wall in-
dicated that the peak metal temperature would be
about 69°F above the interior salt temperature and
would occur close to the outside surface at about
midheight. In making this study it was assumed that the
salt temperature at the inside face was uniform at
1100°F.* A similar study of the 3-in.-thick upper head
gave peak temperatures 20 to 80°F above the inside salt
{emperature (again assumed as 1100°F), also occurring

52

on the outside surface. The lower head has less incident
gamma flux due to the shielding provided by the
internal structures and is cooled by salt closer to the
1050°F inlet salt temperature and thus will operate
somewhat cooler than the upper head.

The caiculated stress intensities n the walls and upper
head are generally within the allowable, or design,
intensity range, since the salt sweeping the inside
surfaces is a bypass stream taken from the reactor inlet
and should not significantly exceed the assumed average
of 1100°F. However, if the metal were bathed by salt
closer to the reactor outlet temperature of 1300°F, it is
possible that some metal temperatures would be unac-
ceptably high in that the allowable, or design, stress
intensity would have to be revised downward. The
vessel has not been designed or analyzed in detail, but it
is considered a possibility that further study would
disclose localized areas, such as the outlet nozzles or the
junction of the top dished head with the cylindrical
portion (where stresses tend to be high), which would
have to be shielded from the flow of hottest salt.
Although the lower head is larger in diameter than the
upper head and thus would have higher stress intensities
in withstanding the internal pressure, the temperature is
sufficiently low to keep the stress intensities in this part
of the vessel within the acceptable range.

3.5.3 Reactor Vessel Stresses
C. W. Collins

A preliminary elastic stress analysis was made for the
reactor vessel using an Air Force computer program®®
which has been modified by ORNL. The analysis was
based on the top of the vessel operating at 1300°F and
42 psig and the bottom at 1100°F and 61 psig. The
maximum stress in the removable head due Lo pressure
alone is 5220 psi. This stress is located in the dished
head near the junction of the head and shell skirt. The
maximum stress in the vessel occurs at the junction of
the lower head and shell and is 16,324 psi. The
cylindrical portions of the vessel are 2 in. thick, and the
dished heads are 3 in. thick.

No analytical work has been done on the nozzles,
closure ~flanges, thermal stresses, or discontinuity
stresses at the necked-down portion of the vessel

*1t is reasonable to assume a 1100°F salt temperature in the
vessel wall coolant passage since the flow through the reflector
is radially inward. The analyses assumed laminar flow of salt
and a heat transfer coefficient of 137 Btu hr ' ft > (°F)~.
Heat transfer from the reactor vessel to the cell environment
was neglected, as was the effect of gamma irradiation from the
primary heat exchangers.
because of the large amount of time that would be
required to develop computer programs. As an al-
lowance for the uncertainty, the stresses were held well
below those allowed by the ASME Boiler and Pressure
Vessel Code for standard Hastelloy N, As described in
Sect. 3.2.4, experimental heats of modified Hastelloy N
are stronger than the standard alloy, and the alloy that
will be used in the MSBR will probably be approved for
higher stresses than the standard alloy. Neutron irradia-
tion to the extent anticipated in the MSBR should not
require a reduction in allowable stress, The graphite
reflector is sufficiently thick to reduce the 30-year
integrated neuntron dose (>300 keV) at the wall to
below 1 X 10*' neutrons/cm?®. At this fluence the
reduction in metal strength is insignificant.

As stated in Sect. 3.2, standard Hastelloy N is
approved for use under Sects. 111 (ref. 56) and VIII (ref.
57) of the ASME Boiler and Pressure Vessel Code. The
design stresses applicable for nuclear vessels at tempera-
tures up to 1300°F were determined through the
following case interpretations.

Case 1315-3 (ref. 30) approves use of Hastelloy N for
pressure vessels constructed in accordance with provi-
sions of Sect. VIII, Division 1. Allowable stresses are
given for temperatures to 1300°F.

Case 1345-1 (ref. 31) approves use of Hastelloy N for
class A vessels constructed in accordance with provi-
sions of Sect. [l of the Code. Design stress intensity
values are provided only to 800°F, in common with
other materials approved for use under Sect. 111 (ref.
56).

Case 1331-4 (ref. 58) provides rules for construction
of class A nuclear vessels that are to operate at
temperatures above those provided for in Sect. 111 (ref.
56). It permits the use of allowable stresses from
Division 1 of Sect. VIII (ref. 57) and the related Code
Case 1315-3 (ref. 30).

In applying these Code cases, it is found that the
allowable primary stress intensity (S,,) is 3500 psi at
1300°F and 13,000 psi at 1100°F. At the juncture of
the heads and shells, where the maximum stresses
occur, paragraph 5 of Case 1331-4 (ref. 58) establishes
the allowable value of the primary plus secondary stress
intensity as three times the allowable design stress
intensity (S,,) for the metal temperature involved. On
this basis, the allowable stress intensity at 1100°F is
39,000 psi and at 1300°F is 10,500 psi. Stresses in the
preliminary design of the vessel have been held well
below these allowable values.

From these preliminary calculations it appears that
the critical stress regions are at the junction of the head
and shell in the removable head and, most particularly,

53

at the outlet nozzles where the highest temperature
occurs and for which no analysis has been attempted,
When a more rigorous analysis is completed, it may be
found necessary to add a thermal barrier in this region
with cooling from the inlet salt stream or to alter the
vessel design in this region to reduce the discontinuity
stresses.

3.6 PRIMARY SYSTEM SALT PIPING
C. W. Collins

Because of the fuel inventory costs, a prime consider-
ation in the design of the pnimary system piping was to
limit the piping volume to the minimum permitted by
reasonable pressure drop and by required piping flex-
ibility. The piping must accommodate the expansion
associated with the high operating temperatures of
1050 to 1300°F. To provide needed flexibility and low
fuel-salt inventory, the fuel-salt piping must probably
be limited to 16 to 20 in. in diameter.

The support scheme for the primary loop is based
upon anchoring the reactor vessel to the concrete
building structure while the other components are
mounted on flexible supports. The pumps, heat ex-
changer, and piping are positioned radially around the
reactor vessel, with essentially the only restraint being
the vertical support by hangers mounted to the roof
structure, thus allowing the components to move freely
without developing excessive piping stresses. The layout
of the primary-salt loop is shown in Figs. 13.7 and 13.8.

The piping system was analyzed at operating tempera-
tures using the MEL-21 “Piping Flexibility Analysis™*”
computer program. It was determined that the piping
meets the requirements of USAS B31.7 “Tentative USA
Standard for Nuclear Power Piping”®? for stresses due
to thermal expansion, weight, and pressure loading of
the system under the operating conditions. The analysis
is incomplete in that no off-design conditions were
considered, nor were uany localized thermal or dis-
continuity stresses taken into account. This would have
involved considerably more effort than was warranted
for this conceptual design study.

The maximum computed expansion stress was 5570
psi, occurring at the point where the pump discharge
pipe connects to the heat exchanger. ASME Code Case
1331.4 (ref. 58) establishes the allowable value of the
primary plus secondary stress intensity as the larger of
three times the allowable design stress intensity (S,,)
or, as an alternate, three times the allowable stress
amplitude (S,) at 10° cycles for the metal temperature
involved. The allowable stress intensities at 1300°F are
thus 10,500 psi, based on 35,,,, or 19,500 psi, based on
3S,, the latter establishing the allowable primary plus
secondary stress intensity. When the ~1500-psi stress
due to pressure is added to the maximum expansion
stress of 5570 psi, the allowable primary plus secondary
stress intensity is not exceeded.

The primary loop is designed to be flexible enough ta
accommodate the large thermal expansions due to the
relatively high operating temperatures, This flexibility
must be controlled during an earthquake or after an
accidental break in the piping that tends to cause
whipping or other movement. Light-water reactors use
spring supports and hydraulic dashpots on equipment
and piping which permit slow movements due to
thermal expansions but dampen the rapid shaking
encountered in earthquakes and resist sudden reactions
that would occur if a pipe ruptured. Very large support
components are required in water reactors to withstand
the reactions that could occur with pipe failure. Smaller
supports can be used in the molten-salt reactors because
the systems operate at lower pressure and have less
stored energy. The MSBR supports, however, must
operate at the high ambient temperatures in the cells.
This can be done either by designing dashpots which
use gases, molten salts, or pellet beds as the working
medium or by installing insulation and cooling systems
for dashpots using conventional fluids.

An engineering consultant®’ made a preliminary
review and evaluation of the ability of the MSBR to
withstand seismic disturbances. His findings were based
primarily on engineering judgment and extensive ex-
perience in seismic engineering. No major problem areas
were indicated for the seismic spectra used in current
designs of reactor plants. The shaking of piping and the
sloshing of fluids in the MSBR vessels do not appear to
be of major concern.

3.7 PRIMARY HEAT EXCHANGERS

C. E. Bettis
H. A. Nelms

M. Siman-Tov
W. C, T. Stoddart

3.7.1 Design Requirements

The overall conditions in the MSBR system impose
several specific design requirements on the primary heat
exchangers:

|. The volume of fuel salt in the heat exchanger must
be kept as low as practical to minimize the fuel
doubling time for the reactor.

2. The entrance and exit salt temperatures, maximum
(or desired) pressure drops, and the total heat
transfer capacity must conform with the overall
system operating conditions.

54

3. The type of heat exchanger, general location of
nozzles, height of the unit, and minimum tube
diameter must be compatible with various design,
layout, and fabrication considerations.

. The heat exchanger must be arranged for relatively
easy tube-bundie replacement by means of remotely
operated tooling.

5. All portions of the exchangers in contact with the
fuel or coolant salt must be fabricated of Hastelloy
N. As in any heat exchanger, the physical properties
of the material establish maximum allowable tem-
perature gradients across walls, allowable stresses,
and the degree of flexibility required to accommo-
date differential expansions.

. Flow velocities, baffle thickness, tube clearance, and
baffle spacing should be selected to minimize pos-
sibilities of vibration,

Within the framework of the above requirements and
guidelines, design procedures®? and a computer
program®? were developed to produce an efficient
design with low fuel-salt volume.

3.7.2 General Description

Four counterflow vertical shell-and-tube-type heat
exchangers are used to transfer heat from the fuel salt
to the sodium fluoroborate coolant salt. The units are
almost 6 ft in diameter and about 24 ft tall, not
including the coolant-sait U-bend piping at the top. A
cross-sectional drawing is shown in Fig. 3.33, and the
pertinent data are given in Table 3.15.

The fuel salt enters the top of each unit at about
1300°F and exits at the bottom at about 1050°F after
single-pass flow through the “%-in.-OD tubes. The
coolant salt enters the shell at the top, flows to the
bottom through a 20-in.-diam central downcomer, turns
and flows upward through modified disk and doughnut
baffling, and exits through a 28-in.-diam pipe concen-
tric with the inlet pipe at the top. The coolant salt is
heated from 850 to 1150°F in the process.

The 5803 Hastelloy N tubes are arranged in con-
centric rings in the bundle, with a constant radial and
circumferential pitch. The tubes are L-shaped and are
welded into a horizontal tube sheet at the bottom and
into a vertical tube sheet at the top. The toroidal-
shaped top head and tube sheet assembly has a
significant strength advantage, simplifies the arrange-
ment for the coolant-salt flow, and permits the seal
weld for the top closure to be located outside the heat
exchanger. About 4 ft of the upper portion of the
tubing is bent into a sine wave configuration to absorb
55

ORNL-DWG 69-6004

SECONDARY
SALT
PRIMARY

SALT

SECONDARY
SALT

Fig. 3.33. Sectional elevation of primary heat exchanger.
Table 3.15. Primary heat exchanger design data

Type One-pass shell and tubes with disk
and doughnut baffles
Rate of heat transfer per unit
MW 556.5
Btu/hr 19 x 10°
Tube-side conditions
Hot fluid Fuel salt?
Entrance temperature, °F 1300
Exit temperature, °IF 1050
Entrance pressure, psi 180
Pressure drop across exchanger, psi 130
Mass flow rate, 1b/hr 23.4 % 10°
Shell-side conditions
Cold Nuid Coolant salt?
Entrance temperature, °F 850
Exit tempersture, ° F 1150
Exit pressure, psi 34
Pressure drop across exchanger, psi 115.7
Mass flow rate, Ib/hr 17.8 % 10°
Tube material Hastelloy N9
Tube OD, in. 0.375
Tube thickness, in, 0.035
Tube length, ft 244
Tube-sheet-to-tube-sheet distance, ft 23.2
Expansion bend radius, in. 9.5
Shell material Hastelloy N
Shell thickness, in. 0.5
Shell ID, in. 67.6
Central tube diameter, OD, in. 20
Tube sheet material Hastelloy N
Tube sheet thickness, in, 4.75
Tube maximum primary (P) stresses, psi 683
Allowed primary stresses, psib 4232
Tube maximum primary and secondary (P + Q) stresses, psi 12,484
Allowed primary and secondary stresses, psi 12,696
Tube maximum peak (P + Q + F) stresses, psi 13,563
Allowed peak stresses, psi (see ref. 12) 25,000
Number of tubes 5803
Pitch of tubes, in. 0.75
Total heat transfer area, ft* 13,916
Basis for area calculation Qutside of tubes
Type of baffle Disk und doughnut
Number of baffles, total 21
Baffle spacing, in. 11.23
Disk OD, in. 54.2
Doughnut 1D, in. 45.3
Overall heat transfer coefficient, U/, Btu he ™' 12 (°F)~* 784.8
Volume of fuel salt in tubes, ft> 79

9Salt and Hastelloy N properties are those listed in Table S.1
bBused on average metal temperature in tube wall of 1244°F.
differential expansion between the tubes and the shell.
Baffles are not uged in this bent-tube portion, the tubes
being supported by wire lacing as needed to minimize
vibration. Without baffles the upper section of the tube
bundle experiences essentially paralle]l flow and rela-
tively lower heat transfer performance.

In the baffled section of the exchanger the tubes have
a helical indentation knurled into the surface to
enhance the film heat transfer coefficients and thus
reduce the fuel-salt inventory in the heat exchanger. No
enhancement was used in the bent-tube portion because
of present uncertainty in the reliability of the tubes if
they were both bent and indented.

The shells of the exchangers are also fabricated of
Hastelloy N. Disk-and-doughnut baffles, modified for
the central downcomer, are used in the shell to a height
of about 20 ft. The baffles produce cross flow and also
help support the tubes to minimize the vibration.
Although testing at conditions as near as possible to
design values is necessary to learn what tube vibrations
may occur, use of thick baffles (equal to, or slightly
greater than, the tube OD) and tube-to-baftle diametri-
cal clearances of the order of a few mils would tend
toward creating a *“fixed-tube’" situation at each baffle
and would be likely to prevent problems due to
vibration.

The upper and lower tube sheets are welded to a
cylinder with a 2%, in. wall thickness, which gives
rigidity to the tube bundle for transport, provides a
gamma shield for the shell, and forms a ‘-in.-wide
passage between it and the shell for downward flow of a
portion of the fuel salt to cool the wall. The top
extension of this inner cylinder, to which the upper
toroidal header is mounted, rests on a projection near
the top of the heat exchanger shell and supports the
tube bundle. The heat exchanger assembly is supported
from the cell roof structure and is mounted at a point
near the center of gravity by a gimbal-type joint that
permits rotation to accommodate unequal thermal
expansions in the inlet and outlet pipes.

Through close material control and inspection the
heat exchangers are expected to have a high degree of
reliability and to last the 30-year life of the plant. If
maintenance is required, a tube bundle can be removed
and replaced using remotely operated tooling, as dis-
cussed in Sect. 12. No specific arrangements are made
for replacement of the shell, although this could be
accomplished during a more extended shutdown of the
plant. A slip joint is provided at the inlet coolant-salt
connection to permit removal of the large U-bend in the
piping at the top. Once this is set aside, the bolting on
the top clamp is loosened and the clamp removed to

57

expose Lhe seal weld. After this is ground away, the
tube bundle can be withdrawn as an assembly.

3.7.3 Design Calculations

The design of the MSBR heal exchanger equipment
has been reported by Bettis et al.82:63 Heat transfer
experience with the primary and secondary salts is
limited, As experimental values for the physical prop-
erties of the salts become more reliable, confidence will
also increase in the heat transfer correlations and in the
overall design. The salt properties used in the MSBR
reference design heat exchange equipment are those
listed in Table S.1.

Since molten fluoride salts do not wet Hastelloy N, it
was suspected that usual heat transfer correlations,
often based on experiments with water or petroleum
products, might not be valid. MSRE experience®® and
recent experiments by Cox®® showed that basically the
fuel salt behaves very similarly to conventional fluids.
His correlations result in heat transfer coefficients
somewhat below those obtained from the Sieder and
Tate correlations for turbulent regions,®® Hansen's
equation for transition regions,” 7 and Sieder and Tate's
correlation for laminar regions.®® The tube-side heat
transfer calculations were made on the basis of correla-
tions recommended by McLain,*® which were based on
Cox’s data.%®

No experiments haye been performed to date for
correlating the heat transfer behavior of a sodium
fluoroborate coolant salt in the shell side of the heat
exchanger. Bergelin’s correlation®® for the baffle zone
and Donohue’s correlation”® for the unbaffled section
were chosen as the most representative available. Since
Bergelin’s correlation is strictly for cross flow situa-
tions, the equation was modified by introducing a
correction factor which depends on the degree of actual
cross flow existing as influenced by the ratio between
the baffle spacing and the shell annular thickness.

The tubes are spirally indented in the baffled zone to
improve the heat transfer performance. Experiments
performed by Lawson et al.”' showed that one can
expect an improvement by a factor of 2 for the
tube-side heat transfer coefficient. Lawson also recom-
mends a factor of 1,3 for the heat transfer coefficient
outside the tube, although no experiments have been
done to substantiate this. Since Lawson's experiment
was limited fo Reynolds numbers greater than 10,000,
there is some uncertainty in the degree of improvement
at numbers less than 10,000. It was assumed that no
improvement can be expected in a truly laminar flow
(Re < 1000). The range in between was extrapolated
using a method recommended by McLain.”?
The shell-side pressure drop was calculated by the
procedure suggested by Bergelin et al.®? The tube-side
pressure drop was calculated by the conventional
friction-factor method, The effect of the spiral in-
denting in the tubes on the pressure drop was assumed
to be in the same proportion as the effect on the heat
transfer performance.

A bypass correction factor due to baffle leakage of
0.5 was used for the pressure drop in the shell side of
the heat exchanger, and a factor of 0.8 was applied in
the heat transfer calculations. These leakage factors
were chosen on the bases of recommendations by
Bergelin et al.”?

A computer program was written which accepts the
design restrictions discussed above, takes into account
the differences in the physical properties of the salts as
they move through the exchanger, recognizes variations
in the flow and heat transfer regimes in the various
sections and applies the appropriate correlations and
correction factors, and, by performing a parametric
study, selects the heat exchanger design with the
minimum fuel-salt volume. Bettis et al. have described
the design procedures and the computer program and
its application.62:63 The reliability of the performance
estimates is assessed in Sect. 3.7.4.

A stress analysis subroutine was incorporated in the
main computer program. It performs a preliminary
stress analysis on the basis of the assumption that the
maximum tube stresses will occur in the curved-tube
region. The subroutine considers pressure stresses,
thermal expansion stresses, and stresses resulting from
thermal gradients across the tube wall. The primary and
secondary stresses are computed and compared with the
allowable stresses given in the ASME Sect. 11l Code.* ®
As additional information becomes available, the stress
analysis subroutine program will be expanded to in-
clude fatigue analysis, tube sheet joints, and the effects
on strength of the tube wall indenting.

3.7.4 Reliability of Design Calculations

[t is believed that the use of the MSBR primary heat
exchanger design program results in an efficient and
reliable design.

Among the input data which significantly affect the
heat exchanger design are the physical properties of the
fuel and coolant salts and their variation with tempera-
ture, the heat transfer correlations applied, the enhance-
ment factors assumed for the indented tubes, and the
leakage factors associated with fabrication clearances.
The most notable uncertainties in the salt physical
property values at the present time are the viscosity and
thermal conductivity of the fuel salt. The average

deviation for the fuel-salt heat transfer correlation is
reported®’ as being about 5.7%. The deviation or error
in the use of Bergelin’s correlation is not certain, but
shell-side heat transfer coefficients might normally have
a deviation of about 25%. Leakage factor deviations
might he about 0% for the pressure drop calculations
and about 10% for the shell-side heat transfer correla-
tion. The enhancement factor deviation might be about
15%.

Two extreme cases were examined: one where all the
pessimistic values were used and the other where the
optimistic values were taken. The result was a deviation
in overall heat transfer area (or fuel-salt volume) of
+38% for the pessimistic case and —28% for the
optimistic case.

3.8 SALT CIRCULATION PUMPS

3.8.1 Fuel-Salt Pumps
L. V. Wilson

The MSBR employs four primary-salt pumps and four
secondary-salt pumps, with one of each located in the
four system loops. In addition, there is a small ancillary
salt transfer pump with the dual purpose of filling the
primary-salt system and pumping the primary salt to
the chemical processing plant. For comparison purposes
the operating requirements for the pumps and tentative
values of some of the pertinent dimensions are shown in
Table 3.16. The secondary-salt pump is discussed in
Sect. 3.8.2 and the transfer pump in Sect. 3.8.3.

The fuel-salt circulation pump in the MSRE ac-
cumulated over 29,000 hr of successful operation, the

Table 3.16. Salt pumps for the 1000-MW(e) MSBR

Primary Secondary  Transfer?
Number required 4 4 1
Design temperature, °F 1300 1150 1300
Capacity, gpm, nominal 16,000 20,000 100 (3)
Head, ft 150 300 100 (25)
Speed, rpm 890 1190 1790 (890)
Specific speed, N 2630 2335 560 (140)
NPSH required,? ft 16 20
Brake horsepower, each ~2350 3230 20 (3)
Impeller diameter, in. 34 35%% 9%
Pump tank diameter, in, 72 72 24
Suction diameter, in. 21 21 3
Discharge diameter, in. 16 16 2

AWhere two values are listed, the first applies to filling the
primary-salt system and the second to circulating the primary
salt to the chemical processing plant.

bNPSH = net positive suction head.
only problem encountered being partial restriction of
the off-gas flow from the pump bowl.”* The pump had
a capacity of 1200 gpm and was driven by a 75-hp
motor. The dependability of this pump, a similar pump
in the coolant-salt system, and many others run for
thousands of hours in test stands has given confidence
that salt circulation pumps for the MSBR do not
present a major development problem.

The conceptual layout for the MSBR primary salt
pump is shown in Fig. 3.34. The lower portion of the
pump (pump tank, impeller, casing, etc.) is located in
the reactor cell, and the drive motor is located on the
crane bay floor, that is, above the concrete shielding.
The bearing housing is recessed into the concrete
shielding to reduce the shaft overhang. The pump shaft
is mounted on two pairs of preloaded oil-lubricated ball
bearings, and the impeller is overhung about 6% ft
below the lower bearing. The first shaft critical speed
will be greater than 1500 rpm to enable the pump to be
run at 1200 rpm when it is to be used for circulating
gas.

Since the reactor is the fixed component in the
system, the primary-salt pumps are subjected to thermal
expansion displacements of about 2 in. horizontally and
about 1 in. vertically at the pump tank when the system
is heated up from room temperature to operating
temperature. During operation at temperature the

coupling will accommodate the approximately “%-in.
horizontal- displacements due to thermal cycling. The
design effects of these displacements on the pump are
apparent in the shield configuration, method of pump
support, cell and/or pump containment, and the
coupling between the motor and the pump. The
shielding around the pump is of the disk-and-doughnut
type and will permit the unhindered displacement of
the pump and also provide adequate shielding of the
lubricant and coolant in the region of the lower bearing
and seal.

A shield plug is provided to protect the lubricant and
other radiation-sensitive elements in the region of the
bearing housing. Approximately a [-ft thickness of
Hastelloy N will limit the accumulated dosage at the
lower seal to 10 rads for the anticipated pump life.
The top of the shield plug will be cooled by an organic
liquid, possibly the same as the bearing lubricant.
Additional shielding will be provided to reduce the
nuclear radiation intensity at the crane bay floor to an
acceptable biological level,

The motor is mounted in a fixed position on the
crane bay floor, and the pump is suspended on
spring-mounted rods that are free to pivot at both ends,
The spring constant of the springs is sufficiently low

59

that the forces on the pump tank nozzles are not
excessive. The coupling between the motor and the
pump is a floating shaft gear type which is installed in
the maximum horizontal displacement position. During
system heatup the pump moves into a position where
the pump shaft is nominally aligned with the motor
shaft for normal pump operating conditions.

The pump has a large seal leakage containment
volume to accept the oil in event of a gross failure of
the lower seal. In addition, a Visco seal, adjacent to the
lower seal, will help to prevent oil from entering the salt
system when the shaft is rotating, When the pump is
stopped, a static shutdown seal can be actuated by gas
pressure to prevent the flow of oil down the shaft
annulus. The primary purpose of the static shutdown
seal is to prevent the leakage of gas-borne fission
products and thus permit the removal of the bearing
housing assembly without removing the shield plug,
shaft, and impeller from the pump tank.

The pump tank provides a volume to accommodate
the anticipated thermal expansion of the fuel salt at
off-design conditions. It is almost completely decoupled
hydraulically from the flowing salt in the impeller and
volute passages by (1) labyrinth seals installed in the
pump casing around the pump shaft and on the
periphery of the casing and (2) bridge tubes that
connect the volute to the inlet and outlet nozzles
attached to the pump tank. The bridge tubes also
eliminate structural redundancies between the pump
tank and the volute and its supporting structure.

The above-mentioned hydraulic decoupling serves (o
minimize the changes that may occur in the pump tank
liquid level if one pump stops when several pumps are
being operated in parallel. Assuming that the gas
volumes of the salt pumps being operated in parallel are
interconnected, that the salt volume in each pump tank
is connected directly to its pump suction, and that all
pumps are being supplied from a common plenum in
the reactors, if one stops, the leve] of salt in the tank of
the stopped pump would try to increase by an amount
equal to the velocity head at the pump suction plus the
head loss in the suction line from the common supply
to the pump tank. This change in level would be 10 ft
or more and would represent an undesirable increase in
the pump shaft length. Also, unless there is sufficient
reserve salt volume in the other pump tanks to supply
the increased salt requirement of the stopped pump, the
system fluid would in-gas when the salt level in the
tanks of the operating pumps is lowered to the level of
the pump suction. However, by connecting the liquid in
the hydraulically decoupled version of the pump tank
to a point in the reactor plenum where the velocity
131 2in,

60

ORNL~DWG 69-6802

MOTOR

BELLOWS ASSEMBLY
COUPLING
,_/
[\, ’.
2 CONCRETE
\ SHIELDING
» Iy S
) N
N
‘ ~ BEARING AND
A , SEAL ASSEMBLY
\\ ¥
1 o
\ N
L
. REACTOR CELL
CONTAINMENT

N AN

Yoy S S
i LAt
s

| O SHIELD PLUG

PUMP TANK
(6ft Qin.0D)

IMPELLER

Fig. 3.34. Primary-salt circulation pump.
changes very little when one pump is stopped and by
making the pressure drop in this connecting line very
low for the salt flow returning from the tank to the
plenum, the level change in the pump tanks probably
can be held to about 2 ft.

The pump tank, its internal structural elements, the
pump shaft, and the lower end of the shield plug are
cooled by a flow of primary salt (at about 1150°F)
which enters a plenum around the inner periphery of
the pump tank and flows upward in an annular liner
(see Fig. 3.34). At the junction of the pump tank and
the outer pump casing the (low splits, with part of it
passing downward between the inner and outer pump
casings and part of it passing across the lower end of the
shield plug and into the annulus between the shaft and
the shaft sleeve. These flows and the fountain flow
from the labyrinth seal then combine with the bulk salt
flow in the pump bowl. Filler blocks may be used in the
pump tank to reduce the parasitic volume of fuel salt.

At each pump the primary cell containment is
extended through the concrete shielding above the
reactor cell to contain the pump drive motor. The drive
motor heat sink is provided by cooling water circulated
through cooling coils attached to the inside of the
motor containment vessel. Internally, a blower attached
to the motor shaft will circulate helium through the
motor and over cooling fins attached to the inside of
the motor containment vessel. The motor is mounted
on a ring through which all electrical, instrument, gas,
coolant, and lubricant lines are connected to the pump.
To obtain a speed range from 10 to 110% of design
speed, each coolant-salt pump drive motor will
probably be supplied with variable-frequency power
obtained from individual solid-state inverters.

3.8.2 Coolant-Salt Circulation Pumps

The design conditions for the primary- and second-
ary-salt pumps are such that the same impeller and
casing design can be used for both. The secondary
pump will operate at higher speed, however, as shown
in Table 3.16. Except for the drive motor and the pump
tank, the two pump designs will be practically identical.

3.8.3 Salt Transfer Pump

The pump used to transfer fuel salt from the drain
tank, etc., could be an updated version of the PKA-2
pump that was designed for use in the ANP program
and has had several thousand hours of successful
operating experience. It will be operated at about 1790
rpm when filling the primary-salt system from the drain
tank and at 890 rpm when circulating salt to the
chemical processing system.

61

3.9 BUBBLE GENERATOR AND GAS
SEPARATOR

R. J. Kedl

3.9.1 Introduction

To enhance the breeding potential of the MSBR, it is
necessary to remove as many neutron-absorbing fission
products as possible from the fuel salt and dispose of
them external to the core. This is particularly true for
135Xe, with its very large absorption cross section.
Several mechanisms for removing xenon (and krypton)
have been studied. The one chosen for the MSBR
involves recirculation of helium bubbles. The theory
and calculations pertinent to this mechanism are
presented in Appendix A of this report. Summarizing
briefly, noble gases, because of their extreme in-
solubility in fuel salt, will migrate readily to any
gaseous interface available. Since they form a true
solution in salt (obey Henry's law), they will migrate in
accordance with the conventional laws of mass transfer.
If small helium bubbles are circulated with the fuel salt,
they will “soak up” xenon and krypton fission
products. The fission-product-rich bubbles may then be
separated from the salt and expelled to the off-gas
system. Xenon migration to the circulating bubbles is in
competition with Xenon migration to the porous
moderator graphite. The graphite is especially of con-
cern because it absorbs xenon and holds it in the core.
This tendency can be counteracted to a great extent by
sealing the surface pores of the graphite with chemically
deposited carbon as discussed in Sect. 3.2.3. In Ap-
pendix A it is concluded that, with moderate success of
the coated-graphite program, the 0.5% target value for
'35Xe poison fraction can be achieved when circulating
helium bubbles 0.020 in. in diameter, (The average void
fraction in the fuel loop would be about 0.2%.) This is
accomplished by bypassing 10% of the fuel salt from
the pump discharge through a bubble separator to
remove the xenon-containing bubbles, then through a
clean helium bubble generator for replenishment of
helium bubbles, and back into the pump suction, as
shown in Fig. 2.1. The average residence time of a
bubble in the fuel loop would be ten circuits.

3.9.2 Bubble Generator

In studying bubble generator concepts, essentially no
industrial experience was found, and very little informa-
tion was available in the literature concerning genera-
tion of bubbles in systems similar to the MSBR. An
exploratory program was therefore undertaken to ex-
amine both mechanical and fluid-powered devices. As a
result, a venturi device was selected for the MSBR, in
which gas is injected into the venturi throat and bubbles
are generated by the fluid turbulence in the diffuser
section.

The experimental bubble generator and its tust
facility are shown schematically in Fig. 335 1t con-
sisted of a teardrop shape inside a 1-in.-ID Plexiglas
tube through which water was flowing. Air was injected
into the annular throat through forty-eight % -in.-diam
holes around the circumference of the teardrop. The
model was tested under a variety of conditions of air
and water flow rates, teardrop shapes, and different
throat widths. Study of high-speed photographs of the
bubble action led to the following observations:

1. A continuous plume developed from each hole in
the teardrop and extended into the diffuser region.
The plume was then broken up into bubbles by the
fluid turbulence in this region.

The bubble size developed was apparently not a
strong function of the hole size used for gas
injection, at least over the range observed.

PHOTOGRAPH
POSIBTION

TWO 90° GLASS ELBOWS

f4-ft 1-in,-ID GLASS PIPE

62

3. The bubble size was independent of the gas flow rate
over the range tested.

The bubble size was a mild inverse function of the
water flow rate.

4,

5. The average bubble size was approximatelv 25% of
the throat width over the range tested. On this basis,
a throat width of about 0.08 in. would provide the
0.02-in. bubble size selected as desirable for the
MSBR.

A conceptual design for the MSBR bubble generator
is shown in Fig. 3.36. 1t consists of a system of linear
venturis formed by arranging air foils in parallel. The
throat width would be about 0.08 in., as discussed
above. The fluid velocity through the throat was
established as 40 fps, thus fixing the total throat length.
A conceptual cross section of a single air foil is also
shown in Fig. 3.36. The helium channel is shown as a
“controlled crack™; that is, one of the mating surfaces is
roughened in such a manner that when the two surfaces
bear against each other, a crack of controlled dimension
is formed through which the helium flow can be

ORNL-DWG 69-54989
WATER

PHOTOGRAFPH
POSLTION

f _

4-ft 1-in.-ID GLASS PIPE

F o5 o

PHOTOGRAPH DRAIN
POSICTION

BUBBLE GENERATOR TEST

AR

fin, ID

*

48-Y%4-in. HOLES

MATNN

—-a— WATER
] -=— AIR

~

BUBBLE GENERATOR DETAIL

Fig. 3.35. Bubble generator test flow diagram,
TRANSITION PIECE

5-in. PIPE

o s FLOW

0.080 in.

12 FOILS

VIEW AA

EVVVVV

)
WL

j ==—He SUPPLY NO. 4

ROUGHENED
SURFACE

" <—He SUPPLY NO, 2

63

ORNL-DWG 70-44914

>

-

12 AIR FOILS

f2in. x 12in. BUBBLE
GENERATOR ASSEMBLY

X

He SUPPLY NO, 4

He SUPPLY NO. 2

CONTROLLED CRACK FOR He FLOW
(CRACK WIDTH SHOWN EXAGGERATED)

Fig. 3.36, Preliminary concept of MSBR bubble generator,

regulated. A crack width of only about 0.001 in. will
probably be needed. The helium channel dimension is
kept small to reduce the likelihood that a pressure surge
in the salt system could push salt into the channel and
plug it. Since the fuel salt does not wet Hastelloy N, a
considerable pressure would be required to force the
salt into a 0.001-in.-wide opening. An alternative to the
controlled-crack method would be to install a narrow
graphite diffuser in the throat region of the venturi,

3.9.3 Bubble Separator

A pipeline bubble separator was chosen to remove the
gas-rich bubbles from the fuel salt, This type was
chosen primarily because of its low volume inventory
and high performance. In addition, there has been
considerable experience with this device at ORNL in
connection with the Homogeneous Reactor Test.”® A

device was tested which consists simply of a straight
section of pipe with swirl vanes at the inlet end and
recovery vanes at the outlet end, as shown in Fig. 3.37.
The swirl vanes rotate the fluid and develop an artificial
gravity field, This causes the bubbles to migrate to the
gas-filled core at the center of the pipe. The gas then
flows down the core and into the takeoff line which is
located in the hub of the recovery vanes. The recovery
vanes straighten out the fluid and recover some of its
energy.

3.9.4 Bubble Removal and Addition System

Figure 3.37 shows a schematic of the MSBR bubble
removal and generation equipment installed in a bypass
stream around the fuel pump. The pump head is in
excess of that needed to operate the system; therefore,
load orifices are required. (The pressures listed have
ORNL—DWG 7O-1{1915

FUEL—-SALT
PUMP
227 psi
) -
227 psi
—_— J'o psi
wz LOAD
hdica ORIFICE
ez | OAD
ad o 136 psi
$ | — ~
< HESEH
' 35 psi
flEpI::ElM 45 psi (NO.GAS CORE)
SUPPLY 85 psl (2 in, GAS CORE)

Fig. 3.37. Schematic flow diagram of bubble removal and gen-
eration bypass in MSBR fuel-salt stream.

64

been estimated from the model studies and are only
approximate.) The load orifice downstream of the
bubble generator is sized so that the generator will
induct helium from a 5-psig supply. The load orifice
between the separator and generator is sized so that the

ressure in the center of the separator, when no gas is
present, is sufficient to force salt into the takeoff line
and into the pump bowl. When gas is present at normal
operating conditions, the gas core will build up to about
2 in. in diameter and the pressure will rise. The load
orifice upstream of the bubble separator is provided to
take up the excess head. For maintenance purposes,
both the bubble generator and bubble separator should
be remotely replaceable, although one could anticipate
more maintenance for the bubble generator than for the
bubble separator.
4. Coolant-Salt Circulation System

4.1 GENERAL
W. K. Furlong H. A. McLain

An intermediate circulating coolant salt is used to
transport the heat generated in the primary system to
the steam-power system rather than to use direct
transfer because:

1. The loop provides an additional barrier for contain-
ing the fission products in the fuel salt in the event
of a heat exchanger tube failure and may provide a
barrier to tritium migration from the fuel salt to the
steam system.

. It links the high-melting-temperature fuel salt
(930°F) to the steam generator inlet feedwater
temperature (700°F) with a salt of relatively low
melting point (725°F), thus reducing the possibility
of freezing the fuel salt.

. The loop isolates the high-pressure steam from the
primary system, making it less likely that the
primary system could be subjected to high pressure
in the event of a steam generator tube failure.

. It guards against entry of water into the primary
system, which could cause oxidation and precipita-
tion of uranium and thorium.

5. It provides an additional degree of freedom in
control of the system through allowing the second-
ary-salt flow rate to be varied.

One of the design features desired for the MSBR is
that the coolant-salt system have natural circulation
capabilities under decay-heat-removal conditions, Multi-
ple loops are also desirable in order to improve the
reliability of the coolant flow.

The coolant-salt circulation system consists of four
independent loops, each containing a salt circulation
pump, steam generators, steam reheaters, coolant-salt
piping, and the shell side of one primary heat ex-
changer. The latter was described in Sect. 3.7, and the
coolant-salt circulation pumps were discussed in Sect.
38.2.

65

The heat transport fluid selected for the MSBR is
sodium fluoroborate salt. The various factors involved
in the selection were discussed in Sect. 3.2.2, and the
salt physical properties used in the design of the system
were listed in Table S.1. In brief, the salt is a eutectic of
NaF and NaBF,, with a melting point of about 725°F
and a low vapor pressure at operating conditions. It is
compatible with Hastelloy N, has satisfactory heat
transfer and flow properties, and has a low cost of less
than 50 cents/lb.

4.2 STEAM GENERATORS
T. W.Pickel W.K. Crowley W.C.T.Stoddart

4,2.1 General

The factors influencing the design of the steam
generators are much the same as those for the primary
heat exchangers, as discussed in Sect. 3.7, except that
the inventory of salt held in the units is not critical.

The total steam generation requirement, including
that needed for feedwater and reheat steam preheating,
is about 10 X 10° Ib/hr. 1t was arbitrarily decided to
divide this load between 16 steam generators, 4 to be
served by each of the 4 secondary-salt circulation loops.
The capacity required of each of the steam generators is
thus about 630,000 Ib/hr, or about 121 MW(t).

The steam generators are operated in parallel with
respect to both the coolant-salt and steam flows, and
they are identical in operation and design, The feed-
water supplied to the steamn generators will be pre-
heated to 700°F and is at a pressure of about 3750 psia
in the inlet region of the unit. (The feedwater heating
system is described in Sect. 5.) The 700°F feedwater
temperature should eliminate the danger of freezing of
the coolant salt in the inlet region, although this is yet
to be determined experimentally.

The water-steam fluid in the tubes is heated to exit
conditions of 1000°F and 3600 psia. The coolant salt is
cooled from 1150 to 850°F as it flows through the shell
side of the exchangers in a direction that is principally
countercurrent to the steam flow. The steam tempera-
ture delivered to the steam turbine will be controlled by
varying the coolant-salt flow rate through the steam
generators and by using a desuperheater, or attemper-
ator, in the outlet steam mains, as discussed in Sect. 5.

The radioactivity induced in the coolant salt in its
passage through the primary heat exchangers will
require biological shielding for the steam generators.
After reactor shutdown and a decay period of about ten
days, however, the generators can be approached for
direct maintenance, as discussed in Sect. 12.

The steam generator conditions analyzed in depth
were those for full-load operation, since this indicates

the size, approximate cost, and general feasibility of the
units. Some of the aspects of partial load and startup
conditions are discussed in Sect. 10. A computer
program was written to arrive at an efficient design for
the steam generators within the established design
parameters. This program accommodated changes in the
properties of the supercritical-pressure water with tem-
perature as it passed through the unit.

4.2.2 Description

The conceptual design of the steam generators is
shown in Fig. 4.1, and the principal data are listed|in

ORNL-DWG 70-1195!

STEAM
OUTLET

c;n;l,lll’“‘,l“
COOLANT SALT T
\ ‘ |

e
‘

FEEDWATER
INLET

Fig. 4.1. MSBR steam generator,
Table 4.1. Each unit is a counterflow U-shell, U-tube
heat exchanger mounted horizontally with one leg
above the other. Both shell and tubes are fabricated of
Hastelloy N. There are 393 tubes per unit, each ‘4 in, in
outside diameter and having a tube-sheet-to-tube-sheet
length of about 76 ft. The 18-in.-diam steam-side
hemispherical plenum chambers are designed for 3800
psia. The coolant salt circulates in counterflow through
segmental baffles in the shell to improve the heat

67

transfer coefficient for the salt film and to minimize
salt stratification. A baffle on the shell side of each tube
sheet provides a stagnant layer of salt to help reduce
stresses due to temperature gradients across the tube
sheets.

As in any once-through type of steam generator, the
feedwater must have the impurities limited to a few
parts per billion. Buildup of solids would only mean
decreased capacity, however, and would not present

Table 4.1. MSBR steam generator design data

Type

Number required

Rate of heat transfer
MW
Btu/hr

Shell-side conditions
Hot fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psia
Exit pressure, psia
Pressure drop across exchanger, psi
Mass flow rate, Ib/hr

Tube-side conditions
Cold fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psia
Exit pressure, psia
Pressure drop across exchanger, psi
Mass flow rate, Ib/hr
Mass velocity, Ib hr ' ft ™2

Tube material

Tube OD, in,

Tube thickness, in.

Tube length, tube sheet to tube sheet, ft
Shell material

Shell thickness, in.

Shell ID, in.

Tube sheet material

Tube sheet thickness, in.
Number of tubes

Pitch of tubes, in. (triangular)
Total heat transfer area, ft?
Basis for area calculation
Type of baffle

Number of baffles

Baffle spacing, ft

Horizontal U-tube, U-shell exchanger
with cross-flow baffles

16

121
4.13 % 108

Coolant salt
1150

850

233

172.0

61

3.82 % 10°

Supercritical fluid
700

1000

3752

3600

152

6.33 % 10°
2,55 % 10°

Hastelloy N
0.50

0.077

76.4
Hastelloy N
0.375

18.25
Hastelloy N
4.5

393

0.875

3929
Outside surface
Cross flow
18

4.02

problems of hot spots or burnout. The steam system
flowsheet, Fig. 5.1, follows established practice and
indicates full-flow demineralizers in the feedwater
system.

Because of the marked changes in the physical
properties of water as its temperature is raised above
the critical point at supercritical pressures, the heat
transfer and pressure drop calculations for the steam
generator were made on the basis of a detailed spatial
analysis with a computer program written for this
study.®® The program numerically integrates the heat
transfer and pressure drop relationships with respect to
tube length. The calculations establish the number of
tubes, tube length, shell diameter, and number of
baffles which are consistent with the specified thermal
capacily, steam pressure drops, and stress limits.

The heat transfer coefficient for the supercritical-fluid
film on the interior of the tubes was determined from
the relationship presented by Swenson et al.”® The
frictional pressure drop on the inside of the tubes was
calculated by using Fanning's equation, with the fric-
tion factor defined’” as

F=0.00140+ 0,125 (y/DG)O.az :

Values for the specific volume and enthalpy of
supercritical steam as functions of temperature and
pressure were taken from the work of Keenan and
Keyes.”® The thermal conductivity and viscosity as
functions of temperature and pressure were taken from
data reported by Nowak and Grosh.”?

The heat transfer coefficient for the salt film on the
outside surface of the tubes and the shell-side pressure
drop were based on the work of Bergelin et al.®®73 A
correction factor was applied to the heat transfer
relationships presented in these papers because of the
large ratio of baffle spacing to shell diameter (approx-
imately 2.7) required in this application. This correc-
tion factor is given by

C=0.77 (2y/B)*1?% ,

where

C = ratio of the corrected heat transfer coefficient to
the heat transfer coefficient calculated by Berge-
lin’s relationship,

y = distance from the center line of the shell to the
centroid of the segmental window area,

B = baffle spacing,

68

The physical properties for the salt used in these
calculations are as listed in Table S.1. The specific heat
and thermal conductivity of the salt were given as
constant values, but the density and viscosity were
functions of temperature. The functional relationships
were included in the computcr prograii

The ~3750-psia fluid pressure on the inside of the
tubes imposes relatively severe requirements on the
heads and tube sheets, This factor was considered in
selecting the number of steam generator units used in
the MSBR, since the relatively small diameter of 18 in.
selected for the shell allows the stresses to be kept
within more tolerable limits.

A preliminary stress analysis was made to establish
the feasibility of the steam generator design concept.
The analysis was based on the requirements given in
Sect. [II of the ASME Boiler and Pressure Vessel
Code’® A complete stress analysis, however, as re-
quired by this code, has not been made. For example,
fatigue analyses were not made in these preliminary
calculations. Additional information on the number and
types of operating cycles and on the effects of transient
conditions is required before a fatigue analysis can be
made. The stresses in the tubes due to steady-state
radial temperature gradients were treated as secondary
stresses rather than as peak stresses. This is the
approach taken in USAS B31.7 (1969) Nuclear Piping
Code®”? and is more conservative than the method of
ASME Sect. I11.5% The results of the stress calculations
are given in Table 4.2. As discussed in Sect. 3,5.3, the
allowable stress values for Hastelloy N were those
prescribed for the standard alloy in the ASME Boiler
and Pressure Vessel Code Cases 1315-3 (ref. 30) and
1331-4 (ref. 58).

4.2.4 Reliability of Design Calculations

The heat transfer and pressure drop calculations are
subject to review due to the empirical nature of the
correlations and the wuncertainties in the physical
properties used in the computations. Although both of
these aspects have been applied without safety factors,
it 1s believed that the preliminary design is a reasonable
one. In any event, the performance data will be
confirmed in test equipment before a final design is
initiated.

The design computer program was modified to permit
steady-state calculations for a specified heat exchanger
design under off-design operating conditions. This
program has been used to evaluate the performance of
the steam generator for operating conditions ranging
from 20 to 100% of design conditions. The calculations
indicate that the steady-state performance of the steam
Table 4.2. Summary of stress calculations

for an MSBR steam generator
Maximum stress intensity ? psi
Tube
Calculated P, =13,900;2, +Q =30,900
Allowable? P, =15500;P, +Q =46,500
Shell
Calculated AN 5800;Pm +Q=13,200
Allowable® P, =8800;P, + Q= 26,400
Maximum tube sheet stress, psi
Calculated <17,000
Allowabled 17,000

9The symbols are those of Sect. Il of the ASME Boiler and
Pressure Vessel Cc.\de,:"s with P, = primary membrane stress
intensity, Q = secondary stress intensity, and §,, = allowable
stress intensity.

PBased on a temperature of the inside tube surface of 1038°F,
which represents the worst stress condition.
“Based on the maximum coolant-salt temperature, or 1150°F.

dBased on the steam temperature of 1000°F and use of &
baffle on the salt side.

generator will be satisfactory over this range of operat-
ing conditions.

The problem of stability in the steam generator has
been considered briefly. As indicated by Goldman et
al®® and by Tong?®' instabilities in steam generators
can arise from two sources: (1) a true thermodynamic
instability where, for a given pressure drop across a
tube, the flow rate through the tube may be changed
from one steady-state value to another by a finite
disturbance, and (2) a system instability which is caused
by fluid ‘“‘resonant” conditions. Krasyakova and
Gluska®? have presented data concerned with the first
type of instability, and Quandt®? and Shotkin®* have
presented information on the second. A qualitative
evaluation of these data indicates that the mass flow
rate, pressure drop, and heat flux used in the horizontal

_~TIE ROD AND SPACER

TUBE SHEET—~. __

69

DOUGHNUT BAFFLE -

U-tube, U-shell design will result in stable operations.
Operation of a test module will provide further infor-
mation about the stability of this design concept.

4.3 STEAM REHEATERS

C.E.Bettis M. Siman-Tov  W.C. T. Stoddart

43.1 General

The design of the reheaters was influenced by most of
the factors that applied to design of the steam
generators, as discussed in Sect. 4.2.

The total steam reheating requirement is about 5.1 X
10° Ib/hr. It was decided to divide this load between
eight units, the capacity of each thus being about
641,000 Ib/hr, or 36.6 MW(t). The steam reheaters
operate in parallel both in respect to the coolant salt
and to steam flow. The coolant salt enters at 1150°F
and leaves at 850°F. The reheat steam is preheated to
about 650°F, as explained in Sect. 5, before it enters
the tube side of the reheaters at about 580 psia. The
exit steam is at 1000°F, the coolant-salt flow rate being
varied to maintain this temperature within a few
degrees.

The 650°F steam temperature entering the reheaters
is below the 725°F liquidus temperature of the coolant
salt, but a study of the heat transfer relationships leads
to the conclusion that there would be no significant
problem with freezing of the salt. This remains to be
verified experimentally, however.

As for the steam generator, a computer program was
written®? to arrive at an efficient design for the
reheater on the basis of the designated parameters.
These studies were based only on full-load conditions.

4.3.2 Description

As shown in Fig. 4.2, the steam reheater is a
22-in.-diam X 30-ft-long horizontal straight-tube single-

ORNL—DWG 66—T118A

|[SaLT ouTLET
: _~DISK BAFFLE :

- DRAIN LINE

STEAM
INLET

&

" INSULATION BAFFLE

Fig 4.2, Steam reheater.
70

Table 4.3. Steam reheater design data“

Type

Number required

Rate of heat transfer per unit
MW
Btu/hr

Shellsside conditions
Hot fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
Pressure drop across exchanger, psi
Mass flow rate, 1b/hr

Tube-side conditions
Cold fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
Pressure drop across exchanger, psi
Mass flow rate, 1b/hr

Tube material

Tube OD, in.

Tube thickness, in.

Tube length, tube sheet to tube sheet, ft
Shell material

Shell thickness, in.

Shell ID, in.

Tube sheet material

Number of tubes

Pitch of tubes, in. (triangular)
Total heat transfer area, ft*
Basis for area calculation
Type of baffle

Number of baffles

Baffle spacing, in.

Disk OD, in.

Doughnut ID, in,

Overall heat transfer coefficient, U, Btu hr ' 1t 2 (°F) ™"

Maximum stress intensity.b psi
Tube
Calculated
Allowable
Shell
Calculated
Allowable

Straight tube and shell with disk
and doughnut baffles

8

36,6
1.25 x 10%

Coolant salt
1150

850

228

168

59.5

1,16 X 10°

Steam

650

1000

580

550

29.9

6.41 % 10°

Hastelloy N
0.75

0.035

30.3

Hastelloy N

0.5

21.2

Hastelloy N

400

1.0

2376

Outside of tubes
Disk and doughnut
21 and 21

8.65

17.8

11.6

306

P, =4582;P,, + Q = 14,090
Ppy = Sy = 13,000; P,y + @ = 38, = 39,000

Py, =5016;P,, + Q= 14,550
Py, = Sy = 9500; Py + 0 = 35, = 28,500

@8alt and Hastelloy N properties are listed in Table S.1.

PThe symbols are those of Sect. [II of the ASME Boiler and Pressure Vessel Code,56 where P),; = primary
membrane stress intensity, Q = secondary stress intensity, and S,,, = allowable stress intensity.
pass shell-and-tube heat exchanger. There are 400 tubes,
%-in.-OD, in a triangular pitch array. Principal data are
listed in Table 4.3.

The tube surfaces are not indented to enhance heat
transfer, as in the primary heat exchanger. The coolant
salt is in counterflow through the disk-and-doughnut
baffles on the shell side. The units are installed in the
steam generating cells, as indicated in Figs. 13.7 and
13.8.

4.3.3 Design Calculations

A computer program was developed for designing the
reheater by modifying the primary heat exchanger
program as it existed in the early stages of development.
The properties of the steam were assumed to be
essentially constant along the length of the exchanger,
although it was recognized that some pain in the
reliability of the estimates could have been attained by
incorporating the steam properties as a function of
pressure and temperature.

The usual Dittus-Boelter equations were used for the
film heat transfer coefficient on the tube side. Other
procedures used in the heat transfer calculations were
described by Bettis et al.82+63

A preliminary stress analysis was made for the
reheaters. This analysis was based on the requirements
of Sect. III of the ASME Boiler and Pressure Vessel
Code;*® however, a complete stress analysis, as required
by this code, has not been made. The calculated stresses
are compared with allowable values in Table 4.3.

4.3.4 Reliability of Design Calculations

The confidence in the steam reheater design cal-
culations is greater than in the primary heat exchanger
because steam is a more familiar fluid than the fuel salt
and because no enhancement factors are involved.
Vibration problems are not likely to be encountered
because velocities are less than 6.5 fps and the tubes are
supported by baffles with relatively close spacing.

Two extreme cases were examined, one where all the
pessimistic values of the heat transfer coefficient were
used and the other where the optimistic end of the
range of possible values was assumed. The maximum
deviation in the overall heat transfer area, relative to the
reference design, was found to be +23% in the
pessimistic case and —13% in the optimistic case.

4.4 COOLANT-SALT SYSTEM PIPING
C. W. Collins

The secondary system piping connects the primary
heat exchangers in the reactor cell with the coolant

71

pumps and steam generators and reheaters in the steam
generating cells. The main piping is 22 in. in diameter,
with branches as small as 12 in. in diameter. The
operating temperatures are from 850 to 1150°F, but in
the design it was conservatively assumed that all the
secondary-coolant system piping would operate at
1150°F. This condition could actually exist only for a
short time, corresponding to temoval of the steam
generators and reheaters from service due to loss of
turbine load.

The piping flexibility analysis for the secondary
system piping was included in the calculations for the
primary system piping, since the two systems are
ronnected and interact with each other all the way to
the anchor points of the steam generators and reheaters.

The maximum expansion stress of 19,510 psi occurs
in one of the coolant return lines from a steam
generator. The operating temperature of this line is 850
rather than 1150°F, as assumed in the calculations. The
highest stress in the 1150°F pump suction line is
13,000 psi. Taking the allowable primary plus sec-
ondary stress intensity to be three times the allowable
design stress intensity (S,,), the allowable stress in-
tensity at 8S0°F is 54,000 psi and at 1150°F is 28,500
psi. The maximum stress due to pressure is approxi-
mately 3600 psi; therefore, the sums of the pressure
stress and the above maximum expansion stresses do
not exceed 385, , as specified by the codes.

Both the pump suction and coolant return lines of
each loop penetrate the reactor containment vessels and
cell walls. Bellows seals are used at these penetrations
on both the reactor cell and steam cell sides to maintain
the containment and permit about 1 in. of thermal
expansion of the piping along each of three axes.
Several flexibility analyses were made with the piping
fixed at the cell wall rather than use of bellows. This
resulted in excessive stresses in both the primary and
secondary loops, and since it did not appear that the
stresses could be reduced substantially without increas-
ing the piping lengths excessively, bellows seals at the
walls were adopted for the MSBR conceptual design.

4.5 SECONDARY-SYSTEM RUPTURE DISKS
J. R. McWherter

Each of the four secondary circulating loops will be
provided with a pressure-relief system to prevent
overpressurization in the event of a failure in the barrier
between the coolant salt and the steam system.

A rupture disk will be located at the secondary-salt
outlet of each steam generator. A preliminary design,
where the rupture disk assembly is set into a 12-in,
ORNL-DWG 70-41952

RUFTURE DISC
ASSEMBLY —ay_

AT ETIAS

A,
v NN

-t

=

. o WOnWAL SALT
: g LEVEL

COOLANT
SALT OUTLET

A

STEAM GENERATOR

Fig. 4.3, Secondary-salt system rupture disk.

vertical tee, is shown in Fig. 4.3. The elevation of the
disk is well above the normal level of the secondary salt
in the system. A gas pocket probably can be provided
to further reduce the possibility of salt contacting the
disk. The assembly is located in the steam generator
cell, which is maintained at about 1000°F, and its
downstream face is exposed to the ambient cell
atmosphere, making it improbable that the opening
would be obstructed with frozen salt, even in the
unlikely event that any of the coolant reached the disk
elevation.

The rupture disk will be fabricated from low-carbon
nickel, ASTM B162. This is a relatively pure metal with
adequate physical properties and corrosion resistance
for the service conditions. The disk will be designed to
rupture at 1000°F with a differential pressure equal to
the design pressure of the secondary-salt circulation
system (200 psi). A commercially available reverse-
buckling disk®* is proposed because of its accuracy
(rupture within £2% of rating) and greater cycle life.
The strength of the metal, and hence the failure
pressure of the disk, increases as the temperature
decreases. At 900°F the disk would fail at an estimated
pressure differential 10% higher than that at 1000°F.
Protective action, such as isolating the affected steam
generators with block valves, would be taken if the
temperature of the rupture disk falls below some
specified value, say 900°F.

72

If one of the ';-in.-diam tubes in a steam generator
were to fail, the pressure at the coolant-salt outlet of
the steam generator could rise from a normal value of
130 psi to about 200 psi in less than 1 sec. In analyzing
the pressure-containing requirements, it is pessimisti-
cally assumed that the ¢y tubes surrcunding a failed
tube will also fail in the estimated 5 sec required to
close the steam-system block valves at the inlet and
outlet of each steam generator. The total steam and
feedwater released to the cell via the rupture disk,
including that trapped between the block valves, is
estimated to be about 1150 lb, representing a heat
release of about 1.2 X 10° Btu. The steam generator
cell has been designed for 50 psig and will accommo-
date this energy release (see Sect. 13.11).

4.6 COOLANT-SALT DRAIN SYSTEM
W. K. Furlong

Four Hastelloy N tanks, each capable of holding 2100
ft* of salt with ample freeboard, are connected in series
to store the ~8400 ft* of coolant salt when it is drained
from the secondary circulation system. Four tanks were
chosen in order for them to be of a more reasonable
size, and the series arrangement was adopted to
facilitate heat removal if the coolant became contami-
nated with fuel salt. The tanks are located in a cell
directly beneath the steam generator cells, as shown in
Fig. 13.3. This cell is heated to about 800°F by electric
resistance heaters in order to maintain the salt above its
melting point.

Freeze valves are used to connect the first of the
coolant-salt storage tanks to the *‘cold™ leg of the
coolant-salt circulation loops. When the freeze valves
are thawed, the bulk of the salt in the coolant system
will drain by gravity, but about 730 ft? in each of the
primary heat exchanger shells will not and must be
removed by gas pressurization of the shell. Each heat
exchanger is provided with a l-in. dip line for this
purpose.

Since the coolant salt will undergo volume changes in
excess of the free volume available in the pump bowl,
each bowl has been provided with an overflow line
directed to the first coolant-salt drain tank. The salt will
be returned from (he tank to the circulation system by
a jet pump arrangement analogous to the arrangement
in the primary system. Gas pressurization can be used
to transfer salt from the other three tanks into the first
tank,

About 400 kW of heat-removal capability is provided
in the first storage tank in the event some fuel salt finds
its way into the coolant by accidental means. Most of
this heat would be transferred by radiation to cooler
surfaces in the cell. It has been estimated that in event
of tube failures in the primary heat exchangers, about
1370 ft* of coolant salt could be drained by gravity
from each coolant loop. In this situation, even with
tube failure, the fuel salt would continue to be
circulated to remove afterheat. (The heating due to
noble-metal deposition on the heat exchanger bundle is
the governing heat load.) During the circulation period
there could be considerable mixing between fuel salt
leaking from the primary system and the approximately
730 ft* of coolant salt remaining in a heat exchanger
shell after the coolant drain. If the shell is pressurized
after about 100 days and the salt mixture transferred to

73

the coolant-salt storage tank, the heat load to be
removed from the tank would be about 400 kW, as
mentioned above. It is recognized that during the
transfer process some of the salt mixture could be
forced back into the primary system through the
accidental opening. This salt would be drained with the
fuel salt into the primary system drain tank, it being
noted that in this type of system malfunction the fuel
salt was probably already contaminated with coolant
salt, since the coolant system normally operates at a
higher pressure than the fuel-salt system. The fuel-salt
drain tank has been provided with extra storage
capacity to accommodate some of the coolant salt, as
discussed in Sect. 6.
5. Steam-Power System

Roy C. Robertson

5.1 GENERAL

The thermal energy released in the MSBR is converted
to electric power in a steam cycle employing once-
through steam-generator-superheaters, a  turbine-
generator, and a regenerative feedwater heating system.
The relatively high operating temperatures in the MSBR
salt systems make it possible to generate steam at
conditions suitable for the most modern and efficient
steam-electric equipment now commonly in use,

Since the steam system components are more or less
conventional, there was no need to study the steam
cycle in any more detail than was necessary to make
cost and performance estimates for the MSBR plant.
There was thus a strong incentive to select a system for
which costs and thermodynamic data were readily
available, such as that used in the nearby Bull Run
steam station of the TVA. This 950-MW(e) plant
supplies steam at 3500 psia and 1000°F to the turbine
throttle, with reheat to 1000°F, and exhausts at 1'% in.
Hg abs. When applied to the MSBR reference design,
the cycle yields an overall net thermal efficiency for the
plant of 44.4%.

A particular requirement of the MSBR steam system
is that the feedwater supplied to the steam generator be
at a temperature high enough to avoid problems of
coolant-salt freezing. The lower limit for the water
temperature has not been established experimentally,
but for purposes of this study it was taken to be 700°F,
Also, for the same reason, it was assumed that the cold
reheat steam must be preheated to 650°F before it
enters the reheaters. These requirements, and the
convenience of using the Bull Run data in the con-
ceptual design study, led to selection of a system in
which the finai stage of feedwater heating is by direct
mixing with high-pressure steam. Although the method
is somewhat unconventional and requires use of
pressure-booster pumps in the feedwater supply, the
arrangement appears feasible and allows use of the Bull

74

Run information with only minor modifications. This
mixing method would probably be practical only if
supercritical pressures are used.

When detailed optimization studies become war-
ranted, several variations in the steam cycle can be
considered. It seems certain that tandem-compounded
single-shaft turbine-generators would be used in future
MSBR stations of large capacity rather than the
cross-compounded type at Bull Run.* Use of sub-
critical steam pressures, although less efficient, may
prove desirable from other standpoints. Use of reheat is
optional and would depend upon the steam conditions
selected, the turbine arrangement, etc. Startup and
partial-load conditions will have an important influence
on the steam cycle design.

The effects on plant performance and costs of use of
wet natural-draft cooling towers rather than the fresh
once-through condensing water supply assumed in the
reference design are discussed in Sect. 16.7 and ex-
plained in Table D, 17.

Although reasonably good efficiencies are attainable
with a variety of arrangements and the feasibility of the
molten-salt reactor concept is not strongly dependent
upon the details of the steam system associated with it,
this section recognizes that the steam-electric equip-
ment represents more than one-half the total station
investment, that it occupies a greater portion of the
plant space, and that even small differences in ef-
ficiency have economic value, all of which are of
interest to a plant owner. Some of the factors de-
veloped in the course of making this study which relate
to these aspects will therefore be briefly discussed.

*The cost of a tandem-compounded unit would not be as
great as for a cross-compounded machine, but its turbine
efficiency would be slightly less. Turbine performance data and
costs for a projected tandem unit were not obtained from a
manufacturer since the information available from the Bull Run
unit appeared adequate.
5.2 DESCRIPTION OF THE REFERENCE DESIGN
MSBR STEAM-POWER SYSTEM

Basic data for full-load conditions in the reference
design steam system are summarized in Table 5.1, and a
simplified flowsheet is shown in Fig. 5.1. Superheated
steam leaves the once-through-type steam generators at
about 3600 psia and 1000°F at a rate of about 10 X
10° Ib/hr. Coolant salt at 1150°F is supplied to the
steam generator at a controlled rate to hold the steam
outlet temperature to within a few degrees of 1000°F.
A steam attemperator, or desuperheater, supplied with
700°F feedwater assists in holding the steam tempera-
ture to within tolerances. (The steam generator was
described in Sect. 4.2.)

Table 5.1, Reference design MSBR steam-power system
and performance data with 700°F feedwater

General performance

Reactor heat input to steam system,® MW(t) 2225
Net clectrical output, MW(e) 1000
Gross electrical generation, MW(e) 1034.9
Station auxiliary load, MW(e) 25.7
Boiler-reedwuter pressure-booster pump load, MW(e) 9.2
Boller-feedwater pump steam-turbine power output, 29.3

MW (mechanical)
Flow ta turbine throttle, ib/hr 7.15 % 10°
Flow from superheater, [b/hr 10,1 X 10
Gross efficiency, (1034.9 + 29.3)/2225, % 47.8
Gross heat rate, Btu/kWhs 7136
Net efficieacy, station, 1000/2250, % 44.4
Net heat rate, Btu/kWhr 7687

Steam generators
Number of units 16

Total duty, MW(1) 1932
Total steam capacity, lb/hr 10,1 % 10°
Temperature of inlet feedwater, " 700
Enthalpy of inlet feedwater, Btu/ll: 769
Pressure of inlet feedwater, psia 3770
Temperature of outlet steam, ‘T 1000
Pressure of outlet steam, psia ~3600
Enthalpy of outlet steam, Btu/lb 1424
Temperature of inlel coolant salt, °F 1150

Temperature of outlet coolant salt, °F 850

Average specific heat of coolant salt, Btu Ib ™" (*F)™* 0.36
Total coolsant-salt flow
1b/hr 61,12 x 10°
ofs 145.5
gpm 65,290
Coolant-salt pressure drop, inlet to outlet, psi 61
Steam reheaters
Number of units 8
Total duty, MW(t) 294
Total steam capacity, Ib/hr 5.13 % 108
Temperature of inlet steam, “F 650
Pressure of inlet steum, psia ~570
Enthalpy of inlét steam, Btu/lb 1324
Temperature of outlet steam, °F 1000

Pressure of outlet steam, psia 557
Enthalpy of outlet steam, Btu/lb 1518
Temperature of inlet coolant salt, " F 1150
Temperature of outlet coolant salt, °F 850

Average specific heat of coolant salt, Bru b~ (°F) " 0.36
Total coolant salt llow
lb/hr 9.28 % 10°
cfs 22.1
gpm 9913
Coolant-salt pressure drop, inlet to outlet, psi 59.4

75

Table 5.1 (continued)

Reheal-steam preheaters

Number of units 8
Total duty, MW(t) 100 <
Total heated steam capacity, Ib/hr §513% 10

Temperature of heated steam, ° ¥

Inlet $52
Outlet 650
Piessure of heated steam, psia
Inlet 595
Outlet 590
Enthalpy of heated steam, Btu/lb
Inlet 1257
Outlet 1324
Total heating steam, Ib/hi 2.92 % 10°
Temperature of heating steam, ° ¥
Inlet 1000
Outlet 869
Pressure of heating steam, psia
Inlet 3600
Outlet 3544
Boiler-feedwater pumps
Number of units 2
Centrifugal pump
Number of stages 6
Feedwater flow rate, total, Ib/hr 715 % 10°
Required capacity, gpm 8060
Head, approximate, ft 9380
Speed, rpm 5000
Water inlet tempesature, “F 358
Water inlet enthalpy, Btu/ib 330
Water inlet specific volume, ft*/ib ~0.0181
Steam-turbine drive
Power required at rated flow, MW (each) 14.7
Power, nominal, hp (each) 20,000
Throttle steam conditions, psla/°F 1070/700
Throttle flow, Ib/hr (each) 414,000
Exhaust pressure, approximate, psia 77
Number of stages 8
Number of extraction points 3
Boiler-feedwuter pressure-booster pumps
Number of units 2
Centrifugal pump
Feedwater flow rate, total, lb/hr 10.1 x 10°
Required capacity, gpm (each) 18,950
Head, approximate, ft 1413
Water inlel temperature, “F 695
Water inlet pressure, psia ~3400
Water inlet specific volume, % /ib ~0.0302
Water outlet temperature, °F ~700
Electric-motor drive
Power required at rated flow, MW(e) (each) 4.6
Power, nominal, hp (each) 6150

9Does not include 25 MW(t) heat losses from reactor system,

Of the steam leaving the steam generator about 2.9 X
10° Ib/hr is diverted for the last stage of feedwater
heating; the remainder enters the 3600-rpm high-
pressure turbine throttle valve at 3500 psia and 1000°F.
After expansion of 1146 psia in the turbine, about 1.5
X 10° Ib/hr is extracted for driving the main boiler
feedwater pump turbines and for the final stage of
regenerative feedwater heating. The remainder of the
steam in the high-pressure turbine expands to about
600 psia and 552°F before exhausting into the two
34-in-diam cold reheat mains leading to the reheat
steam preheater. A portion of this exhaust steam is also
used for feedwater heating in the No. 2 heaters.
76

ORNL-DWE 7011963

51341908
r 540P - 1000° V518.5h {
10,067,550 ¢ '] (5 - S .- L.
3600P - 1000° - 14240 3515 P-1000*— 1424 h F‘—‘—-j
= IP TURBINE GENERATOR ‘A"
[»] s
LA l th o376.4h [ GROSS
= B | —| QuTRUT
. . p
@ 'l p ' ;; L/—/\ .: . 527.2 Mw
af 5 e i
61,420,000 9,280,400 # e l e L B — &
—it —_ ] 5,134,190 # | %
1150 1150° \ : : -4
REHEAT STEAM Teoop: : ) ~
BOILER- AT En PREHEATER e L < M £ ' e e B! e T
SUPERHEATE 2935 Mw(l) 1005 Mwll) r {3559n
19315 Mwll) e e B |
[Figpoaso= - 2 | | -
LOCLANT SALT | LDOLANT SALT —4  13235h ] ' =" 1
850" 850" | ~2,496,400 # I
I e 2,31,990# |
|- 10003 h | coweusu; 8"
1.5 in Hg ABS
s .'3 _ e M| (TR, ,_: 323,390 #
P 2ln P |TURBINE "8” |  GENERATOR 8" 103750 1
10 elT 1L GROSS
| fll:‘ ; 5%‘;1-7”1:47
° o w
g 2 7 1 T T =
' HOTWELL
o 2,915,370 % : { I PUMPS 8"
§ il 3500 P —865.7°~1307.8 h [ | | :
l | | .
724,520 # . ' | | ®2r220e vt I | 2,11300¢ |
| | _.J_L_l_.l,_l.-r_. ]
e r - 427430# | ' % i 203,479 # e
BOOSTER | L T 1 | weosss _ O7i0R_ """ o,
PUMPS 205"3363" a |
4.6 Mwe EACH Y i 5y —— | 33,230 | I
¢ | | ' | ( 12040 k 1 : |
3475 P-695 > <
78641 1-8 { -8 N I : | 4-8 | | 68 1} 7-8 g8 | ok
576,090« Lol
3,576, ' s ~ } | : | = A M5760w PP ~ e | rgga |
B¢ t i | | ol
]! ull i | LT
. |1 | o 1) __Iwa FLASH EVAP
S ] s 4 =
I w— Al —_
1-A 2-A | 4 . |193750¢ Y a-a - | 6-a 7-A 8-A
3,576,090 @ 5201 a80 7" : * |y:°;2_'_’° 362 59 2391 178647 2457880 [14367
> | N o il () "
3500P-550.9°-546.3 h AT57 = 423.1°- I 3aa3n » oz R | S TRV IR 8610660 | 1983 93.0°-610h
O | 4608N OC ] 4043 : ! A136100 :%?ag h 2024w’ |_DC 1913 6630
. T
““MIXING CHAMBER L — 752,180
362,260 # 576,000 # | i A 3,576,090 | _142620¢ | 220,650 ¢ 323380+
4857°-4710h 433 -4t 2 h [ ! 1B34"— 1513 b 1ABE - 1165 N 1436°-1115n
| gy e e .
i BF PUMP B TUR | |BF Pump ['
WATER | () ' i “
— — STEAM ! I OUTPUT f SUMMARY OF PERFORMANCE AT RATED LOAD
—--— COOLANT SALT i N 14.7 Mw i NET OQUTPUT 1000 Mwe
®  Ib/hr (TOTAL FOR ALL UNITS) CONDITIONS FOR “8" EQUIPMENT, pes| |Eapea R (¥ ‘ ! ELEC POWER FOR AUX 257 Mwe
P Ib/in® ABS (EXCERT GEN) SAME AS FOR 2ol 4] BF PRESSURE~BOOSTER PUMPS 92 Mwe
h o Bly/lb i GROSS GEN OUTPUT 1034.9 Mwe
s vF |34,040% BF PUMPS 29,3 Mw
——18s0 REACTOR HEAT INPUT TO GYCLE 2225 Mw!
%54 7300 GROSS EFF OF CYCLE 478 %
g NET EFF OF PLANT 44.4 %
NET HEAT RATE 7687 Blu/kwh
I432,180 # REACTOR HEAT OUTPUT 2250 Mw!

~ 3800 P~3664" 3433 n

Fig. 5.1. MSBR steam power cycle flowsheet.
The minimum temperature for the steam entering the
reheaters was assumed to be 650°F. The 552°F
high-pressure turbine exhaust steam is therefore pre-
heated, or tempered, in the shell side of a surface heat
exchanger using prime steam at 3600 psia and 1000°F
in the tubes. (The preheater is described in Sect. 5.10.)
The high-pressure steam leaves the tubes at about 3500
psia and 866°F and is used for preheating the feed-
water, as described below. The preheated *“cold™ reheat
steam, now at 650°F, then enters eight reheaters, which
are supplied with coolant salt at 1150°F at a controlled
rate to provide 1000°F steam at the exit. (The rcheate_s
were described in Sect. 4.3.) The reheated steam is
supplied to the double-flow 3600-rpm intermediate-
pressure turbine stop valve at about 540 psia and
1000°F.

There are no extraction points on the intermediate-
pressure turbine, Each cylinder exhausts directly into
the two double-flow 1800-rpm low-pressure turbines at
a rate of about 2,5 X 10° Ib/hr per turbine. Steam for
the No. 4 feedwater heaters is also taken from the
intermediate-pressure turbine exhaust.

Each of the four low-pressure turbine cylinders has
three extraction points for feedwater heating. About
2.1 X 10° Ib/hr is finally exhausted from each pair of
low-pressure turbines into four surface condensers
operating at about 1% in. Hg abs. Hot-well pumps
circulate the 92°F condensate through full-flow
demineralizers for the condensate polishing necessary to
obtain the high-purity water required in a once-t* vugh
steam generator. The feedwater flow then splits i, ‘o
two parallel paths for successive stages of feedwater
heating and deaeration. Booster pumps at the bottom
of the deaerators circulate the water through feedwater
heater 4 and to the two main boiler feed pumps. These
barrel-type six-stage centrifugal units have a capacity of
7500 gpm at 10,800 ft of head. Each is driven by an
eight-stage steam turbine with a brake horsepower
capacity of 21,500. The turbines have three extraction
points for feedwater heating and exhaust at 77 psia into
the deaerating feedwater heaters. The turbines normally
operate on 1146-psia steam extracted from the main
high-pressure turbine but can also accept 3500-psia
steam during startup or other times when extraction
steam is not available from the high-pressure turbine.

The feedwater, now at a pressure in excess of 3800
psia, flows through the three top regenerative heaters
and leaves at ~3500 psia and 551°F. Each of the 3.6 X
10° Ib/hr parallel-flow streams then enters a mixing
chamber, where the steam at 3500 psia and 866" F from
the tube side of the reheat steam preheater is mixed
directly with it. (The mixing chamber is discussed in

77

Sect. 5.8.) The resulting mixture, actually compressed
water at about 3475 psia and 695°F, then enters the
boiler feedwater pressure-booster pumps, (Two pumps
are shown on the flowsheet in Fig. 5.1, but as indicated
in Sect. 5.7, more detailed study of the pumps and the
system performance may indicate four or six parallel
units. They are also shown as motor-driven pumps, but
optimization studies would be likely to indicate an
advantage for steam-turbine drives for some of the
units.) The feedwater, now at about 3800 psia and
700°F, is returned to the steam generator at a rate
adjusted to the plant load by controlling the pumping
rate.

5.3 MSBR PLANT THERMAL EFFICIENCY

The steam system efficiency was estimated by using
performance values taken from the TVA Bull Run plant
cycle for the major items, particularly with regard to
pressure and temperature conditions.®® Bull Run mass
flow rates required adjustment, however, in that the
gross generating capacity of the MSBR is about 1035
MW(e) compared with 950 MW(e) for the TVA station.

The gross capacity requirement for the MSBR of
1035 MW(e) is based on an assumed plant auxiliary
electric load of 35 MW(e), of which 10 MW(e) would be
required tc drive the boiler feed booster pumps. The
reactor plant would need to supply about 2225 MW(t)
of energy to the steam-power cycle to deliver this
output. Heat losses from the reactor plant, exclusive of
long-range decay heal in off-gases, etc., have been
roughly estimated at 25 MW(t), making the total
required thermal capacity of the reactor about 2250
MW. The heat rejected by the drain tank heat disposal
system in normal operation is about 18 MW(t). This
decay heat has nolt been included in the thermal
capacity of the reactor. (It is redsonable to assume that
in optimized MSBR systems, a portion of this rejected
heat could be usefully applied.)

Based on a net output of the plant of 1000 MW(e)
and a reactor capacity of 2250 MW(t). the overall
thermal cfficiency of the station is 44.4%.7 The
efficiency based on the 2225 MW(t) of heat input to
the steam system is 44.9%, or a heat rate of 7601
Btu/kWhr.

5.4 SELECTION OF STEAM CONDITIONS FOR THE
MSBR STEAM-POWER CYCLE

If the thermal gradients in the steam generator tubing
walls and the coolant-salt freezing point do indeed
impose the requirements for a high feedwater tempera-
ture of, say, 700°F, the last stage of feedwater heating
78

in an MSBR plant obviously requires an arrangement
not found in a conventional steam power station, and
tenets of performance of the latter would not neces-
sarily apply.

The top temperatures for practical regenerative feed-
water heating could range fiom aboul 556 (0 575°F in
a supercritical-pressure cycle and from 475 to S00°F in
a subcritical-pressure cycle. Heating of the water to
700°F can be accomplished in a relatively simple
manner in the supercritical-pressure system by mixing
supercritical-pressure steam with supercritical-pressure
water, as shown in Figs. 5.1 and 5.2. (A mixing
chamber is discussed in Sect. 5.8.) The resulting
mixture is pumped back up to steam generator pressure
by special low-head high-pressure pumps, referred to as
pressure-booster pumps in Sect. 5.7. As an alternative, a
high-pressure heat exchanger could be used to heat the
supercritical-pressure feedwater to 700°F, with the exit
high-pressure heating steam reintroduced into the cycle,
possibly by heating it to 1000°F in a salt-heated
exchanger, thereby eliminating the pressure-booster
pumps and the 10-MW(e) auxiliary plant load they
imposed. Further study is needed of this alternate
arrangement to determine the extent of the economic
penalty.

Heating the feedwater to 700°F in a subcritical-
pressure cycle by surface heat exchange between steam
generator outlet steam and the water would require an
inordinate amount of steam generator throughput and
surface area. In the subcritical-pressure system, heating
is best accomplished in a Loeffler cycle, where steam
from the steam generator outlet is mixed with incoming
feedwater in a separate drum provided with distribution

nozzles to reduce the sparging effects. In a Loeffler
cycle modified for the MSBR conditions, as shown in
Fig. 5.3, the water would be converted to superheated
steam in the drum and then compressed and blown into
the “‘steam generator.” The latter, in reality, would act

probably be driven by a steam turbine, since the power
requirements could be in excess of S0 MW(e). In this
connection, it may be noted that the higher the initial
pressure of the steam to the compressor inlet, the less
the required compressive work on the steam.

A 3500-psia 1000°F/1000°F cycle with direct mixing
and booster pumps was compared by Robertson®® with
a 2400-psia 1000°F/1000°F Loeffler cycle with steam
compressors. The supercritical-pressure steam cycle
used as a reference was that shown in Fig. 5.1, The
mixing arrangement for the 2400-psia cycle is that
shown schematically in Fig. 5.3; the regenerative
2400-psia steam system flowsheet used for comparison
is taken from ref. 88. Both cycles include facilities for
preheating the cold reheat steam to about 650°F before
it enters the reheaters. As may be seen in Table 5.2, use
of subcritical-pressure steam results in a lower thermal
efficiency; also, the mass flow through the steam
generator would be about twice as great. Since the
specific volume of the steam at 2400 psia is about 1.5
times greater than at 3500 psia, the volumetric flow
rate is two to three times greater for the subcritical-
pressure system. This flow volume would have to be
accommodated by a greater number of tubes in the
steam generator. The expense of the greater number of
tube welds and larger shell diameter probably over-
shadows the cost of the thicker heads and tube sheets
required for the supercritical-pressure system.

ORNL-DWG 70-11954

3500 psia, 1000°F

|

+ TO IP TURBINE

1 HP TURBINE
{5 0CF = =
PREHEATER

850°F«— STEAM

GENERATOR

MIXER
PRESSURE )"
e FEEDWA

Fig. 5.2, Supercritical-pressure cycle with feedwater heated by mixing.
ORNL-DWG 70-41955

2400 psiy, 1000°F

e = TO IP TURBINE
— HP TURBINE
>
HSO"F—‘ S A R
< ,?.\TEHE TE
COOLANT SALT 2 )+
BS0°F - STEAM SEE.
GENEROTON — STEAM COMPRESSOR
\-—MIXING DRUM AND SUPERMEATER

4B0°F

FEEDWATER

Fig. 5.3. Modified Loeffler cycle for feedwater heating.

Table 5.2. Comparison of performance of
supercritical-pressure direct-mixing cycle
with subcritical Loeffler cycle to attain

Table 5.3. Properties of superheated steam at 900" F4

2400 psia 3500 psia

700°F feedwater
Supercritical Subcritical
cycle cycle
Nominal feed temperature to steam 700 ~700
generator, °F
Mixing pressure, nominal, psia 3500 2600
Booster pump or steam compressor 3800 2900
discharge pressure, psia
Booster pump or steam compressor 74 52
power requirement, MW(e)
Steam flow through steam generator, 9.5 x 10° 19 x 10°
[b/hr (total)
Overall thermal efficiency of heat- 44 .5 41.1

power cycle, %

Thermal conductivity, Btu hr ™ ft 0.052 0.062
°F)7!

Viscosity, Ib sec ft > X 107 63 67

Specific volume, ft3/lb 0.285 0.176

Specific heat, Btu Ib ™! (°F) ™! 0.74 0.91

Relative film resistance to heat transfer? 1.9 1.0

Use of supercritical-pressure steam also has some
advantages with regard to the heat transfer coefficient
on the steam side of the tubes in the steam generator.
Essentially all the heat transferred is in the superheated
regime, and the steam-side coefficient is largely con-
trolling. The physical properties of steam at 900°F for
3500 and 2400 psia are briefly compared in Table 5.3.
It can be seen that the film coefficient for heat transfer
in the 3500-psia system is about twice that in a
2400-psia system, and the surface area requirement
would therefore be significantly less.

In summary, the supercritical-pressure system pro-
vides a higher thermal efficiency, appears lo offer a

21967 ASME steam table values (ref. 78).
b Assuming the same tube diameters and velocities.

more direct means of attaining 700°F feedwater, and
could require 4 less expensive steam generator. The
higher efficiency not only affords a lower electric
power production cost but means less fuel processing,
less accumulation of fission products, and less heat
discharge to the environment.

5.5 USE OF REHEAT IN THE MSBR
STEAM CYCLE

Reheat would probably be profitable in the MSBR
steam cycle, particularly if plant layouts could be made
having shorter reheat steam lines than those used in the
reference design. More study is needed, however, before
it can be said conclusively that the improved efficiency
gained by use of reheat offsets the added complexity
and cost of the system. In considering reheat vs
nonreheat cycles, it should be noted that if reheat is not
used, external moisture separators are required o
prevent excessive moisture in the last stages of the
low-pressure turbines and that reheating does provide
somewhat better turbine performance than moisture
separation. These factors have not been evaluated
because this would involve obtaining rather precice
comparative information on equipment costs and tur-
bine performance, a refinement which to data has not
been warranted in the MSBR conceptual studies.

It is interesting that a study made for the LMFBR®?
comparing moisture separation with reheat for a 2400-
psig 900°F/900°F steam cycle concluded that the
economic gain for reheat (using sodium as the heat
source) was not sufficient to offset the added com-
plexity and reduction in plant reliability. These condi-
tions do not necessarily apply to the MSBR, however,
because the MSBR can attain 1000°F top temperatures
and does not require a relatively expensive reheater
design to accommodate exothermic reactions, as would
have been required for the LMFBR.

If future economic studies should indicate that reheat
for the MSBR cycle is indeed marginal, the system
could be simplified by elimination of the reheaters,
reheat steam preheaters, and the flow proportioners
that divide the coolant-salt flow between the steam
generators and reheaters.

5.6 EFFECT OF FEEDWATER TEMPERATURE ON
THE MSBR STEAM-POWER CYCLE

As previously mentioned, a feedwater temperature as
high as 700°F may be required for the steam genera-
tors, and an entering steam temperature of 650°F or
more may be needed for the reheaters. The special
equipment necessary to achieve these temperatures and,
more importantly, the loss of available energy in the
cycle are distinct disadvantages of the arrangement. In
the unlikely event that an even higher feedwater
temperature would be required, say 800°F, the dis-
advantages would become strikingly greater. It is
therefore of interest to briefly discuss the magnitude of
the cost penalties involved in order to compare them
with possible development costs for an improved
arrangement.

An MSBR steam cycle with 700°F feedwater and
650°F cold reheat steam was compared with one with
S80°F feedwater and 552°F reheat steam in ORNL-
3996 (ref. 4) and with a cycle with 800°F feedwater
and 650°F cold reheat steam.®® The results are
summarized in Table 5.4. The S80°F temperature was
selected primarily on the basis that this was about the
highest temperature that could be reasonably attained
by regenerative feedwater heating, In this case no

80

Table 5.4. Elfect of feedwater temperature on
performance of MSBR supercritical-pressure
steam-power cycle?

Nominal Booster Steam Net
feed pump generator plant
Wwmiporatia wutk fow rate emmeciency
(°TF) [MW(e)] (Ib/hr) (%)
x10°
5807 None required 7.4 449
70049 7.4 9.5 44.5
800°¢ 87 28 41.3

ABased on net plant output of 1000 MW(e) and reactor heat
of 2250 MW(t).

bassumes extra stage of regenerative feedwater heating and
no mixing or booster pumps required.

CFeedwater healed by mixing with steam from rcheat steam
preheater.

dThis case represents the performance now cited in MSBR
literature. Small variations exist due to different steam tables
used in the calculations.

special mixer or booster pump would be required, and
it was assumed that the reheat steam would not require
preheating.

Comparing the 580 and 700°F cases in Table 5.4, the
lower temperature affords a higher efficiency, which
can amount to about 10 MW(e) of additional output
capacity. An additional high-pressure feedwater heater
is required to obtain the 580°F water, but this cost is
more than offset by the expense of the mixing
chamber, pressure booster pumps, and reheat steam
preheaters needed in the 700°F cycle. As a result, the
S80°F cycle is estimated to have a total construction
cost, including indirect charges, of about half a million
dollars less than for the 700°F system.®*® Taking fixed
charges at 13.7% per annum, the saving amounts to
about $68,500 per year, This saving is small, however,
in comparison with the value of a better thermal
efficiency. Based on power worth 4 mills/kWhr, the
value of 10 MW(e) at 80% plant factor is about
$280,000 per year. The total yearly saving of the lower
temperature system is thus about $350,000. The
present worth (discounted at 6%) over & 30-year plant
life of this yearly sum is equivalent to roughly $5
million for an MSBR station. In a power economy with
many molten-salt reactors in operation, there would
thus be a strong incentive to develop a means for
lowering the required feedwater temperature, either
through use of a different heat transport fluid or
improved steam generator design, or both. (With regard
to use of a different secondary coolant, however, it
should be noted that the sodium fluoroborate pro-
posed in the reference design MSBR has an estimated
cost of less than 50 cents/lb. Since the coolant
inventory is about 900,000 Ib, if a different coolant
costs as much as about $3 per pound, the increased
inventory cost could nullify the cost advantages of the
lower temperature cycle.)

5.7 PRESSURE-BOOSTER PUMPS FOR
MIXING FEEDWATER-HEATING SYSTEM

After the feedwater is heated to about 700°F in the
mixing chamber used in the reference design (described
in Sect. 5.8), about 38,000 gpm of the mixture must be
raised to the steam generator inlet pressure of about
3800 psia. Canned-rotor pumps are currently in use
which operate under much the same pressure and
temperature conditions as those required. Preliminary
information obtained from pump vendors indicates that
development may be needed to produce multistage
variable-speed pumps, as may be required for the
MSBR, but no major extensions of the technology
appear to be involved.

5.8 MIXING CHAMBER FOR FEEDWATER
HEATING

The reference design provides 700°F feedwater by
direct mixing of supercritical-pressure steam at about
866°F with supercritical-pressure water at about 550°F.
The problems associated with the mixing of steam and
water at lower temperatures are well known; the rapid
formation and collapse of vapor bubbles causes noise,
vibration, and erosion similar to those found in pump
cavitation. At supercritical pressure, however, there is
no phase change or bubble formation, and the mixing
can be accomplished in a simple device.

At the TVA Bull Run steam plant, supercritical-
pressure steam and water are mixed in a 42-in,-diam
sphere, with the steam brought in at the top and the
water entering tangentially at the equator. The mixture
leaves at the bottom after passing through a screen with
%-in.-diam holes. The total pressure drop is said to be
less than 25 psi. One sphere handles a flow of over
4,000,000 1b/hr. Other mixing chamber configurations
may be possible, such as a simple pipe tee. Choice of
this method of feedwater heating for the MSBR cycle
dees not appear to impose major development
problems.

5.9 SUPERHEAT CONTROL BY ATTEMPERATION

Coarse control of the outlet steam temperature from
the steam generators will be by adjustment of the

81

coolant-salt pumping rate. Fine control, and more gross
control under certain loading conditions, will be
achieved by attemperating the steam with 700°F
feedwater injection. The attemperator design has not
been studied in any detail. The possible problem of
moisture in the throttle steam is alleviated to a large
extent because there would be approximately 150 ft of
high-temperature steam piping downstream of the
attemperator before the steam reached the turbine. A
major steam turbine manufacturer has stated that this
suggested method of superheat control by attempera-
tion is acceptable in principle.

5.10 REHEAT STEAM PREHEATERS
T. W. Pickel

5.10.1 General Description

The reference design requires that about 5.1 X 10°
Ib/hr of 551°F steam leaving the high-pressure turbine
exhaust be preheated to about 650°F before it enters
the reheaters. The proposed arrangement is to heat the
steam by heat exchange with steam at steam generator
exit conditions of 3600 psia and 1000°F. The capacity
required in each of eight preheater units is thus about
630,000 Ib/hr, or 12.3 MW(t).

There are eight identical preheater units operating in
parallel. The supercritical-pressure heating steam enters
the tube side at about 3600 psia and 1000°F and exits
at about 3535 psia and 869°F. The turbine exhaust
steam enters the shell side at about 595 psia and 551°F
and leaves at about 590 psia and 650°F.

A conceptual design for the preheater is shown in Fig.
5.4, and the principal data are given in Table 5.5. The
units are vertical single-pass U-shell, U-tube, with an
overall height of about 15 ft. The legs of the shell are
about 21 in. in diameter and are surmounted by
25-in.-ID spherical plenum chambers for the super-
critical-pressure heating steam. Each unit has about 600
tubes, % in. in outside diameter, located in a triangular
array. There are no flow baffles used on the shell side,
but bypass preventer rings are installed at intervals
around the tube bundle to prevent channeling of flow
in the clearance space between the bundle and the shell.
A baffle plate on the shell side of each tube sheet
provides a stagnant layer to help reduce stresses due to
the temperature gradient across the sheet.

5.10.2 Design Considerations

The preheaters may be constructed of Croloy since
they are not in contact with the fluoride salts, The units
will not be exposed to any radioactivity and will be
SUPER CRITICAL
FLUID OUTLET

STEAM INLET
—_—

ORNL DWG 65-23824A

u\\\\‘;i\\i:‘g:;c\)$ ~

SUPER CRITICAL
FLUID INLET

TUBE SHEET

BAFFLE

STEAM
OUTLET
fi

Fig. 5.4, Reheat steam preheater.

located in the feedwater heating bay, where direct
maintenance can be performed.

The high pressure of the heating steam prompted
selection of a U-shell rather than a divided cylindrical
shell, since it permits smaller diameters for the heads
and reduces the thicknesses required for the heads and
tube sheets, The same pressure considerations led to
selection of the spherical plenums for the high-pressure
steam.

The preheaters have been shown as vertical units, but
there is no compelling reason why they could not be
used horizontally. Gravity drainage is not considered
mandatory.

The heat transfer coefficient for the supercritical-fluid
film inside the tubes was calculated by using the
Dittus-Boelter equation,

hd.
—;‘fl = 0.023(Ng )08 (Np)04 .

The film heat transfer coefficient for the lower
pressure reheat steam flowing outside of and parallel to
the tubes was calculated by a correlation reported by

Short,”? given by
d,G\6 [ Coup\0-3?
Hp k

Pressure drops in the tubes and in the shell were
calculated by using the Darcy equation for the friction
loss; four velocity heads were associated with the inlet,
exit, and reversal losses; a correction factor was used for
changes in kinetic energy between the inlet and exit of
the exchanger.

An analysis was made of the stress intensities in the
tubes, tube sheets, shells, and high-pressure heads and
of the discontinuity-induced stresses at the junction of
the tubes and tube sheets. The results are shown in
Table 5.5. The calculated stresses are within the
allowable values.

o =0.16 —
Table 5.5. Design data for the reheat-steam preheater

Type

Number required

Rate of heat Lransfer, each
MW
Btu/hr

Shell-side conditions
Cold fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
Pressure drop across exchanger, psi
Mass flow rate, 1b/hr
Mass velocity, Ib h ™' £t ™2

Tube-side conditions
Hot fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
Pressure drop across exchanger, psi
Mass flow rate, Ib/br
Mass velocity, b he ™' £t ™2
Velocity, fps

Tube material

Tube OD, in.

Tube thickness, in.

Tube length, tube sheet to tube sheet, ft
Shell material

Shell thickness, in.

Shell ID, in,

Tube sheet material

Tube sheet thickness, in.
Number of tubes

Pitch of tubes, in. (triangular)
Total heat transfer area, fi?
Basis for area calculation
Type of baffle

Overall heat transfer coefficient, U, Btu he ! £t ™2

Maximum stress intensity,” psi
Tube
Calculated
Allowable

Shell
Calculated
Allowable

Maximum tube sheet stress, psi
Calculated
Allowable

One-tube-pass, one-shell-pass
U-tube, U-shell exchanger
with no balfles

8

12.33
4,21 x 107

3,56 X 10°

Supercritical water
1000

603

0.75

781
Tube OD
None
162

Py = 10,503; P, + @ = 7080
Py, = Sy, = 10,500 at 961°F;
Py + 0 = 38, = 31,500

Py =14,375; Py, + @ = 33,081
Py =8, = 15,000 at 650°F;
Py + Q =38, = 45,000

7800
7800 at 1000°F

“The symbols are those of Sect. [11 of the ASME Boiler and Pressure Vessel Code (ref. 56), where
Py, = primary membrane stress intensity, Q = secondary stress intensity, and S,,, = allowable stress
intensity.
6. Fuel-Salt Drain System

W. K. Furlong

6.1 GENERAL DESIGN CONSIDERATIONS

The preferred mode of MSBR operation is that the
fuel salt remain in the primary system after reactor
shutdown so that circulation can be continued through
the primary heat exchanger for afterheat removal.
There are some circumstances, however, either planned
or unplanned, which will require that the salt be
drained. Intentional drains are usually associated with
maintenance operations, such as reactor core graphite
replacement and servicing of pumps, heat exchangers,
etc. In these instances the salt circulation can be
continued as long as necessary prior to the drain to
allow the activity to decay to the necessary level for the
maintenance task. There is a low probability of un-
scheduled drains, but they must be accommodated in
the design. Examples of unplanned situations are:

1. massive failure of a primary system pipe or vessel,

2. a slow loss of salt from the primary system so that
pumps would eventually be unable to maintain

circulation,
3. loss of heat-removal capacity in the steam system,
loss of coolant or circulation in the secondary loops,

5. loss of power or mechanical failure of primary
pumps,

inadvertent thawing of- the frecze valve which holds
the fuel salt in the primary loop.

The principal function of the fuel-salt drain system is
to provide a place where the salt can be safely
contained and cooled under any of the accidental or
intentional situations. The drain system must, there-
fore, have a highly reliable cooling system capable of
removing the afterheat even with a sudden drain after
long-term operation at full reactor power. In designing
the cooling system the overall objectives were:

84

1. It must be able to keep the maximum drain tank
temperature well within the safe operating range
even under the worst condition of transient heat
loads.

2. The system must be reliable, with a minimum of
reliance on the electric power supply or operator-
initiated actions.

3. If only a single barrier is provided between the tank
coolant and the fuel salt, leakage of the coolant into
the salt should not require chemical processing to
prevent adverse nuclear or chemical effects.

4. The cooling system should impose a minimal risk for
freezing of either the fuel salt or the cooling system
coolant.

Several methods of cooling the drained fuel salt were
considered. One was to store the salt in a long pipe with
radiant heat transfer to cooled plates. Another possible
method was the use of heat pipes to cool fuel-salt-filled
tanks. Since a storage tank with a convective cooling
system was used with good results in the MSRE, it was
decided that the above objectives would be best met by
storage of the salt in a tank having a coolant circulated
by natural convection to a water-cooled heat exchanger.
A variety of heat-transport fluids were studied. The salt
originally selected as having the most promise was
TLiF-BeF,, and a drain system using this salt was
studied in some detail, as described in Sect. 6.3. Late in
the study, however, the apparent advantages of an
NaK-cooled system led to consideration of an alternate
drain tank cooling system using NaK as the coolant, as
discussed in Sect. 6.4. Unfortunately, the NaK system
study could not be developed in time to be reported as
comprehensively as the salt-cooled system.

Without impairing the above-mentioned principal
function of the drain system, the drain tank can be
conveniently used for other purposes, such as a holdup
volume for off-gases to allow about a 2-hr decay time
before the gases are processed. The drain tank cooling
system can continuously remove the decay heat load of
these gases and at the same time provide assurance that
the cooling system is operable and could accommodate
a major drain. With this arrangement, internal surfaces
in the drain tank, particularly cooled ones, may act as
sites for deposition of noble metals in the off-gas and
will possibly eliminate the need for a particle trap in the
off-gas system. The decay heat load in the drain tank,
estimated to total about 18 MW(t), is discussed in more
detail in Sect. 6.3.2.

The drain tank also serves usefully as a surge volume
to which salt can be continuously overflowed from the
primary pump bowl. The supply and return connections
to the chemical processing facility will be made at the
drain tank. The same jet pump arrangement used to fill
the primary system from the drain tank can be used to
transfer salt to the chemical facility, eliminating the
need for pressurizing the tank for salt transfer. With this
arrangement, salt can be taken from the tank for
processing independently of reactor operation.

It was also decided that the reference MSBR design
would provide a backup container if the drain tank
should develop a leak. In addition, a second safe storage
tank was provided for the salt to permit the primary
drain tank to be drained for repairs,

6.2 FUEL-SALT DRAIN LINES

Although draining the fuel salt from the reactor is a
positive shutdown mechanism, it is not necessary to rely
on this as an emergency procedure, and rapid drainage
is not a primary design criterion. The drain tank is
connected to the bottom of the reactor vessel by a 6-in.
drain line equipped with a freeze-plug type of “‘valve”
which can be thawed to allow gravity drainage of the
entire primary circulating system in about 7 min. A
small circulation of fuel salt is normally maintained in
the drain line between the reactor and the freeze valve
to prevent overheating due to stagnant salt, as indicated
in the drain system flowsheet, Fig. 2.3.

During normal operation of the reactor about 150
gpm of fuel salt overflows from each circulating pump
bowl. The gases stripped from the fuel salt at the gas
separator, laden with highly radioactive fission product
gases and particulates, are combined with the overflow
salt from the pump bowls in a small tank (4 in Fig. 2.3)
before flowing to the drain tank. The 2-in. overflow line
has a 3-in.-diam counterflow cooling jacket supplied
with 1050°F fuel salt from the reactor inlet. This salt,
in flowing upward through the jacket, also cools the
small mixing tank and the lower portion of the pump
bowl] before mixing with the bulk salt flow in the bowl.

85

The overflow gas-salt mixture, which reaches the drain
tank at an estimated temperature of about 1200°F,
enters the top of the drain tank and is first directed
beneath the top head and then downward through a
!, -in.-wide annulus between the tank wall and an
internal liner (used as @ gamma shield) to cool the drain
tank and the internal liner.

6.3 PRIMARY DRAIN TANK WITH SALT-COOLED
HEAT-DISPOSAL SYSTEM

6.3.1 Description

The drain tank is a vertical cylinder about 14 ft in
diameter and 22 ft high with torispherical heads and
internal U-tubes, All portions in contact with salt are
constructed of Hastelloy N. Plan and elevation views are
shown in Figs. 6.1 and 6.2, and the principal data are
listed in Table 6.1. The layout of the drain tank and its
cooling system is shown in Fig. 6.3.

The storage volume of the tank is about 2500 ft®
The tank dimensions were based on the following
volume requirements (1%):

Total fuel-salt volume 1720
Volume of coolant salt that could 730
reach drain tank in event of tubeg
failure in one primary heat exchanger
Volume occupied by U-tubes and 250

other components in drain tank

After considering various means of cooling the tank
walls and heads, it was decided to use the internal liner
with a continuous fuel-salt flow to remove the heat. A
flow of 150 gpm of fuel salt from each of the primary
circulation pumps, after being cooled to about 1200°F
by a counterflow of “cold”-leg salt, as mentioned
above, will enter the drain tank and flow down the
annulus between the liner and the wall. The annulus is
orificed at the bottom to ensure that it remains full of
salt. The maximum steady-state wall temperature is
estimated to be 1260°F, occurring at the bottom. The
liner is separated from the walls by standoifs to provide
a 0.5-in. radial cooling passage and to make it struc-
turally independent of the tank. The liner also provides
support for internal baffles, which are provided to
impart a circuitous path for the off-gas and also to
stiffen the U-tubes. Since there are no structural
connections between the tank and the inner liner, the
status of the tank as an ASME Code Sect. 111, class AS°®
vessel is not impaired by this approach.

The 0.75-in.-diam U-tubes through which the cooling
salt circulates to remove the heat generated in the
86

ORNL DWG 63-10495A

f4ft 2in. OD -

b2y o
bz@ @L@@@?@m (@==t’

C

i) fifi%é‘:‘.’pfi!x@ e “

‘ E@ PRIMARY SALT / =
DRAIN LINE (&in.) —

.COOLANT SALT

e
&
= =
: SA S
(40 CIRCUITS, 3-in. PIPE) é N

Fig. 6.1. Plan view of fuel-salt drain tank with salt cooling,
87

ORNL-DWG 69-10493A

COOLANT SALT rz!MARY SALT TYPICAL PENETRATION FOR

N | Mgy f

in. PIPE) GAS AND JET PUMP PIPING

el ft Qin

I 10009

15t &in

SALT
LEVEL

1311 Qin

L s 3/4~0D x 0042-in-WALL "U" TUBES (1500 TOTAL)

e

{C
1

g

L— 5t Qin 10—

- 'Iz in.

13f1 9in 1D

Fig. 6.2. Elevation of fuel-salt drain tank with salt cooling.
Table 6.1. Principal design parameters and data for primary
drain system using salt cooling system?®

Drain tank
Outside diameter, It 14
Overall height, ft 22
Wali thickness, in, 1
Bottom head thickness, in. 11/2
Liner thickness, in. 1
Material Hastelloy N
Storage capacity, f t* ~2500
Design conditions, psig/°F 40/1300
Number of internal U-tubes ~1500
U-tube OD X wall thickness, in, 0.75 % 0.042
Off-gas flow rate, cfm at 10 psig and 1000°F 18
Flow rate of overflow salt, gpm 600
Entering temperature of overflow salt, °F 1200
Fraction of total noble metal yield found in off-gas 0.5
Off-gas holdup time, hr ~2.3
Equilibrium heat generation in off-gas and noble metals, MW(t) 18
Heat absorbed in tank liner and walls, MW(t) 4
Maximum heat release from salt after sudden drain, MW(t) 53
Maximum steady-state heat load, MW(t) 18
Maximum steady-state wall temperature, °F ~1260
Estimated time for primary system to drain, min 7
Heat disposal system

Drain tank coolant fluid 7LIF-BCF2
Coolant composition, mole % 67-33
Number of autonomous cooling circuits 40
Total coolant volume, fit3 ~400
For normal steady-state operation at 18 MW(t) heat release in drain tank:%

Temperature of coolant entering drain tank, °F 900

Temperature of coolant leaving drain tank, °F 1050

Coolant circulation rate, gpm at av temperature 714
For conditions after sudden drain of salt, heat release of 53 MW(1):

Temperature of coolant entering drain tank, °F 900

Temperature of coolant leaving drain tank, °F 1163

Coolant circulation rate, gpm at av temperature 1200
Number of salt-to-water heat exchangers 40
Number of tubes in each exchanger 333
Tube size, length (ft) X OD (in.) 10 % 0.625
Arca in each exchanger, ft? 544
Water pressure, psia 100
Distance of heat exchangers above drain tank midplane, ft 60
Stack size, height X diam, ft 400 X 60

2Due to decay of gases and noble metals only.

stored salt are divided into 40 separate circuits. The
choice of the number of circuits was somewhat arbi-
trary, the primary objective being to have a large
number so that in event of failure, any one of them
would represent only a small loss in capacity. There
were also space limitations in arranging the header
circuits at the top of the drain tank. It may be noted
that all welds for the coolant system tubes and headers
are well above the normal fuel-salt level in the drain
tank,

Salt flows into the drain tank by gravity. It is
transferred from the tank by salt-actuated jet pumps
located in a salt reservoir provided by a depression in
the bottom of the tank, Four jet pumps, one in parallel
across each primary salt pump, return the overflow salt
to the “hot™ leg of each primary loop. Some internal
cooling of the drain tank wall can be maintained even if
three of the four primary salt pumps should fail. An
ancillary salt circulation pump is used in conjunction
with a fifth jet pump in the bottom of the drain tank to
ORNL-DWG 69-10487A

W ."
AIR COOLED RADIATORS IN STACK
REACTOR
CONDENSATE PRIMARY SALT Pumpj s o
STEAM . |
PRIMARY |
HEAT EXCHANGER i
T.-\ w ¥ Y W i3 30 B¢ 30 3C 0 - |
Fi
' '{ LB
: l |" Il" I!I'."l‘ L !. I -3
RGN |
" o= Ay == f
(" - i) :.B\ l' 2 = T
i | ol N |
s : | ] J
1 A L (= S = "
50N N TC SN 3NN S SE SC 00 B ) OFF-GAS NEW_CORE
: 2 CELL REPLACEMENT CELL
\ -
\l m
B =
DRAIN TANK COOLANT SALT
(LBz, 40 CIRCUITS) —._
DRAIN LINE —
e (6-in)
o |
FREEZE VALVE CELL \
|
WASTE STORAGE CELL
DRAIN TANK AND
GAS HOLDUP

Fig. 6.3. Elevation of drain tank cell with salt cooling system.
transfer salt to the chemical processing facility. By
thawing a freeze valve, indicated as A in Fig. 2.3, this
jet pump can also be used to transfer salt from the drain
tank to fill the primary system.

One feature not shown in Figs. 6.1 and 6.2 is an
external shell arannd the side walls and bottom of the
drain tank which acts as a backup container in the
unlikely event of a failure of the tank below the salt
level. This shell, sometimes referred to as a “crucible”
in the MSR literature, is made of stainless steel and is
open at the top. The annular space between the shell
and the tank is filled with tightly packed copper rope,
the purpose of which is twofold: to minimize the salt
volume which can occupy the annulus and to provide a
good conductor for heat to the tank wall.

6.3.2 Heat Sources in Drain Tank

In normal operation the drain tank receives ~11 scfm
of off-gas containing radioactive gases and metals.”
Besides tritium, the gases are primarily Kr and Xe, and
the noble metals are Nb, Rh, Mo, Ru, Tc, and Te. Heat
is also produced by decay of the daughters of Kr and
Xe, notably Ba, La, Cs, Rb, Sr, Y, and Zr. Assuming
that all of the noble metals present in the system
deposit on the U-tube walls and other internal surfaces
of the drain tank, the equilibrium value for the heat
source would be about 9 MW(t). Decay of the radio-
active gases and daughters contributes a maximum of
another 9 MW(t), making a maximum total of about 18
MW(t) generated in the drain tank for a reactor which
has run several weeks at full power.

The heat sources in the tank were assumed to be
uniformly distributed over the drain tank volume, and
the methods of Rockwell’' were applied to estimate
the source strength in the liner and tank walls. It was
noted that approximately 40% of the off-gas energy is
released as betas and hence is deposited locally.
Similarly, about 40% of the energy due to the noble
metals is from beta emission.® The gamma source per
unit of homogenized tank volume then becomes 3857
W/ft®. This converts to 2150 W/ft?* impinging on the
liner. Close agreement is obtained between cylindrical
and spherical models.

Estimates of the internal energy absorption by the
U-tubes and other internals were based on a linear
energy absorption coefficient of 0.82 in.™", which was
determined for attenuation of reactor spectrum gamma
radiation in the reactor vessel wall using a gamma
transport calculation (ANISN). Assuming the same
absorption coefficient, 56%, or a heat flux of 949
W/ft?, is absorbed in the 1-in.-thick liner, leaving 782

90

W/ft* to be absorbed in the 1-in.-thick tank wall from
this source. The rest of the energy will be absorbed in
the backup vessel in which the drain tank sits. Since the
tank walls and head have about 1000 ft* of surface

area, a heat load of about 2 MW(t) must be accommo-
ntad

M

The drain tank will be used as a salt repository during
shutdown for core graphite replacement or other
maintenance. The design basis for such a drain has been
taken as 10°% sec, or 11.6 days, after reactor shutdown.
During this interval the salt is circulated with the
primary-salt pumps to remove afterheat, including that
associated with sources adsorbed on and diffused into
the graphite in the reactor. The heat load due to the salt
and noble metals in the drain tank at the end of this
period, and immediately after the drain, is about 4
MW(1).

The most severe heat loads imposed on the drain tank
would be an inadvertent thaw of the freeze valve or an
emergency shutdown and drain. (Possible causes for
such shutdowns were discussed in Sect. 6.1 above and
by Furlong.®?) The maximum heat load that could
occur in such circumstances is estimated to be about 50
MW(t), if about 7 min is allowed for the drainage to
take place. The maximum possible heat release in the
tank, with no credit taken for heat sources retained by
the graphite, is shown in Fig. 6.4. In general, the
afterheat rejection requirements decrease by a factor of
10 during the first day.

6.3.3 Heat Transfer in Drain Tank Walls

During normal operation the tank walls and liner are
cooled by overflow salt from the reactor, as mentioned
above. A value of 150 gpm per pump was chosen
somewhat arbitrarily for the overflow rate. It was
desirable to have a value large enough to give adequate
cooling and also be large compared with the discharge
rate from chemical processing to assure good mixing of
processed salt as it returns to the reactor. On the other
hand, an upper constraint was the jet pump size, The
mixture of overflow salt and off-gas flows in a 2-in. pipe
located concentrically inside a 3-in. pipe. The annulus
between the pipes is connected to the drain line
upstream of the freeze valve, Cold (1050°F) salt from
this source cools the overflow lines, the mixing chamber
(A in Fig. 2.3), and the walls of the pump bowl. About
150 gpm will cool the overflow mixture to 1213°F
(average of four lines) upon entering the drain tank and
will have a temperature range (depending upon the line
length) of 1124 to 1167°F upon entering the mixing
chamber and slightly higher temperatures upon entering
the pump bowl. A higher value may be desirable,
2 ORNL-DWG 70-11956
— = NS hr =1

megawatts

ol 108 1P

TIME (sec)

10°

Fig. 6.4. Total afterheat production.

depending upon the relative importance of the colder
drain tank wall coolant and the ability to keep it cool
with less than four pumps running vs the acceptable salt
temperature impinging on the pump tank walls.

'Vith the heat sources described in Sect. 6.3.2 and
with a total flow of 600 gpm cooled to 1213°F, it is
estimated that the maximum drain tank wall tempera-
ture will be ~1260°F and that the maximum liner
temperature will be about 1300°F during normal
steady-state operation. These temperatures appear to be
acceptable. However, if necessary, they can be lowered
by appropriate adjustments in the flow rates of over-
flow salt and/or counterflow salt.

6.3.4 Heat-Removal System

A qualitative comparison of the coolants considered
for the drain tank heat-removal system is given in Table
6.2. The fused salts, NaK (see Sect. 6.4), and the
steam-water systems were considered to be most
worthy of further consideration. The most likely salt of
the candidates were (1) sodium fluoroborate
(NaBF4-NaF), the same salt used in the MSBR second-
ary system; (2) "LiF-BeF, of the peritectic compo-
sition 66-34 mole %, and (3) Hitec, a commercial
nitrite/nitrate heat transfer salt. The significant physical
properties of these three salts are listed in Table 6.3,

91

—— 317 years

and each is compared with a steam-water system in
Table 6.4.

Although water appears to be a very attractive
coolant, provided a double barrier is used in the drain
tank cooling tubes to avoid thermal shock following a
salt drain and to give better assurance that water could
not reach the fuel salt, calculations showed that it
would be difficult to fit the required number of tubes
into the drain tank head. A compromise was therefore
reached which employs natural circulation of an
7LiF-BeF, salt mixture through the drain tank tubes
and then cooling of the salt by radiative heat transfer to
boiling water. Heat transfer from the gas in the tank to
the "LiF-BeF; is by conduction and some internal heat
absorption; heat transfer from the salt is by convection,
conduction, and internal absorption. Selection of this
compromise arrangement was motivated largely by the
desire to have chemical compatibility between the
coolant and the fuel salt.

The layout of the drain tank cooling system is
indicated in Fig. 6.3 and in the flowsheet, Fig. 2.3. The
pertinent data are listed in Table 6.1,

The steady-state natural circulation flow rate of the
drain tank coolant salt was calculated as a function of
the heat load on the system. The method of calculation
involved iterating between the calculated thermal driv-
ing head and calculated head losses due to piping and
fittings until a flow rate was obtained which made those
two quantities equal. The coolant inlet temperature was
fixed at 900°F (freezing point 856°F). The other
system temperatures are functions of flow rate for a
given heat load. Salt density and viscosity were reeval-
uated for each successive value of flow rate (and hence
temperatures) and then used in determining the heads
mentioned above. Figure 6.5 is a plot of fraction of
design flow (that corresponding to design heat load)
and salt temperature at the U-tube outlet as functions
of fraction of design heat load. During normal reactor
operation the heat load on the drain tank due to the
off-gas and noble metals is about 16 MW, or about 25%
of the design value. It is noted from the figure that
about 55% of design flow is obtained at this heat load.
This is particularly advantageous because a drain will
not require the system to be accelerated from a very
low flow or from a static condition, as would be the
situation if the drain tank were not used for off-gas
holdup.

The drain tank coolant salt is cooled in 40 salt-
to-water heat exchangers located about 60 ft above the
drain tank to provide the thermal driving head for
natural circulation. The heat exchange is entirely by
radiation from salt tubes to a plate (or tubes) in which
Table 6.2. Qualitative comparison of fuel-salt drain tank coolants

General ability 1o Ease and 2
transfer heatat IE“"I ::‘ reliability Compatibility Compatibility e coc“‘ end R"::“ - i G :‘m
Coolznt temperatures and " of retention if mixed with if spilled in plexity Corrosion Conclusions
5 required coolant of components thermal and
rates required for ol in closed fuel salt cell e M tability controlling
an MSBR system b - - system g -~
Water and steam Excellent as Excellent  Good; pressure Poor Poor; generates high Very low  Low Good Good; protect against Requires Inconel o~ Sutiable for cooling,
coolant; poor can get high pressure and some freezing and high stainless steel and possibly best
for adding heat H; + 03 gas steam pressure special treatment
Liquid metals, Good for cooling  Good Excellent; low Poor Good if N4 in cell; Low Medium to high Excellent Excellent; protect Stainless steel or Excellent for heating
NaK and Na or heatiny pressure, low poor if ir against freezing possibly carbon or cooling mn absence
corrosion steel of air
Fused salts, fuorides, Good [or cocling  Good Excellent; low Excellent to Excellent; protact Low to Medivm 10 high Excellent to Good; protect against Requires stainless Excellent for heating
carbonates, and or heating pressure, low good against toxic vapors high good freezing steels or Hastelloy N or cooling
nitrate-nitrites corrosion
Organics, diphenyls  Good for coaling; Excellent  Good; pressure Poor Poor; generates high- Low Medium; requires Fair; tends to  Good; protect against Carbon steels Fair
and polyphenyls poor for heating can get high pressure, flamable, purification system decompose freezing and high
and toxic vapor above 700°F  pressure
Gases, COy, Na, Poor for heating  Poor Good; pressure Exesllent Good to excellent, Low Medium to high; Excallent Excellent Carbon steels Fair
and argon or cooling is high depending on ratio of large high-pressure
velume of pressucized systems
gas to volume of cell
or provisions for

pressure reliel

O
[38)
93

Table 6.3. Properties of possible fused-salt coolants for drain tank system

NaBF4-NaF TLiF-BeF, KNO;-NaNO;-NaNO,
eutectic peritectic eutectic

Composition, mole % 92-8 66-34 5340-7
Viscosity, Ib ft™! hrt

At 900°F 4.1 404 3.1

At 1150°F 2.6 18.7 2.5
Liquidus temperature, °F 725 856 228
Density, 1b/ft?

At 900°F 1194 125.5 108.0

At 1150°F 112.9 121.9 103.0
Specific heat, Btu Ib~! (°F) ! 0.36 0.57 0.37
Thermal conductivity, Btu hr ™! ft™! (°F) 7! 0.27 0.58 0.33

Table 6.4. Evaluation of salt-type coolants and water-steam for primary drain tank cooling system

Coolant Desirable features Undesirable features
NaBF4-NaF Inexpensive (~$70/ft3) Reactor must be shut down if coolant gets into
High melting point means reduced thermal shock fuel salt, and the fuel must be processed
on drain tank High melting point makes freezing in stack
Relatively low viscosity more likely
Hastelloy N would be required in coolant circuil
7LiF-BeF, No processing of fuel salt is required in event of Very expensive (~$1500/ft3)

leak
Least thermal shock on drain system
Extensive experience with this coolant in MSRE
Hastelloy N may not be required in coolant circuit

High melting point
High viscosity

No volume change on freezing

KNO3-NaNQ,-NaNO; Inexpensive

Carbon steel can be used up ta 850°F; stainless

steel for higher temperatures
Low melting point

Water-steam Has least danger of freezing
Lowest cost

Used in MSRE drain tank

Of doubtful stability at high temperatures and
in radiation (eld
Salt processing on leak may be required

Requires double barrier tubes (e.g., bayonet)
Relatively large number of tubes required

Relatively easy to get natural circulation

low-quality steam is produced at about 100 psia. Steam
separators divert the steam to one of several air-cooled
condensers located in a natural-draft stack. The con-
densate is returned by gravity to provide a circuit which
operates entirely by natural circulation. (A similar
system demonstrated satisfactory performance in the
MSRE drain tank.) Preliminary calculations indicate a
stack height of about 400 ft and an average diameter of
about 60 ft. Use of elliptical-shaped tubes and an
increase in longitudinal pitch would possibly permit
reduction in the stack height. During some months it

may be necessary to preheat the air or use other
methods to prevent freezing of the condensate in the
coils during periods of light load on the plant. Water-
cooled condensers could obviously be substituted for
the air-cooled coils and stack if an assured source of
cooling water were available at a particular site.

In calculating the transient temperature behavior of
the drain tank and associated cooling system, the
system was divided into a number of nodes, and
appropriate energy and momentum balances were
written. Allowance was made for the time variation of
ORNL-DWG 70~-11957

1.0 1300

&
wl W
& o8 1200 &
g / )/ =
> ! T
5 ._Af = &
> = 1100 5
o / o
: / =
(=) >

w
5 1 - 1000

n |
& @
o ©
g -
& 900 5

o

o

Q

0 BOO
0 0.2 0.4 0.6 0.8 1.0

FRACTION OF DESIGN HEAT LOAD

Fig. 6.5. Effect of heat load on drain tank cooling salt flow
rate and exit temperature.

the heat sources and of the salt volume in the tank as it
fills. The allowable stress limitations on the Hastelloy N
were based on long-term creep restrictions. The dura-
tion of the transient temperatures exceeding the present
design maximum of 1300°F was found to be only a few
hours due to the rapid decay of the heat sources in the
salt.

One area of uncertainty in the above analyses is the
heat transfer coefficient between fluid and tubes.
Although the literature contains some work on the
subject of heat transfer with fluids containing sources,
little information is available for the case of open
lattices. This is an area of experimental investigation
which remains to be done before reliable temperatures
can be calculated for the drain tank. Another area of
uncertainty is the amount of noble metals which will
adhere to surfaces in the drain tank. For heating
purposes, it was assumed that all the noble metals
present in the off-gas remain in the drain tank, When
the salt is drained, some of these could be washed off
the vertical surfaces and agglomerate on the bottom of
the tank. If the salt becomes more oxidizing, the noble
metals will go back in solution. Niobium is the first to
be oxidized. After its oxidation is complete, the other
noble metals will oxidize more or less together, A final
area of uncerfainty is the disposition of the daughters
of the gases which decay in the drain tank. Heat from
the decay of daughter products has been included in
calculations. These daughters would be expected to go
back into the salt when it is drained, and some fraction

94

should dissolve in the 600 gpm which is circulated
through the drain tank for cooling purposes.

Calculations for the salt-to-boiling-water heat ex-
changers indicate that 0.3 kW of heat per foot of salt
tube can be transferred. This assumes surfaces which are
well oxidized Lo give emissivities of 0.8, a sait tube
surface temperature of 1200°F at maximum heat load,
and a steam tube or plate-coil surface temperature of
250°F. The salt was assumed to be located inside %-in.
tubes spaced on a 1.5-in. square pitch, with the plate
coils interspersed between tube rows. Each of the four
heat exchangers would require 333 tubes, each 10 ft
long, to handle a total of 40 MW of power. It is
estimated that under the worst-case transient conditions
a maximum heat load of about 40 MW is all that these
exchangers could experience. This is due to the rapid
decay of the sources during the drain time and the
relatively long transit time of the drain-tank cooling
loops.

Calculations for the natural-draft stack were first
made assuming that the hot fluid was saturated vapor at
atmospheric pressure and using an ambient air inlet (dry
bulb) temperature of 100°F. Size and pressure drop
data were based on the data of Zimmerman®? for
commercial fin-tube heat exchangers. Results indicated
that it was not feasible to use a natural-draft stack of
reasonable height because of the small draft available
from the low temperature differences involved. By
pressurizing the water system to 100 psia, the saturated
vapor temperature was raised to 327.8°F. This resulted
in a stack height of 400 ft, assuming a 100°F inlet
temperature and a maximum heat load of 40 MW. Such
a stack would have to be about 60 ft in diameter to
accommodate the commercial units on which the
calculations were based. Because of the low gas temper-
atures, only a minimum amount of stack insulation
would be needed.

6.4 FUEL-SALT DRAIN TANK WITH
NaK COOLING

6.4.1 Introduction

There are several features which could be improved in
the salt-cooled primary drain tank system described
above, such as the general complexity and the need for
a relatively tall natural-draft stack. A restudy of the
conceptual design of the drain tank system led to
favorable consideration of an alternate arrangement in
which NaK is vsed as the coolant. In addition to
eliminating the stack, the revised concept is believed to
provide a more dependable emergency cooling system
and to offer other improvements. There was not
sufficient time, however, to carry the study of the
alternate design to as great a detail as was possible for
the initial system.

NaK can be heated to relatively high temperatures
and can experience significant radiation fluences with-
out problems of dissociation or high vapor pressure.
Since its density and viscosity variations with tempera-
ture are favorable for natural circulation in the system,
no auxiliary power or action by the plant operators is
required to initiate and maintain circulation. The use of
NaK and placing primary emphasis on radiant heat
transfer (which varies as the fourth power of the
absolute temperature) accommodate the wide ranges of
temperature and heat loads which may be encountered,
such as the factor of 3 difference between the normal
off-gas heating load and the maximum transient after a
sudden salt drain. The NaK is compatible with Croloy
or stainless steel and does not require the more
expensive Hastelloy N used in the salt systems. Since it
has a eutectic temperature of about 10°F, it is liquid at

95

room temperature, and no preheating of the NaK
circuits prior to filling is required.

6.4.2 Description

A schematic flowsheet for the NaK-cooled system is
shown in Fig. 6.6, and the design data are given in Table
6.5.

The alternate drain tank design is for a vessel about
14 ft in diameter and 20 ft high, with the bottom head
containing a plenum for the jet pumps used to transfer
the salt. It is constructed of Hastelloy N, and the vessel,
internals, and supports are designed for the reference
earthquake loading referred to in Sect. 14. There are
1028 Hastelloy N thimbles extending vertically down-
ward into the tank. Each 2-in.-ID thimble contains a
concentric Croloy 2'/; (or stainless steel) bayonet tube
in which eutectic NaK circulates by natural convection
to a bank of NaK-filled tubes inserted in horizontal
tubes which are immersed in a pool of water at an
elevation about 60 ft above the drain tank. The

ORNL—DWG- 70~ 11558

PRIMARY SALT PUMP

T REACTOR [ A l “ |
VESSEL — } Y a
J =) 3 T —e-T0O HEAT EXCHANGER
i Y ~GAs H
D- "X SEPARATOR LI/ "
~— — “ BUBBLE ' P —T
ol GENERATOR
| [—‘" HEAT c:gr%n—ol
./ EXCHANGER PROCESS -
_MAKE UP - — L——m SUPPLY
g WATER — — e —
B R P ennd) CATCH BASIN
?J WATER POOL \‘:4"—"'-—" T0

-4 CHEMICAL It

= G PROGESS

— E M PUMP H -

e —— | — :
== - l - )
VENT ==}y L — TO OFF=GAS SYSTEM l !——;
" I
A o
PRIMARY SALT DRAIN TANK CHEMICAL PROCESSING
:3 AND GAS HOLDUP
(I f
= !i

E

-G

Fig. 6.6. Simplified flow diagram of primary drain tank and heat-removal system using NaK as the coolant.
Table 6.5. Design data for NaK-cooled drain tank

Salt capacity, ft? 2500
Outside diameter of vessel, ft 13.8
Overall height, [t 20
Design heat load, MW(1)
Due to offigac decay 18
Maximum transient after sudden salt drain 53
Emissivity of Hastelloy N surface 0.55
Emissivity of surface coated with iron or 0.90
calcium titanate
Emissivity of water-tube surface (oxidized 0.75
steel or copper)
Conductivity of nitro%en gas in annuli, 0.026-0.031
Btuhr ! ft ™! (°F)”
Thimble surface area, based on temperature of 10,700
850°F under off-gas load conditions, ft?
Thimble 1D, in. 2
Number of thimbles required 1028
Average NaK temperature under off-gas load, °F 400
Average NaK temperature under maximum 640
load, °F
Maximum thimble temperature, °T 1400
Number of autonomous NaK circuits selected 10
for this study (number is optional)
NaK flow rate per circuit under off-gas heat 801
load of 18 MW(t), gpm
NaK flow rate per circuit under maximum heat 1081
load of 53 MW(1), gpm
Hot- and cold-leg temperatures, °F
Under off-gas load of 18 MW(1) 436-350
Under maximum heat load of 53 MW(1) 726-550
NaK circuit pipe size (sched 40), in, 12
Assumed length of each leg of circuit, t 100
Assumed elevation difference, center of drain 60

tank to center of water pool, It
Temperature of receiving surface of thimble in 232
water tank, °F

Heat transfer area of NaK tubes in water 83,1007
tank, ft?

Total water boiled from water tank, ft3
2 days after shutdown 24,000
10 days after shutdown 81,000
35 days after shutdown 192,000

Makeup water required for 18 MW(t) heat 126

load, gpm

AThimbles with internal fins can be considered as a means
of reducing the total length required, but fabrication of this
special tubing would probably require development by the
manufacturer.

arrangement provides a double barrier between the fuel
salt and the NaK and between the NaK and the water.

The drain tank is surrounded by essentially an
open-topped stainless steel vessel about 14% ft in
diameter and with a 3- to 4-in.-thick wall provided with
two autonomous internal cooling channels for circula-
tion of NaK. The outside tank serves as a backup in
event a salt leak develops in the drain tank and also as a

gamma shield. Heat transfer by radiation across the
3-in.-wide annular space between the vessels cools the
walls and lower head of the drain tank. In event of a
major leak of fuel salt into the external vessel, the NaK
circulating through its walls would provide the neces-

H winntama nf tha radao: nnd Avnies +nels
sary CCC!.!‘.g. An ad-unsubv Oi ull T8GLGIENCA Graiil wan

and use of the cooled outer vessel is that the internal
liner for the drain tank, as was used in the salt-cooled
system, is eliminated.

Heat generation in the drain tank due to radioactive
decay of off-gases and entrained particulates is about 18
MW(t) during normal full-load operation of the MSBR,
as discussed in Sect. 6.3.2. The maximum transient heat
release is about 53 MW(t), which would occur after a
sudden salt drain. About 60% of the energy release is in
the form of gamma rays, much of which will be
absorbed by the vessel walls or by the bayonet tubes
and thus be directly transferred into the NaK. Most of
the generated heat is removed by the cooling thimbles.
Heat is transferred from the thimble wall to the
NaK-cooled bayonet tube by radiation, although some
will be conducted by the nitrogen which fills the
0.1-in.-wide annular space between the two. The
thermal-radiation-receiving surface on the NaK tubes is
assumed to be coated with iron or calcium titanate to
afford an emissivity factor of about 0.9. Since the
thimbles and bayonet tubes are not in physical contact
either in the drain tank or in the water pool, a leak in
any system is unlikely to contaminate another.

The NaK cooling system is arranged with several
autonomous circulating loops, so that failure of one
circuit would not cause a severe loss of cooling capacity
and necessitate an immediate shutdown of the plant.
Ten separate loops were assumed in the preliminary
study reported here. As indicated in Fig. 6.6, an
electromagnetic pump (acting as a brake) is installed in
each of the NaK circuits to retard or stop the NaK
natural circulation as necessary to protect against
freezing of the fuel salt in the drain tank. This
arrangement is particularly advantageous during startup
or partial load operation.

The arrangement of the heat transfer surface in the
water pool has not been studied in detail but would
probably be somewhat as indicated in Fig. 6.6. Heat
transfer would be by radiation and gas conduction from
the outside surface of the NaK-filled tubes to the inner
surface of the concentric tubes which are submerged in
water. The water would boil and require either con-
densation and return or a continuous makeup of about
126 gpm of treated water under normal full-load
reactor power and even larger amounts under con-
ditions of a sudden reactor drain. Its water storage
capacity, however, can be made large enough to
accommodate the decay heat for a protracted period
even without water makeup. This arrangement provides
a reliable heat sink, is not dependent upon a power
source, and may be more earthquake resistant than the
natural-draft stack used with the salt-cooled drain tank
system.

6.5 FUEL-SALT STORAGE TANK

A storage tank is provided for the fuel salt in event it
is necessary to carry out repairs on the fuel-salt drain
tank or its associated components or piping. Although
the tank is located in the chemical processing cell, it is
not used as a part of the chemical system, since the
tank does not have a heat-removal system capable of
handling the high volumetric heat sources in the
chemical system. The storage tank will be the same
regardless of the type of cooling used for the primary
drain tank.

97

The tank has a storage capacity of about 2500 ft* and
may be constructed of 304L stainless steel rather than
Hastelloy N, since the tank will have a low use factor.*
The tank is connected into the drain tank system as
shown in the flow diagram, Fig. 2.1. Centrifugal and jet
pumps will transfer salt into and out of the storage
tank.

The tank has a heat-removal capacity of about 1
MW(t), which is provided by boiling water in 12-ft-long
U-tubes, with the steam being condensed in an air-
cooled condenser in the same manner as was used in the
MSRE system.”**?® The heat-removal capacity is based
on allowing about a 100-day decay period for both the
salt and the noble metals.

*To be conservative in the feasibility study, Hastelloy N was
specified for several portions of the MSBR systems where
stainless steel would probably have been acceptable. A test loop
constructed of 304L stainless steel has operated with 1200 to
1300°F fuel salt for more than 60,000 hr with a corrosion rate
of 1 mil/year, or less, and the rate is decreasing. !
7. Reactor Off-Gas System

A. N. Smith

7.1 GENERAL

The function of the primary off-gas system is to
reduce the concentration of undesirable contaminants
in the primary system off-gas stream to a level low
enough to permit continuous recycle of the helium
carrier gas to the primary system. The term “un-
desirable contaminants” includes gaseous and gas-borne
fission products, fission product daughters, water,
oxygen, hydrocarbons, etc. The off-gas system also
includes the equipment for handling all the associated
functions, such as dissipation of decay heat, collection
and storage or disposal of stable and long-lived gases,
liquids, and solids, and recompression of the recycle
gas. As shown in Fig. 7.1, the boundaries of the off-gas
system on the upstream side are defined as the outlet of
the particle trap in the gas flow leaving the fuel-salt
drain tank and, on the downstream side, as the outlet of
the accumulator tanks supplying helium to the bubble
generators and the purge flow for the salt-circulation
pumps.

The fission yields of noble gases (krypton and xenon)
are such that nearly one atom of gas is produced for
every atom of 223U which fissions. Since the fission of
1 g of uranium is roughly equal to 1 MWd, the MSBR at
2250 MW will produce more than 1 kg of noble-gas
fission products per day. About 15% of the gaseous
fission products are relatively short-lived and will decay
in the fuel-salt system. The remaining 85% are either
stable or have half-lives which are long enough for them
to be removed at the gas separator along with the
helium carrier gas. Continuous decay processes will
produce nonvolatile or slightly volatile daughter
products which may deposit on duct or vessel surfaces
or which may be carried along with the gas stream in
the form of smokes or mists until removed by filtration
or adsorption. In addition to the kryptons and xenons,
the carrier gas which leaves the gas separator is expected
to contain Iritium, oxygen resulting from fluorine

98

burnup, noble metal fission products, and a small
amount of entrained fuel salt.

The nonvolatile fission products either will deposit in
the primary system drain tank or will be removed by
the filter at the outlet of the drain tank, so that the
off-gas stream at the inlet to the off-gas system will
consist primarily of gaseous components. On a volume
basis, the contaminants in the stream are expected to be
on the order of 0.1%, or about 1000 ppm. This number
is based on a flow from the gas separator of 11 scfm,
stable noble-gas yields of 7% for krypton and 21% for
xenon, and a recycle rate of 80% from the 47-hr xenon
holdup system to the bubble generator. As the gas
stream passes through the off-gas system, the decay of
the radioactive noble gases and daughters will continue,
as will also the attendant necessities for heat dissipation
and materials collection and disposal. The amount of
decay heat per unit volume will be high at first but will
drop off rather quickly during the first hour due to the
rapid disappearance of the short-lived isotopes, as
shown in Fig. 7.2.

An estimate was made of the distribution of fission
product decay heat in a 1000-MW(e) MSBR off-gas
system. The calculations were based on the following
model:

1. The flux of krypton and xenon into the off-gas line
was to be as calculated by Kedl for a 0.56% poison
fraction (see Table A.2), Solid daughters of krypton
and xenon were assumed to plate out at the point of
formation,

. A 2-hr residence time in the drain tank was assumed
between the outlet of the reactor system and the
inlet to the 47-hr xenon holdup system.

. Krypton delay in the charcoal beds was assumed to
be one-twelfth of the xenon delay.

. The off-gas system was divided into 20 regions in
which the radioactive noble gases were assumed to
PURGE FLOW TO
OTHER PUMES PRIMARY SYSTEM —~=—

99

ORNL-DWG 70-41959

- QFF-GAS SYSTEM

I
| |
|
l

ACCUMULATOR |
c ? | SURGE GAS CLEANUP
FROM OTHER TANK 8 MONITOR
SEPARATORS ———e— I = 0
ENTRAINMENT
SEPARATOR
LONG-DELAY
%e HOLDUP
FUEL-SALT |
CIRCULATING .~
PUMP — l
GAS
$ SEPARATOR i e
e |
!/PmncLE TRAF
1 (IF REQUIRED)
| 47-Hr Xe 1 cfm
! I HOLDUP SYSTEM
i
FUEL-SALT I l
DRAIN TANK ! |
: I S cfm
‘ | EREIRATNY —E_ SURGE CHEM. TRAP
TANK B MONITOR
TO OTHER BUBBLE '
GENERATORS COMPRESSOR
Fig. 7.). Schematic flow diagram MSBR off-gas system.
decay exponentially in accordance with an assigned 7.2 BASIC ASSUMPTIONS AND
delay or residence time. The 2-hr volume holdup and DESIGN CRITERIA FOR STEADY-STATE
the 47-hr xenon delay charcoal bed were divided OPERATION

into compartments with various delay times in an
attempt to obtain approximately equal heat loads.
The delay times for the pipe sections were arbitrarily
set at 18 sec each. The results of this calculation,
shown in Fig. 7.3, were used in estimating heat loads
in the various sections of the off-gas system.

With regard to iodine in the MSBR, the iodine
isotopes produced directly by fissions will remain with
the fuel salt. Much of the tellurium (the precursor of
iodine) will probably deposit on surfaces as noble metal
particulates, but significant amounts could be swept
into the off-gas system. Here, upon decay of the
tellurium, the iodine will be quickly trapped as it
contacts the charcoal in the adsorber beds. Effluent gas
from the beds is normally recycled, and none is vented.
(The decay heats from the iodine nuclides of concern —
those with half-lives greater than 10 min — are shown in
Table 7.1.)

The following assumptions were made in the design
study of the off-gas system:

1. Reactor power is 2250 MW(t), and the fuel is
233
U.

2. The carrier gas is helium, with a total flow to the
off-gas system of 11 scfm. This total is the combination
of flows from each of the four pump loops, consisting
of 2.25 scfm from each of the gas separators and 0.5
scfm of purge gas for each of the pump shafts. Net flow
of fission products and materials other than helium is
about 0.1%, or 0.01 scfm.

3. The atom flow rates of krypton and xenon into
the off-gas system are based on calculated atom flow
rates at the gas separator discharge, with appropriate
corrections for a 2-hr residence time in the fuel-salt
drain tank. All solids which are gas-borne at the outlet
of the drain tank (including noble metals, salt mist, and
100

ORNL-DWG 70-11960R
2 L LB 1 1
NOTE: {. PLOT IS BASED ON CALCULATIONS
BY BELL USING INPUT CALCULATED
BY KEDL. (REF, 9)
2. DAUGHTERS ARE ASSUMED TO PLATE
OUT AT POINT OF FORMATION.
NOBLE METAL DECAY HEAT IS NOT

L L1

- INCLUDEU.
.
-
wn
z
@ 2 i \
o
.~ Kr AND Xe + DAUGHTERS AT
3 10—\ EQUILIBRUM
- A \\
> —p N
% 5 AN N
N
X o
2
o
5 \\
Kr AND Xe ONLY '\
0 \
L [9) > ~3
.. — >
5

0 0.2 0.4 0.6

DECAY TIME (hr)

0.8 1.0

Fig. 7.2. Decay heat vs time for 1000-MW(e) MSBR off-gas
stream using 233 U fuel.

solid daughters of the noble gases) will be removed by a
filter before the gas stream enters the off-gas system.
The total yield of tritium (> H) from all mechanisms will
be 2400 Ci/day, and all tritium will remain in the
off-gas stream; that is, for the purpose of studying the

RESIDENCE TIME (hr) 0.0l 0ol 0.008

DECAY HEAT (Mw) GAS + DAUGHTERS 0.002 0.002 0,001

DECAY HEAT (Mw) GAS ONLY 0,001 0.004 0.0005
SURGE

TANK sl

off-gas system, the rate of diffusion of tritium through
vessel and pipe walls is assumed to be zero. (Tritium
diffusion rates are discussed in Sect. 3.3.7.)

4. The gas will enter the off-gas system at 15 psig; 9
scfm will be returned to the bubble generators at 5 psig,
and 2 sefm will be returned {0 the puige gas hedder at
45 psig,

5. At least two barriers, or containment walls (one of
which is the wall of the gas duct or vessel), will be
provided to guard against leakage of radioactive off-gas.
Shielding will be provided for attenuation of penetra-
ting radiation to permissible levels. Instrumentation will
warn of excessive leakage of gas or penetrating radia-
tion,

6. The target reliability of the system is 100%; that
is, spare units will be provided, and the maintainability
of units will be such that predictable failures in the
off-gas system will not result in shutdown of the reactor
or loss of the contaminants to the environment.

7.3 SUMMARY DESCRIPTION OF
OFF-GAS SYSTEM

The flow of gas in the primary system can be
represented by two recycle loops, a 47-hr xenon holdup
loop and a long-delay (~90-day) xenon holdup loop.*
The 47-hr loop circulates through the bubble generator
and gas separator to strip the ' 25 Xe from the fuel salt;

*These holdup times do not include the 2-hr residence time of
the off-gas stream in flowing through the primary-salt drain
tank.

ORNL-DWG 69-54868R

0.0 0.005 6 M § B 2 4
0003 0.0014 0.42 056 050 037 047 OM
0.00! 0.0006 0.24 034 033 0.25 042 0,07
SURGE " 47=Nr XENON HOLOLP
TANK PIPE CHARGUAL BED
1 4 4
~2160 720
0.23 0.004
043 0.004
LONG DELAY Gas
CHARCOAL BED CLEANUP

Fig. 7.3. Distribution of decay heat in MSBR off-gas system. The residence time for krypton in the charcoal beds is about
one-twelfth of that for xenon. The flux of atoms into the off-gas line was calculated by Kedl (see Table A.2 and ref. 9). Decay heat
calculations, assuming a 2-hr holdup in the primary drain tank, were made by Bell and were based on previously reported values for

1-hr holdup,?
101

Table 7.1. Decay heat from iodine nuclides

¥y Total heat :
. . Fraction Decay
lodine Half-li cumulative per b
" a -life . o3 > " gamma heat
isotope fission yield disintegration heat (MW)
(%) (MeV)
131 8.0 days 2.90 0.7 0.57 0.23
132 24 hr 4.54 2.7 0.78 1.36
133 20.8 hr 5.78 1.1 0.51 0.71
134 52.5 min 5.75 33 0.81 2.11
135 6.68 hr 5.05 23 0.84 1.29
Total 5.70

40nly those nuclides with half-lives greater than 10 min are included.

bEquilibrium conditions are assumed, that is, the decay rate is equal to the
fission yield. The yield in atoms per day = 6.22 X 10?2 Yi, where Y7 is the yield
in percent. The decay heat in MW =0.11 YiQ, where Q = MeV/disintegration.

Table 7.2. Flow of principal gaseous components to off-gas system of 2250-MW(t)

single-fluid MSBR
Siotoie . Docay Cux;l;:lntive Flow X 107>% (atoms/hr)
Half-life U Out of Entering Leaving
Mass constant . d
Element N (ty)2) (hr ) fission yield reactor 47-hr Xe 47-hr Xe
At (%) system holdup holdup
H 3 112.26 years 645%x 107 0.8 0.21 0.21 0.21
Kr 82 Stable 0 0.3 0.35 0.35 0.35
83 Stable 0 1.14 1.50 1.50 1.50
84 Stable 0 1.10 1.45 1.45 1.45
85 10.76 years 735 107° 2.49 3.28 3.28 3.28
86 Stable 0 3.28 4.32 4.32 4.32
87 76 min 0.55 4.50 1.15 0.63 3.0x 107
88 2.80 hr 0.25 5.70 1.75 1.33 0.40
89 3.18 min 13 6.23 0.92 1.2 x 10712 0
T Kr 14.7 12.9 11.3
Xe 128 Stable 0 0.02 0.025 0.025 0.025
129 Stable 0 2.10 2.76 2.76 2.76
130 Stable 0 0.10 0,13 0.13 0.13
131 Stable 0 3.85 5.05 5.05 5.05
132 Stable 0 5.48 7.20 7.20 7.20
133 5.27 days 548 % 107 6.48 4.30 4.25 3.30
134 Stable 0 6.83 9.00 9.00 9.00
135 9.14 hr 0.0753 6.16 1.50 1,29 0.040
136 Stable 0 7.00 9.20 9.20 9.20
137 4.2 min 9.9 7.16 1.16 5.1% 10710 0
138 17 min 245 6.63 1.45 0.01 0
T Xe 41.7 38.9 36.7
Jotes:

1. Tritium (SH) yield includes yield from lithium burnup and assumes zero diffusion through vessel and pipe walls.

2. Isotopes with half-lives of less than 2 min are not shown.

3. A residence time of 2 hr is assumed between outlet of reactor system and inlet to 47-hr xenon holdup system.

4. Fluxes for krypton and xenon are taken from calculations by Kedl and Bell (ref. 9). The stable and long-lived isotopes appear
at five times yield due to the 80% recycle stream,
the long-delay loop carries the balance of the gas flow
in the fuel system. The two loops are joined together at
the salt entrainment separator and flow cocurrently
through the primary drain tank and the 47-hr holdup
system, as shown in Fig. 7.1.

The cocurrent stream enters the primary off.gac
system at the discharge from the fuel-salt drain tank.
The tank will probably serve as an efficient collector of
particulates in the gas, but if it proves necessary a
particle trap, or filter, can be added, as shown in Fig.
7.1. At this point the gas will have been stripped of
nongaseous components (noble metals, salt mist, and
nongaseous daughters of the noble gases), so that the
primary contaminants are Kr, Xe, and *H. About 2 hr
will have elapsed since Lhe gas first left the fuel-salt
system. The gas first passes through the 47-hr xenon
holdup system to provide a residence time for xenon
molecules sufficient to permit the '*Xe to decay to
about 3% of the inlet amount. The 47-hr holdup system
will utilize charcoal for the dynamic adsorption and
holdup of krypton and xenon. The decay heat will be
transferred to boiling water.

At the outlet of the 47-hr system the gas stream is
divided into the two recycle loops. In the 47-hr recycle
loop, 9 scfm, or about 80% of the total flow, passes in
succession through a chemical trap and alarm system, a
compressor, and a surge tank. From the surge tank the
gas is metered to the bubble generators at the four
circulating pumps. In the second recycle gas stream, 2
scfm, or 20% of the total flow, passes first through the
long-delay xenon holdup system, where the residence
times for krypton and xenon are sufficiently long to
allow all radioisotopes except the ten-year ®SKr to
decay to insignificant levels. The gas then passes
through a purification system which reduces the level of
any remaining contaminants (*¥Kr, *H, stable isotopes
of Kr and Xe, water, hydrocarbons, etc.) to an
acceptable level, then through a surge tank, a com-
pressor, and an accumulator, and finally is returned to
the primary system.

Table 7.2 shows the flow of tritium and noble-gas
isotopes at the outlet of the reactor system and at the
inlet and outlet of the 47-hr xenon holdup system. The
flow rates at the outlet of the reactor system are based
on calculations by Kedl® assuming a 0.56% Xenon
poison fraction. The second and third flow rate
columns in Table 7.1 are based on calculations by Bell®
using the first column as input and assuming: (1) simple
exponential decay, with a 2-hr residence time between
the reactor system outlet and the inlet to the 47-hr
xenon holdup system; (2) a residence time for krypton
one-twelfth that for xenon; and (3) 80% of the flow

102

from the 47-hr xenon holdup system to be recycled to
the primary system. (For stable and long-lived isotopes
the effect of the recycle flow is to increase the total
flow by a factor of about 5.)

7.4 THE 47-hr XENON HOLDUP
SYSTEM

The 47-hr xenon holdup system provides residence
time for xenon isotopes to reduce the concentration of
135Xe in the effluent. The design criteria for the
system are as follows:

1. The residence time for xenon is 47 hr. This time is
exclusive of the volume holdup in the primary system
drain tank and other vessels and ducts. A 47-hr delay
time permits 97% of the 9.14-hr ' ?5Xe to decay.

2. The estimated heat load, based on Fig. 7.3,is 2.14
MW, 42% of which is due to daughter-product decay.
The design capacity of the heat-removal system is 125%
of calculated, or 2.7 MW.

3. A dynamic adsorption system is used for delay of
the xenon. The adsorbent is activated charcoal, with
transfer of the decay heat to boiling water. The design
temperature of the charcoal duct wall is 250°F. The
average temperature of the charcoal is 340°F.

4, The assumed charcoal properties are: bulk density,
30 1b/ft*; thermal conductivity, 0.03 Btu hr™' ft™2
(°F)' ft; and size range, 6 to 14 Tyler sieve series (%
to 3/64 m) -

5. The decay heat distribution is obtained from the
calculations by Kedl and Bell,” as shown in Fig. 7.2,

6. The efficiency of the bed is assumed to decrease
with time due to accumulation of solid daughters, Spare
capacity is provided, and provision is made for replace-
ment of modules by remote maintenance techniques.

7. Carrier-gas flow is 11 scfm, and the overall
pressure drop is 5 psi. An estimate of the size of the
charcoal bed is obtained by using the empirical relation-
ship developed by Browning and Bolta:® ¢

k
th =-f4n-: (l)

where ¢;; is holdup time, m is mass of charcoal, [ is
volume flow rate of carrier gas at local conditions, and
k is a proportionality factor which is known as the
adsorption coefficient and which varies with the
carrier-gas composition, the adsorbent, the adsorbate,
and the temperature. For typical commercial charcoals,
Ackley and Browning” 7 have determined the following
relationship between k& and temperature for xenon at
temperatures between 32 and 140°F:
5880
k(Xe)=3.2X 107 exp R

ft*/1b . (2)
Equation (1) indicates that the holdup time increases
directly with k. However, an increase in holdup time
increases the heat generation, which results in an
increase in charcoal temperature and a decrease in k, in
accordance with Eg. (2). Note also that an increase in
temperature causes an increase in f (local flow rate),
which results in a decrease in holdup time. For any
given section of the bed, k and 7;, will seek equilibrium
values which are a balance between the opposing forces.

For the purpose of this estimate, the assumption was
made that Eq. (2) is valid up to 500°F and that the
average charcoal temperature is 340°F. Equation (2)
indicates that this temperature would be equivalent to
an adsorption coefficient of 0.5 ft*/Ib. For a holdup
time of 48 hr and a flow of 11 scfm, Eq. (1) indicates
that the required mass of charcoal would be 63,360 Ib.
[t should be noted that, within limits, the average
charcoal temperature can be adjusted by the pipe
diameter and the heat-removal capability. Due to the
complex interaction of variables, however, the optimum
system would not necessarily be the one with the
smallest mass of charcoal.

The physical concept for the 47-hr charcoal bed
would be similar to that proposed by Burch et al.®®
Hairpin tubes filled with charcoal are suspended in large
tanks. The decay heat is transferred to boiling water.
The steam is passed through an external condenser, and
the condensate is recycled. In an actual system, one
would use the largest diameter pipe which would permit

an acceptable average charcoal temperature. Smaller
diameters may be necessary at the inlet end, where the
decay heat rate is high. For this system, it is estimated
that 1'%4-in. pipe may be required for the inlet end, but
that 2-in. and possibly 3-in. pipe would be suitable for a
large portion of the bed. Figures 7.4 and 7.5 show how
the charcoal bed might be arranged, assuming the use of
2-in. pipe throughout and an excess charcoal capacity

ORNL-DOWG 70-11962

L & CONDENSER

'c" .

A .'}.)r.f\eg‘-:$.-‘|$' oy T . E
d \ SIS -
AN Tl AL e - y

'V‘;J

STEAM, ~230°F

- —— = ~ i
- . g 2t |

TR

=
— - R -
._‘
—

e i

OUTLET
HEADER

HEADER

-

- == = B N

25 ft

| - —

Fig. 7.5. Cross section of one bank of 47-hr Xe holdup char-
coal bed for MSBR (see Fig. 7.4).

ORNL-DWG 70~11964

EACH UNIT: :——-———--.—‘———..__‘
)

12 VERTICAL HAIRPIN SECTIONS

OF 530 LINEAR {t PER UNIT)
PACKED WITH CHARCOAL (BULK
DENSITY = 30 Ib/ft3)

He FLOW = 0.06 scfm/UNIT

Ap = 00| psi/ft PIPE
Apg/UNIT = ~5 psi
EXCESS CAPACITY = 30%

Hefm FROM
VOLUME HOLDUP

CELL:
6 x 65 x 25 t+ DEEP

TOTAL OF BOTH BANKS:

CHARCOAL = 82,900 |b
2-In. PIPE = 127,200 ft
HEAT REMOVAL CAPACITY = 2.7 Mwt

OF 2-in, PIPE 20 ft HIGH (A TOTAL ) e e

T F-|—==STEAM
: ! BANK NO. |
i !-Lp---wmen

—r )
k
L.. 9 sefm TO PRIMARY

4 SYSTEM
¢ e _FROM OTHER BANKS
— = STEAM

a2 scfm TO LONG-DELAY
CHARCOAL BED

BANK NO. 2
(IDENTICAL TO BANK NO. 1)
e WATER

(EXCESS CAPACITY = 25 %)

Fig. 7.4. Plan view MSBR charcoal bed — 47-hr Xe holdup (see Fig. 7.5).
Table 7.3. Accumulation of nonvolatile materials in
the 47-hr xenon holdup system?

Gaseous

Accumulation rate

til t

parent atoms/hr g-moles/day g/year NOTRIRIRUGRI
X 1044

8%y 0.33 0.13 4,015 6.2 % 10'? year ®Rb

88kt 0.65 0.26 8,395 Stable %88

133xe 0.97 0.39 18,970 Stable *33Cs

135xe 1.24 0.50 24,820 3% 10° year 35Cs

@The accumulation rate of the four isotopes is 56,210 g/year, If this amount is
distributed uniformly over the 30 tons of charcoal, the concentration is 0,002 g

of nonvolatile per gram of charcoal.

of 30%. There are 240 parallel units, arranged in banks
of 60 units each, with each bank containing 530 lin ft
of pipe. The mass of charcoal is 82,400 Ib, and the
length of pipe is 127,200 ft. The overall plan area
required is about 32 by 65 ft, and the pipes are
suspended in cells about 25 ft deep. The valves and
headers are located in smaller ducts, as shown in Fig.
7.5. A minimum of two containment barriers are
provided to guard against leakage of the radioactive
fission gas into areas which would be hazardous to
personnel. The condenser capacity is 2.7 MW, which is
25% over the maximum estimated heat load. The
estimated accumulation of nonvolatile materials is
shown in Table 7.3.

7.5 LONG-DELAY CHARCOAL BED

At the outlet of the 47-hr xenon holdup system the
off-gas flow is split into two streams, as shown in Fig.
7.1. One stream of 9 scfm is returned to the primary
system by way of the bubble generator, and the other
stream, of 2 scfm, is fed to the long-delay charcoal bed.

The function of the latter is to provide a relatively long
residence fime, so that the heat load and penetrating
radiation in the ensuing gas cleanup system will be at a
reasonable level. Table 7.4 lists the isotopes which have
the longest lives and hence are controlling in the bed
design. Figures 7.6—7.8 show the activity load and the
heat load in the gas cleanup system as a function of the
holdup time in the long-delay system. The assumed
design residence time is somewhat arbitrary since
whatever load is not handled by the long-delay bed
must be dissipated by the gas cleanup system. The
incentive, however, is to handle as much as possible
with the long-delay bed, since its construction and
operation would probably be more simple than that of
the gas cleanup system. The following criteria were used
in the design of the long-delay charcoal bed.

1. Holdup time for xenon is 90 days.

2. The heat load is 0.25 MW, based on calculations by
Bell and using input data provided by Kedl, as
shown in Fig. 7.2. The average heat load is 2 X 1072
kW per minute of holdup time.

Table 7.4, Longer-ived noble-gas fission products? exclusive of *H and **Kr

Average energy per disentegration

233
Half-life, 1) /5 Decay U (MeV)

Isotope constant fission

Days Hours (hl‘-l) yield (%) Beta Gamma Total

X102

13imy, 12,0 288 2.4 0,023 0 0.16 0.16
13850 5.27 126.5 5.5 5.78 0.12 0.08 0.20
1399 0.38 9.13 7.6 6.16 0.30 0.27 0.57
aSmy, 0.18 4.36 159 2.43 0.23 0.18 0.41
B8k, 0.12 2.77 250 5.84 0.33 2.1 2.44

AIncludes only the fission products having significant fractions remaining at the inlet to the gas cleanup system.
1020 ORNL-DWG 70-11963

19 X\ \ 14,8 days
10 —4

5 o
) 5
{ N\
2 —
1018 as
% X §
Y
o ¥

e e
1 Y
<1 L \ N
|

atoms/hr ENTERING GAS CLEANUP SYSTEM

‘O as > s
N
X
5 \
| |
[ 2.77 hr 88kr |
X i
10'®
0 20 40 60 80 100

Xe HOLDUP TIME IN LONG-DELAY CHARCOAL BED (days)

Fig. 7.6. Effect of charcoal bed holdup time on atoms per
hour of xenon and krypton entering gas cleanup system.

3. The physical properties of the charcoal are the same
as those noted in the description of the 47-hr xenon
holdup system, Sect. 7.4.

4. The gas flow rate is 2 scfm at an inlet pressure of 5
psig, and the design Ap is 5 psi.

5. The gas composition is 99.9% helium, with trace
quantities of contaminants, as described in Sect. 7.1.
Since noble-gas daughters will be deposited on the
charcoal during operation, there will be gradual
reduction in the effectiveness of the charcoal. About
30% spare capacity is provided to offset this loss in
effectiveness.

6. The heat will be transferred to cooling water. The
average temperature of the charcoal duct wall is
80°F.

The size of the long-delay bed was estimated using a
method similar to that used for the 47-hr xenon holdup

105

ORNL~-DWG 70-11964

10* ‘-\\
11
L
5 4,36 hr 2°Mkr
, \\
s \
W 403 \
w
9 \
5 1
o
& [BAY
Z \l\ sl
uJ \ 5.27 days '>3xe
3 A S — _
3. \
@ 9 !1_2 days ;?
e l “1
= )
G ) N \
8 ., J‘_\ X
g \
3 ot
° 10 ‘ == mmear T 3 =
5 \ k‘ AN
= \ N
, | 277 e Sk XN
10° \
0 20 40 60 80 100

Xe HOLDUP TIME IN LONG-DELAY CHARCOAL BED (days)

Fig. 7.7. Effect of charcoal bed holdup time on curies pe:
hour of xenon and krypton entering gas cleanup system.

charcoal bed. The results indicate that 3-in. pipe is a
reasonable duct size, and on this basis the average
charcoal temperature is 125°F, the mass of charcoal is
18.5 tons, the volume of charcoal is 1234 ft3, the
length of pipe is 24,060 ft, and the average heat flux is
41 Btu hr ™ ft™2. Figure 7.9 shows a proposed layout
for the system in an arrangement that requires a cell
about 60 ft long, 25 ft wide, and 25 ft deep. The unit
design is based on a calculated Ap of 0.005 psi/ft for a
helium flow of 0.07 cfm. Thirty-three percent spare
capacity is provided, and any unit may be isolated from
the rest of the system. The estimated charcoal pipe wall
temperature is 80°F with the decay heat transferred to
circulating water. The principal nonvolatiles ac-
cumulating in the long-delay charcoal are the four
isotopes listed in Table 7.5.
106

ORNL-DWG 70-11965
103

EACH UNIT:

e | 20 VERTICAL HAIRPIN
SECTIONS OF 3~in, PIPE

20 ft HIGH (A TOTAL OF
8OO LINEAR f+ PER UNIT)
PACKED WITH CHARCOAL
R ¥ I:\I-:NQITV = X0 u_\/n
A TOTAL OF 1233 |b/UNIT).
He FLOW IS 0.07 scfm /UNIT,
TOTAL Ap/UNIT = 5 psi
EXCESS CAPACITY = 30 %

\ ]
102 , rs e

e

\ BASED ON ADSORPTION
COEFFICIENT OF
= \7 £t/ Ib—atm

OVERALL CHARCOAL
BED DIMENSIONS:

\ ~25 % 60 % 25 ft DEEP

HEAT LOAD IN GAS CLEANUP SYSTEM (kw)

ELEVATION VIEW 1S
SIMILAR TO FIG.7.5
10° ISR P S
et HEAT REMOVAL

a3 CAPACITY = ~0,25 Mw(t)

107!

0 20 40 60 BO 100
Xe HOLDUP TIME IN LONG~-DELAY CHARCOAL BED (days)

ORNL—=DWG 7011966

FROM 47-hr Xe
\ HOLDUP SYSTEM 2 sctm"’[;:

g TOTAL OF 20

PAHALLEL
UNITS PER
BANK

3

- WATER

DA NU Y

~<«WATER

—e=T0 GAS

CLEAN-UP
SYSTEM

TOTALOF 20
PARALLEL
UNITS PER
BANK

- WATER

BANK NO. 2
(IDENTICAL
TO NO. 1)

~—-WATER

Fig. 7.9. Plan view of long-delay MSBR charcoal bed.

8SKr and the 12-year *H have decayed to negligible

amounts, Thus, if one assumes a delay time of 90 days

Fig. 7.8. Effect of charcoal bed holdup time on heat load in
gas cleanup system.

for xenon, the longest-lived isotope (12-day *?*!™Xe)
would be reduced to 0.6% of its original value, and the

reduction for the shorter-lived isotopes would be
proportionally greater. The stable noble gases, as well as

7.6 THE GAS CLEANUP SYSTEM

essentially 100% of the 35Kr and *H, will be carried

into the gas cleanup system at a rate equal to the rate of

After leaving the long-delay charcoal bed, the off-gas
stream enters the gas cleanup system. At this point, all
the radioactive fission product gases except the 10-year

Table 7.5. Accumulation of nonvolatiles in the long-delay

charcoal bed?
Accumulation rate
Gaseans Nonvolatile daughter
parent atoms/hr g-moles/day g/year
X 10%?
87k r 0.075 0.03 952 6.2 % 10'? year ®7Rb
88p, 0.51 0.20 6351 Stable 28y
L3850 3.3 1.3 63108 Stable '3
193 % 0.04 0.016 788 3 x 10° year ?5Cs

2The cumulative total for the four isotopes is 71,200 g/year. If this quantity is
distributed uniformly over 18.5 tons of charcoal, the concentration is 0.004 g of

isotope per gram of charcoal.

production in the reactor (assuming that no tritium is
lost to other parts of the reactor system by diffusion
through duct and vessel walls).
The function of the gas cleanup system is to process
the carrier gas to reduce the residual contaminants to a
level which will permit the effluent carrier gas to be
recycled to the reactor purge-gas system, Design criteria
for the gas cleanup system were as follows:

1. Carrier gas is helium at a flow of 2 scfm and an inlet

pressure of 20 psia. The design pressure drop is 4 psi.

The level of each contaminant in the effluent gas is
not more than 1% of the value at inlet. Table 7.6
shows the calculated isotopic flow rates at inlet for
the stable and very long-lived isotopes.

The gas contains some '*'"Xe, which is negligible
from a mass flow standpoint but which must be
considered in the design of shielding and the heat
dissipation system. The tritium values are based on the
assumption that the gas cleanup system receives all the
estimated total yield of 2400 Ci/day. This is un-
doubtedly a maximum figure, since a significant frac-
tion of the tritium may be transferred to other parts of
the reactor system by diffusion through duct and vessel
walls, as discussed in Sect. 3.3.7.

Upon entering the gas cleanup system, as shown in
Fig. 7.10, the off-gas first passes through a preheater,
which raises the gas temperature to 1500°F. It then
passes through an oxidizer, which converts the tritium
to *H,0, and then through an aftercooler, which
reduces the gas temperature to 100°F. (Both the

107

preheater and the aftercooler have heat loads of 3 kW
and are designed for negligible Ap due to flow. The
function of the aftercoolers is to reduce the heat load
on the ensuing components,) The off-gas then passes
through a charcoal-packed adsorber which is maintained
at 0°F. The *H,0 and the kryptons and Xxenons are
retained on the charcoal, while the carrier gas passes
through the bed. After leaving the refrigerated adsorber,
the carrier gas is recompressed and recycled to the
reactor purge system. In normal operation, two ad-
sorbers are alternated on a fixed cycle. A regeneration
process is used to transfer the adsorbed gases in the
off-stream unit to a receiver cylinder for permanent
storage.

The tritium oxidizer is 2 in. in inside diameter and 3
ft long, is packed with 13 lb of copper oxide, and
operates at 1500°F, The tritium flow is 0,036 ft*/day
with an allowable Ap of 2 psi. The CuO consumption at
breakthrough is 60%, and the operating life of a unit is
estimated to be 1000 days. Development work will be
needed to confirm the efficiency and pressure drop
estimates, however.

Each adsorber is made up of 16 pieces of charcoal-
packed 8-in. pipe with 1%-in. interconnections. The
total length of 8-in. pipe is 288 ft, arranged in two
branches to provide a Ap of 2 psi. The pipes are closely
stacked inside a 3- to 4-ft-diam pipe with a heated or
cooled fluid circulated in the interstitial spaces to

Table 7.6. Flow of isotopes into gas cleanup system

Isotope
. Yield Flow to gas cleanup Concentration
Mass e %) 3 (ppm. by volume)
Element No. atoms/hr g-moles/day ft* /day ppm, by volume
x 10%3
Kr 83 Stable 1.14 0.029 0.12 0.092 31
84 Stable 195 0.049 0.20 0.15 52
85 10.76 years 0.66 0.017 0.068 0.052 18
86 Stable 341 0.085 0.34 & 94
Total 0.73 0.56 195
Xe 131 Stable 3,39 0.085 0.34 0.27 94
132 Stable 4.54 0.11 044 0.36 125
134 Stable 5.94 0.15 0.60 0.47 163
136 Stable 6.89 0.17 0.68 0.55 191
Total 2.06 1.65 573
H 3 12.26 years 0.8 0.02 0.04 0.032 11
NOTES:

1. Calculations of flow to gas cleanup system based on carrier gas flow rate of 2 scfm.
2. Yield values for Kr and Xe isotopes may differ slightly from values shown in Table 7.1,
3. Tritium values are based on the assumption that all of the H production (estimated at 2400 Ci/day) goes to the gas cleanup system.
FROM LONG-DELAY
CHARCOAL BED

108

ORNL- DWG 70- 4967

TO HELIUM
SUPPLY SYSTEM

AGGUMULATOR .T
ey
PREHEATERS X !
I I TRAP NO. 3 |—L = N_W
|
3
H X ol
o
BT OXIDIZERS ouet i }
u |
I F[ ""N_:—_T“—\ TRAP NO, 2 je—t— | —pg—
AFTER- , X :
COOLERS L_’ 4 et
100°F | I : x (B G—X _‘i 2
—{><1—'——|———‘ TRAP NO. | : :><)-T
| | x
L.~ (TRAPS CAN BE HEATED s
: OR COOLED. NO. 3 SHOWN W o=
i IN STANDBY ) | |
TRAP TEMPERATURES: l ON-LINE '
—— ON STREAM = 0°F | Q-2 3cfm ot s !
—— REGENERATING = 500°F i |
3 Earaa |
' “%“ e i i St G e ‘V‘WI—‘E““
S ON-LINE [ |
: MONITOR
, LIQUID Ny TRAP
I

7.10. MSBR off-gas cleanup system,

provide an average on-stream operating temperature of
0°F and a temperature of 500°F when on the regenera-
tion cycle. Using an adsorption coefficient of 4.8 ft*/Ib,
the estimated total charcoal requirement is about 3000
Ib. The operating cycle is eight days — four days on
stream and four days regenerating.

The helium gas used for regeneration is taken from
the helium purge header, as indicated in Fig. 7.10.
During regeneration the gas flow is about 10% of
normal on-stream flow and moves through the adsorber
unit in the opposite direction. After leaving the heated
adsorber bed, the regenerating gas, now laden with
3H,0, krypton, and xenon, passes through a storage
bottle maintained at a liquid-nitrogen temperature of
—325°F. The water, krypton, and xenon are trapped in

the bottle, and the purified effluent is returned to the
main carrier-gas stream. Assuming a storage bottle
similar to a 1.5-ft® high-pressure gas cylinder, each
container would be kept on line for 12 cycles, or 48
days. About 30 Ib of xenon, 6 1b of krypton, and 0.1 1b
of tritiated water would be accumulated in each bottle.
Each freshly filled bottle would contain about 240 Ci
of ®B5Kr, equivalent to a decay energy of about 0.4 W
per bottle. The bottle pressure after equilibrating to
room temperature would be ~1000 psi. About 230
bottles would be filled during the 30-year life of an
MSBR station. Each filled container would be trans-
ferred to long-term storage, where, after a period of about
100 years, the *H and ®5Kr would decay sufficiently
for the contents to be released or sold without
radiological protection.
7.7 COMPRESSORS

A compressor is used to return the effluent of the gas
cleanup system to the purge-gas cycle. The compressor
has a capacity of about 2 scfm of helium, with an inlet
pressure of 14.7 psia and an outlet pressure of 60 psia.
A major requirement for the compressor is to provide
positive sealing for the pumped fluid so that the highly
purified gas is not recontaminated.

The 47-hr xenon recycle system will be designed to
operate on the available pressure drop, so a compressor
probably will not be required. However, if one is
needed, the flow will be 9 scfm, and the compression
ratio will be fairly low, about 1.4 to 1. Positive sealing
will be essential to prevent outleakage of the highly
radioactive gas. Other requirements will be radiation
resistance and remote maintainability.

7.8 PIPING AND VALVING

Double containment, or better, is provided in all parts
of the system where outleakage could cause a hazard to

109

personnel, In especially critical areas, favorable pressure
gradients are provided, for example, by use¢ of a
high-pressure inert gas blanket in an annulus surround-
ing the radioactive gases, The off-gas system layout
recognized the necessity to minimize the effects of
solids accumulations at valve seats, pipe bends, etc.,
where fission product decay heating would tend to
cause hot spots, and additional study and development
will be required.

All valves are provided with welded bellows for
positive stem sealing. Positive-sealed end connections,
either buffered O-rings or butt welds, are also used.
Where necessary, provisions are made for remote
maintenance of valving.

Gas system piping and components are provided with
a controlled-circulation ambient air system, which
assures prompt detection of gas leaks and the channel-
ing of such leaks to an absolute filter system.
8. Fuel-Salt Processing System

L. E. McNeese

8.1 GENERAL

The principal objectives of fuel processing are the
isolation of 2?3Pa from regions of high neutron flux
during its decay to 23U and the removal of fission
products from the system. It is also necessary to remove
impurities from the reactor fuel salt which may arise
from corrosion or maloperation of the reactor system.

The fuel processing system is an integral part of the
reactor system and will be operated continuously. This
allows processing of the reactor on a short cycle with
acceptably small inventories of salt and fissile materials.
The reactor can continue to operate even if the
processing facility is shut down, however, although at a
gradual decrease in nuclear performance as the poisons
accumulate.

The processing methods are based on reductive
extraction, which involves the selective distribution of
materials between salt and bismuth containing reducing
agents such as thorium and lithium. The isolation of
protactinium by reductive extraction is relatively
straightforward since there are significant differences in
chemical behavior between protactinium and the other
components of the fuel salt (U, Th, Li, and Be), as is
evidenced by the distribution ratios? of these materials
between fuel salt and bismuth containing a reductant.
Extraction of the protactinium into bismuth requires
the prior and complete removal of uranium from the
fuel salt. Two methods (described below) are available
for accomplishing this.

In the older protactinium isolation method,”® the salt
stream from the reactor was fed directly into a bismuth
contactor, and sufficient reductant was fed counter-
current to the fuel salt to not only isolate the
protactinium but to also reduce all of the UF, present
in the fuel salt. The UF; concentration in the fuel salt
is relatively high (0.003 mole fraction), and the
quantity of reductant required (10* gram equivalents
per day) was sufficiently large that its purchase would

110

be uneconomical. For this reason a relatively large
electrolytic cell was used to reduce LiF and ThF,; from
the fuel salt to provide the required reductant.

In the preferred protactinium isolation system, only
recently devised, fluorination is used for removing most
of the uranium from the fuel salt prior to protactinium
isolation. With this system, the quantity of reductant
required is such that it can be purchased economically,
and an electrolytic cell (which presents unusual devel-
opment problems) is not required.

The removal of the rare-earth fission products from
the fuel salt is more difficult because the chemical
behavior of the rare-earth fluorides is similar to that of
thorium fluoride, which is a major component of the
fuel salt. Two rare-earth-removal systems, both based
on reductive extraction, have been considered.

In the older rare-earth-removal system,’® the fuel
carrier salt containing rare-earth fluorides was counter-
currently contacted with bismuth in order to exploit
the small differences in the extent to which thorium
and the rare earths distribute between the fuel carrier
salt and bismuth containing a reductant. Since the
distribution behavior of the rare earths and thorium is
quite similar (j.e., rare-earth—thorium separation factors
near unity),?*? it was necessary to use a large number
of stages in the extraction columns and high metal-to-
salt flow ratios. The system used a large amount of
reductant (about 4.5 X 10* gram equivalents per day)
which was provided by electrolytic reduction of LiF.

The preferred rare-earth-removal method, known as
the metal-transfer process,'’ was also devised only
recently. This process exploits the relatively large
differences in the extent to which rare earths and
thorium distribute between bismuth containing a reduc-
tant and lithium chloride.!' The new process does not
require an electrolytic cell; this is an important advan-
tage over the earlier process.

The remainder of this section describes a system
incorporating the fluorination—reductive-extraction
process for protactinium isolation and the metal-
transfer process for rare-earth removal.

8.2 PROTACTINIUM ISOLATION

The fluorination—reductive-extraction system for iso-
lating protactinium is shown in Fig. 8.1, The salt stream
from the reactor first passes through a fluorinator,
where about 95% of the uranium is removed. The salt
stream leaving the fluorinator is countercurrently con-
tacted with a bismuth stream containing lithium and
thorium in a multistage contactor in order to remove
the uranium and protactinium from the salt., The
bismuth stream leaving the column, which contains the
extracted uranium and protactinium as well as lithium
and thorium, is contacted with an HF-H, mixture in
the presence of a molten-alt stream in order to remove
these materials from the bismuth. The salt stream which
flows through the hydrofluorinator also circulates
through a fluorinator, where about 95% of the uranium
is removed, and through a tank which contains most of
the protactinium. Uranium produced in the tank by

oy Tl

[
|

111

decay of protactinium is removed by the circulating salt
stream, Reductant (lithium) is added to the bismuth
stream leaving the hydrofluorinator, and the resulting
stream is returned to the extraction column. The salt
stream leaving the column is essentially free of uranium
and protactinium and is processed for removal of rare
earths before being returned to the reactor.

Calculations have shown that the system is quite
stable with respect to variations as large as 20% for
most of the important parameters: flow rates, reductant
concentration, and number of extraction stages.'®®
The required uranium-removal efficiency in the initial
fluorinator is less than 95%. The number of stages
required in the extraction column is relatively low, and
the metal-tosalt flow ratio (about 0.14) is in a range
where the effects of axial mixing in packed column
extractors will be negligible.'®':'°? Since the protac-
tinium-removal efficiency is very high and the system is
quite stable, materials such as 2®'Pa, Zr, Ni, and Pu
should accumulate in the protactinium decay tank.

Operating conditions that will yield a ten-day protac-
tinium removal time include a fuel-salt flow rate of 0.88

ORNL—DWG 70-11032A

‘ REDUCTANT
ADDITION

e L_|

1

Y

REACTOR EXTRACTION
UFg Bi UFyg
} [ ¥ }
——— 1 FLUORINATION L HYDROFLUORINATION =g FLUORINATION - Pa DECAY
Fy HF Fp \
Fz Pa DECAY UFg
— AND —
FLUORINATION
SALT
TO WASTE

Fig. 8.1. Protactinium isolation with uranium removal by fluorination.
gpm (ten-day processing cycle), a bismuth flow of 0.11
gpm, and five stages in the extraction column. The
required quantity of reductant is 371 equivalents per
day, which will cost $288 per day, or 0.012 mill/kWhr,
if 7Li is purchased at $120 per kilogram.

8.3 RARE-EARTH REMOVAL

Rare-earth and alkaline-earth fission products can be
removed effectively from the fuel salt by the metal-
transfer process. In this process, bismuth containing
thorium and lithium is used to transport the rare-earth
fission products from the reactor fuel salt to an
acceptor salt. Although LiCl is the preferred acceptor
salt, LiBr or LiCl-LiBr mixtures could also be used.

Both thorium and rare earths transfer to the bismuth;
however, because of favorable distribution coefficients,
only a small fraction of the thorium transfers with the
rare earths from the bismuth to the LiCl. The effective
thorium—rare-earth separation factors for the various
rare earths range from about 10% to about 10%. The
final step of the process is removal of the rare earths
from the LiCl by extraction with bismuth containing
0.05 to 0.50 mole fraction lithium,

The conceptual process flowsheet (Fig. 8.2) includes
four extractors that operate at about 640°C. Fuel salt
from the protactinium isolation system, which is free of
uranium and protactinium but contains the rare earths
at the reactor concentration, is countercurrently con-
tacted with bismuth containing approximately 0.002
mole fraction lithium and 0.0025 mole fraction thor-
ium (90% of thorium solubility) in extractor 1. Frac-
tions of the rare earths transfer to the downflowing
metal stream and are carried into extractor 2. Here, the
bismuth stream is contacted countercurrently with
LiCl, and fractions of the rare earths and a trace of the
thorium transfer to the LiCl. The resulting LiCl stream
is routed to extractor 4, where it is contacted with a
bismuth solution having a lithium concentration of 0.05
mole fraction for removal of trivalent rare earths.
About 2% of the LiCl leaving extractor 4 is routed to
extractor 3, where it is contacted with a bismuth
solution having a lithium concentration of 0.5 mole
fraction for removal of divalent rare earths (samarium
and europium) and the alkaline earths (barium and
strontium). The LiCl from extractors 3 and 4 (still
containing some rare earths) is then returned to
extractor 2.

Calculations were made fo identify the important
system parameters.'' It was found that there is
considerable latitude in choosing operating conditions
which will yield a stated removal time. The number of
stages required in the extractors is low: less than six in

112

ORNL=DWG 70—-4517A

PROCESSED SALT

TO REACTOR ‘—?_‘;——

EXTRACTOR |
3-6 STAGES

FUEL SALT FREE } [
OF U AND Pa Bl

EXTRACTOR 2
3-6 STAGES

A

i=Li
(0.5 mole fraction Li)

EXTRACTOR 3

LiG! 2-3 STAGES
l——. Bi—-Lj
+ DIVALENT
' RARE EARTHS
j—————Bi-Li
r (005 mole fraction Li)
EXTRACTOR 4
\ STAGE
|=————®Bi—Li
L N + TRIVALENT
S RARE EARTHS

Fig, 8,2, Mass transfer process for removal of rare earths from
a single-fluid MSBR.

extractors 1 and 2, three or less in extractor 3, and one
in extractor 4. The process appears to be essentially
insensitive to minor variations in operating conditions
such as flow ratios, reductant concentrations, and
temperature. The required salt and bismuth flow rates
depend on the desired rare-earth-removal times.

8.4 INTEGRATED PLANT FLOWSHEET

The flowsheet that has been adopted for the MSBR is
a combination of the processes described in the two
previous sections. Figure 2.4 shows the integrated
flowsheet. A description and analysis follow.

A small stream of fuel salt taken from the reactor
drain tank flows through a fluorinator, where about
95% of the uranium is removed as gaseous UF,. The
salt then flows to a reductive-extraction column, where
protactinium and the remaining uranium are chemically
reduced and extracted into liquid bismuth flowing
countercurrent to the salt. The reducing agent, lithium
and thorium dissolved in bismuth, is introduced at the
top of the extraction column. The bismuth stream
leaving the column contains the extracted uranium and
protactinium as well as lithium, thorium, and fission
product zirconium. The extracted materials are re-
moved from the bismuth stream by contacting the
stream with an HF-H, mixture in the presence of a
waste salt which is circulated through the hydrofluori-
nator from the protactinium decay tank. The salt
stream leaving the hydrofluorinator, which contains
UF, and PaF,, passes through a fluorinator, where
about 90% of the uranium is removed. The resulting salt
stream then flows through a tank having a volume of
about 130 ft*, where most of the protactinium is held
and where most of the protactinium decay heat is
removed. Uranium produced in the tank by protac-
tinium decay is removed by circulation of the salt
through the fluorinator. Materials that do not form
volatile fluorides during fluorination will also accumu-
late in the decay tank; these include fission product
zirconium and corrosion product nickel. These materi-
als are subsequently removed from the tank by periodic
discard of salt at a rate equivalent to about 0.1 ft?/day.
This salt is withdrawn to a storage tank on a 220-day
cycle (eight **3Pa halfdives) in order to ensure suffi-
ciently complete decay of the protactinium. After this
decay period a batch fluorination of the 22-ft® salt
volume is carried out in the storage vessel for removal
of residual uranium. The salt is then discarded.

The bismuth stream leaving the hydrofluorinator is
then combined with sufficient reductant (lithium) for
operation of the protactinium isolation system. Effec-
tively, this stream is fed to the extraction column of the
protactinium isolation system; actually, it first passes
through a captive bismuth phase in the rare-earth-
removal system in order to purge uranium and protac-
tinium from this captive volume.

The salt stream leaving the protactinium extraction
column contains negligible amounts of uranium and
protactinium but contains the rare earths at essentially
the reactor concentration. This stream is fed to the
rare-earth-removal system, where fractions of the rare
earths are removed from the fuel carrier salt by
countercurrent contact with bismuth containing lithium
and thorium. The bismuth stream is contacted with
LiCl, to which the rare earths, along with a negligible
amount of thorium, are transferred. The rare earths are
then removed from the LiCl by contact with bismuth
containing a high concentration of 7 Li. Separate extrac-
tors are used for removal of the divalent and trivalent
rare earths in order to minimize the quantity of "Li
required. Only about 2% of the LiCl leaving the
trivalent-rare-earth extractor is fed to the extractor in
which the divalent materials are removed.

113

Calculations have been made'®? for a range of
operating conditions in order to evaluate the flowsheet
just described. In making these calculations the MATA-
DOR code was used to determine the reactor breeding
ratio for each set of processing plant operating condi-
tions examined. Data are not available on the cost of
processing for this flowsheet or for the reference
flowsheet for the processing system that uses electro-
lyzers in both the protactinium- and rare-earth-removal
systems. In the absence of these data, processing
conditions were examined which would result in the
same reactor performance (i.e., the same breeding ratio)
as that obtained with the previous reference flowsheet.

Although the optimum operating conditions which
will result in a breeding ratio equal to that of the
reference reactor and processing system (1.063) have
not been determined, the following conditions are
believed to be representative, The reactor was processed
on a ten-day cycle, with the complete fuel-salt stream
(0.88 gpm) passing through both the protactinium
isolation system and the rare-earth-removal system. The
resulting protactinium removal time was ten days, and
reductant requirement was 371 equivalents per day, or
$230 per day, which costs 0.012 mill/kWhr. The
protactinium isolation column is 3 in. in diameter, and
the total number of required stages is about 5. The
protactinium isolation system also results in a ten<ay
removal time for materials that are more noble than
thorium but do not have volatile fluorides, These
include zirconium, 2%'Pa, plutonium, the seminoble
metals, and corrosion products.

The rare-earth-removal system consists of three pri-
mary contactors: (1) a 7.1-in.-diam six-stage column in
which the rare earths are transferred from the fuel salt
to a 12.5gpm bismuth stream, (2) a 13-in.-diam
six-stage column in which the rare earths are transferred
from the bismuth to a 33.4-gpm LiCl stream, and (3) a
12.3-in.-diam column in which the trivalent rare earths
are transferred from the LiCl to an 8.1-gpm bismuth
stream having a lithium concentration of 0.05 mole
fraction. Two percent of the LiCl (0.69 gpm) leaving
the trivalent-rare-earth extractor is contacted with a
bismuth stream (1.5 ¢m®/min) having a lithium concen-
tration of 0.5 mole fraction for removal of the divalent
fission products such as Sm, Eu, Ba, and Sr. The total
lithium consumption rate for the rare-earth system is
119 moles/day, or $81 per day, which costs 0.0042
mill/kWhr.

The rare-earth-removal times range from 15.5 days for
cerium to 50.4 days for europium. The distribution
data for neodymium, which are believed to be conserva-
tive, were used for rare earths for which distribution
data were not available (ie., Y, Pr, and Pm).
The costs for reductant in both the protactinium
isolation system and in the rare-earth-removal system
constitute only a small fraction of the total processing
costs: however, they indicate that one can purchase
reductant rather than use an electrolytic cell for
producing this imaieiial. As duaia become availabie on
processing costs, the optimum conditions will be
determined for the most economic operation of the
processing plant,

8.5 SALT-BISMUTH CONTACTORS

Salt-metal contactors are required at several points in
the flowsheets. Where multistage contactors are needed,
packed columns operated with the salt phase con-
tinuous are the preferred type of contactors. In cases
where only a single stage is required, mixer-settlers
could be used instead.

Studies have been made of pressure drop, flooding,
dispersed phase holdup, and axial mixing for columns
packed with both solid cylindrical and Raschig ring
packing ranging in size from % to % in.! 11199193 Bor
most applications the preferred packing is %-in. Raschig
rings. Sufficient data are available for determining the
required column diameter for stated throughputs of salt
and bismuth, but additional data are needed on the
column height equivalent to a theoretical stage (HETS).
The HETS values for the required contactors are
assumed to be 20 to 24 in. The column diameters range
from 3 to 13 in.

8.6 FLUORINATORS

Uranium is removed from the salt streams as UF4 by
countercurrently contacting the salt with fluorine gas in
a salt-phase-continuous systemn. Because this process
involves quite corrosive conditions, it is carried out in
columns whose walls are protected from corrosion by a
layer of salt frozen on all surfaces that potentially
contact both fluorine and salt.?

The fluorinators are envisioned as open columns, and
axial mixing in the salt phase caused by rising gas
bubbles tends to reduce fluorinator performance. Axial
dispersion data have been obtained during counter-
current flow of air and water in columns having
diameters of 1.5, 2, 3, and 6 in, These data were
combined with previous data on uranium removal in a
l-in-diam continuous fluorinator in order to predict
the performance of fluorinators having larger diameters.
The two continuous fluorinators used in the processing
system, which remove 95% of the uranium from salt
streams having flow rates of about 170 ft3/day, are 6
in. in diameter and 10 ft high.

114

8.7 FUEL RECONSTITUTION

Uranium is removed as UFg at two points in the
process, and it is necessary to return most of this
uranium fo the fuel salt returning to the reactor. This is
accomplished by absorbing the Ul mnio the processed
salt and reducing the resulting mixture with H, to
produce UF,. Although the overall reaction is straight-
forward,

UF, + H, = UF, + 2HF ,

it is believed that intermediate uranium fluorides such
as UFs, which are soluble in the salt and nonvolatile,
are responsible for the rapid absorption reaction which
occurs when UF, is contacted with salt containing
lower valence uranium fluorides. The rate at which UF4
must be reduced to UF, is about 700 moles/day. It is
believed that the reaction can be carried out continu-
ously with the H, and UF4 added either to the same
vessel or to different vessels between which the salt is
circulated. Conditions in the system are likely to be
corrosive, and frozen wall corrosion protection may be
required.

8.8 SALT CLEANUP

Before the processed salt is returned to the reactor,
the concentration of impurities which may be harmful
to the reactor system must be reduced to safe levels. It
will also be necessary to ensure that the U**/U** ratio
in the salt has the proper value so that conditions in the
reactor will be noncorrosive to Hastelloy N,

Since nickel is quite soluble in bismuth and Hastelloy
N is a nickel-base alloy, bismuth is the most important
potential impurity in the salt. Bismuth could be
dissolved or entrained in the salt or could be present as
a soluble bismuth compound. Few data are available
with which to assess the magnitude of the bismuth
problem. The solubility of bismuth in the fuel salt is
believed o be no greater than about 2 ppm and may be
much lower. Entrainment is not considered a serious
problem. Also, the bismuth concentration which can be
tolerated in the reactor is not known. Until additional
data are obtained, however, the problem of bismuth
being present in the salt will be regarded as significant.
The concentration of other impurities such as FeF, and
NiF; must also be reduced to low levels since these
materials will interact with chromium, a constituent of
Hastelloy N.

The presently envisioned salt cleanup system consists
of a 2-in.-diam, 50-ft-long vessel packed with nickel
mesh. Salt flowing through the vessel is contacted with
a countercurrent flow of H, at a rate of about 34 scfm.
The salt then passes through a porous metal filter prior
to its return to the reactor.

8.9 PUMPS

Several small pumps will be required for both molten
salt and bismuth throughout the processing plant. The
capacities for bismuth pumps range from about 0.15 to
12.3 gpm and for salt pumps from about 1 to 33 gpm.

8.10 MATERIALS

The MSBR chemical processes impose severe limita-
tions on containment materials. Compatibility with
liquid bismuth and molten salt fuels at 1200°F (650°C)
is required. Conventional nickel- and iron-base alloys
are not satisfactory because of their susceptibility to
dissolution and mass transfer in bismuth. The most
promising materials appear to be molybdenum, tung-
sten, rhenium, tantalum, and graphite. Of these, molyh-
denum, tungsten, rhenium, and graphite are difficult to
fabricate into complex shapes, and tantalum has a high
reactivity with environments other than ultrahigh vac-
uums. In addition, it is necessary to consider the
possible effects of lithium or thorium in bismuth and a

115

high fluoride ion concentration in the molten salt on
compatibility. With these factors in mind, it was
concluded that molybdenum has the highest probability
for success in this application.

Molybdenum vessels can be fabricated by the back-
extrusion process, which involves the flow of metal into
a die and extrusion back over an advancing plunger. The
advantages of this process are that the final diameter is
as large as or larger than the starting blank, the
geometry can be changed by relatively simple changes
in die and mandrel design, and deformation can be
accomplished below the recrystallization temperature,
so that a wrought structure having good mechanical
properties is produced. By this technique, vessel heads
can be produced with integral bosses for pipe connec-
tions.

Brazing produces joints in molybdenum systems with
good mechanical properties, but commercially available
brazing alloys for molybdenum are not compatible with
both bismuth and fluoride salts. Molybdenum can be
welded by either a gas tungsten-arc process or by an
electron beam technique, Welding has the disadvantage,
however, that the recrystallized region is very brittle,
The most satisfactory joint may be a butt weld backed
up by a brazed sleeve which limits the stress on the
brittle zone.
9. Liquid-Waste Disposal System

Radioactive liquid wastes accumulated from decon- systems have not received any conceptual study, but it
tamination operations and other sources will be col- is anticipated that the design will be straightforward
lected in the chemical processing facility for treatment. and will not pose major development problems. An

The concentrated waste will be stored for decay and allowance was made in the cost estimate for these
eventual disposal. The wasle treatment and storage facilities.

116
10. Plant Operation, Control, and Instrumentation

10.1 GENERAL

Operation of the MSBR power station embraces all
phases of startup from either cold or standby con-
ditions, reliable delivery of electric power at any
demanded load between about 20 and 100% of capac-
ity, and procedures for both scheduled and unplanned
shutdowns. An overriding consideration at all times is
safe operation of the plant to protect the public from
possible radioactive hazards and to prevent injury to
operating personnel and major damage to equipment.

Thre controls system must recognize the different
requirements for the various operating modes and
establish safe and appropriate operating conditions. The
systems must coordinate the reactor, the primary- and
secondary-salt loops, the steam generators and re-
heaters, the turbine-generator, and the several as-
sociated auxiliary systems. In general, the load demand
is the primary signal to which the controls subsystems
are subordinate, unless overridden by safety considera-
tions. The controls should minimize temperature fluctu-
ations at critical points, such as at the turbine throttle,
should limit rates of temperature changes to keep
stresses in materials within the acceptable ranges, and
should guard against freezing of the fuel and heat-
transport salts in the systems.

It may be noted that the steam conditions to be
maintained at the turbine throttle cannot be realized by
simply controlling the power produced in the reactor,
since the transport lag, or time delay, between a change
in reactor power and a corresponding change in the heat
transferred to the steam is about 10 sec under most
conditions. A faster adjustment can be made by
controlling the coolant-salt flow to the steam generator.
Salt flow regulation can be accomplished either by
valves in the salt lines or by varying the speed of the
coolant-salt circulating pumps. Since the pump rotation
can be varied with sufficient speed of response to
accommodate anticipated load changes, this is the
control method selected for the MSBR reference design.
Although valves for salt service have received relatively

117

little development to date, it is to be noted that flow
control valves for salt service are relatively simple in
concept as compared with mechanical-type shutoff
valves,* and the problems in developing the flow
control device are not necessarily great. Fluidic valves
were briefly studied at ORNL and appear to have
considerable promise for proportioning flows in
molten-salt systems.

To establish the general feasibility of the MSBR
concept, estimates were made of material stresses under
transient conditions to determine whether the allowable
rates of load change would be acceptable. Analog
simulations were carried out to indicate whether the
systems were stable and whether the basic control
conditions and requirements could reasonably be met.
Standby, startup, and shutdown modes were explored
sufficiently to suggest a flowsheet, to outline the special
equipment needed, and to generally evaluate this aspect
of plant operation.

10.2 MSBR REACTIVITY CONTROL

John L. Anderson  S. J. Ditto

Two types of rods are planned for the MSBR core:
(1) control rods, which have both regulating and
shimming functions for normal load following and
shutdown, and (2) safety rods, which are primarily for
backup to assure adequate negative reactivity for
emergency situations.

The control rods are movable graphite cylinders about
3% in. in diameter with axial passages through them for
a cooling flow of fuel salt, as shown in Fig. 3.1.
Withdrawal of the graphite leaves an undermoderated
region at the center of the reactor and causes a
reduction in reactivity. It may be noted that the
graphite has considerable buoyancy in the fuel salt;

*Positive shutoff is achieved in the MSBR drain line by a
freeze-plug arrangement, a concept proven to be satisfactory in
the MSRE.
118

thus, if a rod should break, the graphite pieces would
float out of the core and reduce the reactivity. The
total worth of each rod, as calculated by Smith,'" in
moving the full core height from fully inserted to the
fully withdrawn position is about 0.08% 6k/k. Based on
a higher anticipated werth than this, two contiol 1uds
and two safety rods were originally planned, as shown
in Fig. 3.1. On the basis of later estimates, however, it
now appears that a total of four control rods and two
safety rods may be required to achieve satisfactory
control.

MSBR reactor control simulations reported by
Sides' °* indicate that a reactivity rate of change of
about 0.01%/sec 6k/k is adequate for normal control of
the reactor. This would require linear velocities for four
rods acting together of 04 fps. It is reasonable to
expect that this velocity could be attained with a
relatively simple rod-drive system using electric motors.

Circumstances could arise which would require a
faster rate of reactivity decrease than the 0.01%/sec
mentioned above, such as sudden large load reductions
or loss of load. Such transients may require negative
reactivity rates as high as 0.05 to 0.1%/sec 8k/k. One
method of attaining the fast rate of control rod
withdrawal would be by an air turbine and an electric
motor coupled to the control rod drive through
differential gearing. The electric motor would be used
to increase the reactivity at a relatively slow rate, and
the air turbine would be capable of fast withdrawal.
The inherent unidirectional characteristics of the
turbine would make it impossible for it to run
backward to insert reactivity at a fast rate. More study
will be required to arrive at definitive designs, but the
control rod drives appear to be within established
technology.

Long-term reactivity adjustments will be ac-
complished in the MSBR by varying the fuel concentra-
tion. Initial fuel loading will be done by gradually
increasing the concentration in circulating barren salt.
Subsequent refills of the reactor system may be with
already enriched salt from the drain tank. The normal
fuel-addition rates will.be slow and manageable, so that
very modest control of reactivity rates can oversee the
process. The possibilities for misoperation of the
fuel-addition process have not been assessed at this
stage of the MSBR design study, but a reasonable
allowance in shutdown control reactivity will be made
for this eventuality.

Temperature changes in the primary salt will affect
the reactivity. The mean temperature of the salt could
possibly increase about 150°F from startup to full-load
conditions. With a nominal temperature coefficient of

reactivity of —5 X 107%/°F,!' a net reactivity change
of about 0.075% 8k/k must be accommodated. Tem-
perature changes will normally be made slowly in order
to minimize thermal stresses in the system, but there is
the possibility that on stopping and restarting of a
fuel sult pump a cooler stug of saii {rom the heart
exchanger could be carried into the reactor core to
produce a relatively rapid increase in reactivity. The
amount of reactivity involved, however, is not likely to
be great because of the improbability that all the
primary pumps would be stopped and then restarted
simultaneously.

In normal MSBR operation there is a reactivity loss
due to delayed neutron precursors being carried out of
the core by the circulating fuel salt. At the present time
it is planned to operate the MSBR with 2 constant
circulation rate for the fuel salt, but if the flow rate
were decreased or stopped, this effect would cause an
increase in positive reactivity, It is estimated that total
flow stoppage would result in a reactivity change of
about +0.2% 8k/k.* °3

Since the amount of gas entrained in the fuel salt
affects the reactivity, changes in the salt circulation
rate, the system pressure, salt chemistry, and perform-
ance of the stripping gas injection and removal systems
could cause relatively rapid insertion or removal of
reactivity. Maximum rates are related to the velocity of
the fuel salt in the core. Extrapolation of MSRE
experience to the MSBR indicates that the maximum
total reactivity effect due to gas entrainment will be less
than 0.2% &k/k. A change in gas entrainment from the
expected normal level of 1% to a level of 2% is
calculated to produce a reactivity change of about
—0.04% Sk/k." !

The amount of reactivity needed to override xenon
reactivity transients associated with changes in reactor
power is quite small in the MSBR compared with other
reactor types in that a large fraction of the xenon is
continuously removed by the gas purging and stripping
system. The total equilibrium xenon effect from low
power to full power is estimated to be about 0.3%
8k/k." Transient effects can, of course, vary widely,
depending upon the amount and duration of the power
changes.

In summary, although the sum of the reactivity
effects discussed above is about 0.85% 6k/k, all the
effects will not have maximum importance occurring
simultaneously, and some will be of opposite sign. A
total of 0.3% &k/k provided by the graphite control
rods is expected to be adequate to cover short-term
reactivity effects in the MSBR. As previously men-
tioned, long-term effects will be compensated by fuel
concentration changes.
10.3 REACTIVITY CONTROL FOR
EMERGENCY SHUTDOWN

John L. Anderson  S. J. Ditto

Over and above the normal reactivity control needs
discussed above, additional shutdown capacity is neces-
sary to take care of unforeseen situations or emergency
conditions, such as major changes in salt composition or
temperature effects when filling the primary system,
flow stoppages in the circulating loops, gross tempera-
ture changes, malfunctions in the control rod system,
etc.

Safety rods consisting of boron carbide clad in
Hastelloy N can be used at the center of the core to
furnish an independent shutdown capability. Each of
these absorber rods would have an estimated worth of
about —1.5% Ak/k, and two to four rods would
probably be sufficient.!®5 The presence of neutron-
absorbing material in the core is undesirable during
normal operation; therefore the rods would be for
safety purposes only and would normally be fully
withdrawn. Since there would be times, however, when
it might be preferable to operate for short periods with
the absorber rods partially inserted, they should have
full adjustment capability in addition to a fast-insertion
action.

10.4 PLANT PROTECTIVE SYSTEM
John L. Anderson S, J. Ditto

104.1 General

The plant protective system includes those com-
ponents and interconnection devices, from sensors
through final actvating mechanisms, which have the
function of limiting the consequences of specified
accidents or equipment malfunctions, The minimum
requirement of the plant protective system is protection
of the general public. In addition, the protective system
should limit the hazard to operating personnel and
provide protection against major plant damage.

This section briefly outlines specific protective
actions considered necessary for the MSBR, together
with some of the requirements for their initiation. The
plant protective system would function by three
primary mechanisms: reactivity reduction, load reduc-
tion, and fuel-salt drain.

104.2 Reactivity Reduction

The protective system must be capable of coping with
reactivity disturbances beyond the capability of the

119

normal control system. As discussed in Sect. 10.3, such
postulated conditions include malfunction of the con-
trols system, accidental large additions of reactivity,
sudden loss of plant load, gross loss of core cooling
capability, etc.

The safety rods provided for the MSBR must have a
time response, reliability, and a total worth adequate
for the worst-case accident. A dynamic system analysis
will be necessary to establish the performance required.
The necessary reliability is a function of the estimated
frequency of need and the consequences of failure to
perform as planned.

10.4.3 Load Reduction

The relatively high melting temperatures of both the
fuel and coolant salts make freezing of the salt in the
heat exchangers a concern, since loss or reduction of
salt flow in any loop can lead to overcooling if
appropriate steps are not taken. In addition, failure to
maintain a proper balance between reactor power and
heat removed by the steam system can lead to system
cooldown. The MSRE, however, demonstrated that
prevention of freezing of salts in a molten-salt reactor is
not a particularly difficult controls problem.

Loss of temperature control through failure of the
controls system or by other accidents must be pro-
tected against. The need for protective action will be
sensed by measuring appropriate temperatures, flow
rates, and power balances. The action taken will be
dependent upon the type of condition existing and will
probably involve stopping circulation in various salt
loops as well as shedding parts of the load. A particular
problem exists when an emergency shutdown of the
reactor occurs. Immediate reduction of the load to the
afterheat level is required so that the salt systems can be
held at acceptable temperature levels.

10.4.4 Fuel Drain

While draining of the fuel salt into the drain tank is an
ultimate shutdown mechanism for the MSBR system, it
is anticipated that sudden drains would be required
only if the integrity of the primary system were lost. In
general, the best place for the fuel salt is within the
primary circulation system, but if through pipe rupture
or other failure circulation within the system cannot be
maintained, the drain mechanism will be used. While
the drain system must be very reliable, it is not
mandatory that it be capable of being initiated rapidly
in the “dumping’ sense.
10.5 AVAILABILITY OF INSTRUMENTATION
AND CONTROLS FOR THE REFERENCE
DESIGN MSBR

R. L. Moore

As was reported by Tailackson,'®® the MSRE pro-
vided valuable design and operating experience with
molten-salt reactor nuclear and process instrumenta-
tion. One of the important differences between the
MSRE and the MSBR concept, however, is that the
high-temperature cells planned for the MSBR could
subject some of the instruments to ambient tempera-
tures as high as 1000°F unless they are provided with
special cooling.

Nuclear detectors are not now available which could
operate at temperatures in excess of 1000°F. [nasmuch
as Ruble and Hanauer’®” were of the opinion that
there was a practical upper limit of about 900°F for
electrical insulating materials for ionization chambers
and counters, development work in this area, and in the
location of the detectors, will be needed for an MSBR.
In this connection, neutron fluctuation analyses may
prove to be a valuable tool for monitoring and
predicting anomalous behavior, 108,109

Process instrumentation located inside the MSBR cells
will tend to require development because of the high
ambient temperatures, as mentioned above. Thermo-
couple temperature measurements in the MSRE were
generally satisfactory, although more work was needed
on measurement of small differences at the higher
temperatures. Ceramic-insulated platinum resistance
thermometers and ultrasonic methods of temperature
measurement could have application in the MSBR.

Direct and differential pressure measurements in the
MSBR can probably best be accomplished by NaK-filled
pressure transmitters, In addition to the venturi-type
flowmeters used in the MSRE, turbine and magnetic-
type flowmeters can be considered for the MSBR, The
gas bubblers and the conductivity-type probes'®®
used for liquid level indication in the MSRE worked
adequately, but supplementation by float-type in-
strumentation would be desirable. The pneumatic
weighing system used to determine MSRE tank in-
ventories would require adaptation to the higher tem-
peratures in an MSBR. The containment penetration
seals, gas-system control valves, electrical disconnects,
and wiring and insulation associated with all the
above-mentioned devices will also require study and
development.

Effort is needed in many areas to arrive at detailed
designs and specifications for MSBR control system
components,” ! but it may be noted that work being

120

accomplished for other reactor types will probably have
application in the MSBR.' 2

The aspects of the MSBR instrumentation and con-
trols systems requiring significant development have
been discussed in detail in ORNL-TM-3303.' '3

10.6 ALLOWABLE RATES OF LOAD
CHANGES

R. B. Briggs

To design the controls system for an MSBR station it
15 necessary to know rates of change which should not
normally be exceeded when varying the plant load. A
major consideration is the rate that the temperatures of
the fuel and coolant salts can be allowed to change. The
factor most likely to govern is the thermal stresses
generated in the Hastelloy N in contact with the salts.
Changing the temperature of the salt will cause the
metal surface temperature to change more rapidly than
the interior, resulting in a greater temperature gradient
and increased stresses. The magnitude of the stress will
depend upon the thickness of the metal, the sait-film
heat transfer coefficients, the rate of change of tem-
perature, and, for many situations, the total range of
temperature change.

The results of a simple study’'* to provide prelim-
inary information are given in Table 10.1. In this study,
computer calculations were made of stresses induced in
Hastelloy N plates 2 to 4 in. thick, with various heat
transfer coefficients and with varying rates of change of
salt temperature. The latter were selected to represent
the conditions providing maximum stress that would
occur due to load changes of 10, 20, and 40% of full
load, with the reactor inlet temperature held constant
at 1050°F and with full design flow of fuel salt across
one surface of the plate but with no heat flow through
the other surface. The temperature distribution through
the plate was calculated for various times after initiating
changes in the salt temperature, and the corresponding
stresses were determined. The calculated maximum
stresses were compared with an allowable stress value of

Table 10.1. Effect of metal plate thickness
on allowable rate of change of MSBR plant load

Allowable rate of change (%/min) for

Plate
thickness total change in load of —
(in.) 10% 20% 40% 100%
2 >40 40 ~6 4
4 >40 ~2 <1 <1

121

18,000 psi, which is based on the assumption that the
MSBR will be designed for combined stresses and will
experience no more than about 10,000 cycles of 20% or
more in power over the plant life. On this basis the
effect of plate thickness on the allowable rate of load
change is as shown in Table 10.1. These values are
believed to be conservative in that the thicker plates
will probably be cooled to some extent from both
surfaces. In addition, the walls of the reactor vessel are
cooled by the inlet flow of salt, so that the heavy
sections do not have to change through the full range of
temperature when the power changes through the full
range.

Since the estimated allowable rates of load change,
even when based on these somewhat pessimistic as-
sumptions, are much the same as those presently used
in thermal power stations, it can be concluded that
operation of the reference design MSBR is not uniquely
restricted in this sense.

10.7 CONTROL OF FULL AND PARTIAL
LOAD OPERATION

W. H. Sides, Jr.

The power operating range for the 1000-MW(e)
MSBR station is from 20 to 100% of full design load.
Throughout this load swing the steam temperature to
the turbine throttle must be held essentially constant,
the primary- and secondary-salt temperatures and flow
rates must be kept within acceptable limits, and the
resulting stresses due to induced thermal gradients must
remain within the acceptable ranges. Also, the system
temperature and flow profile at 20% load must be
compatible with the conditions existing in the plant in
the upper portion of the startup range.

A master load programmer would probably be used to
divide the required load demand among the four
primary-coolant loops and among the steam generators
and reheaters associated with each primary-coolant
loop. It should be possible to operate the plant at, say,
75% of full load by operating three of four primary
loops (and their associated secondary-salt loops) at
100% capacity each. Although perhaps not mandatory,
it seems reasonable that all parallel loops should operate
under essentially identical conditions, sharing the exist-
ing load equally. This is, in part, because all parallel
loops always have identical salt conditions at their
inlets.

A scheme for dividing the load should be capable of
making load allotments to the various loops on the basis
of total power demand and number of operable loops.
It should also be capable of recognizing a power

demand exceeding the capability of operable loops and
correcting such conditions, by shedding load, in a way
that does not jeopardize plant operation at its current
maximum capacity. Presuming all operating loops
operate under similar conditions, closed loop control
can perform normally for the appropriate percentage of
design point power as described for full system opera-
tion.

Plant load control may be accomplished by the use of
two basic control loops: a steam temperature controller
and a reactor outlet temperature controller, as indicated
in Fig. 10.1, The steam temperature may be controlled
by varying the secondary-salt flow rate in the steam
gencrator., For example, if the mass flow rate of the
steam is decreased, the outlet steam temperature tends
to increase. A steam temperature error is generated by
comparing the measured value with its set point of
1000°F. The error reduces the secondary-salt flow rate
and thus the heat input to the steam generator. This
control loop continues to adjust the salt flow rate
appropriately to maintain the steam temperature at
1000°F. Results of analog simulations®® have shown
that accurate steam temperature control may be ac-
complished in this way. A change in plant load from
100 to 50% at a rate of 5%/min produced a maximum
simulated steam temperature error of about 2°F. The
maximum required rate of change in secondary-salt
flow to accomplish this was about 9%/min.

The temperatures and flow rates in the salt system
required to produce 1000°F, 3600-psia stearn at part
loads, using the reactor outlet temperature controller
considered here, were determined by specifying the
reactor outlet temperature as a function of load and the
primary-salt flow rate as constant. The remaining
temperatures and the secondary-salt flow rate were
calculated from heat balance considerations through the
plant.

The reactor outlet temperature controller is similar to
that used successfully on the MSRE.'°® Specifically, a
load demand signal determines the reactor outlet
temperature set point. The measured reactor inlet
temperature is subtracted from the reactor outlet
temperature set point, and since the primary-salt flow
rate is constant, a reactor power set point is generated
by multiplying this AT by a proportionality constant.
The measured value of reactor power (from neutron
flux) is compared with the reactor power set point, and
any error is fed to the control rod servo for appropriate
reactivity adjustment. The reactor power set point,
generated from the outlet temperature set point and the
measured reactor inlet temperature, is a function of the
reactor inlet temperature during a transient and thus a
122

ORNL-DOWG 71—247

Poemano

T

stge 1000°F)

STEAM
GENERATOR

ROD

DRIVE
SERVO

&

/FD PRIMARY
HEAT
EXCHANGER
REACTOR
nl -~

T2

Fig. 10.1. Simulation model of plant and control system.

function of dynamic load. Analog simulations of the
plant employing abbreviated models for the reactor
core, primary heat exchanger, and steam generator''®
indicate that plant load control can be accomplished in
this way. The control system also is capable of
canceling small reactivity perturbations.

The small isothermal temperature coefficient of re-
activity in the reactor core implies that only modest
amounts of control reactivity are needed to accomplish
plant load maneuvering. For a normal load change of
100 to 50% at a rate of 5%/min, the maximum amount
of reactivity required was 0.06% 6k/k, reduced at a rate
of —0.0053% 8k/k per minute.* ! 3

10.8 CONTROL FOR FAST SHUTDOWN
W. H. Sides

A fast-acting load and power reduction system may
be required to enable the plant to remain in operation if
failures occur in the heat transfer system. Such a system
could avoid total shutdown of the plant and also
facilitate resumption of normal operation when condi-
tions permit.

Upon loss of primary- or secondary-salt flow in a loop
due to the failure of a primary or secondary pump or
due to some failure of piping or components which
necessitates reduction of flow, care must be taken to
prevent undesirably low temperatures of the salts. For
example, if the flow of secondary salt in a loop is
stopped or greatly reduced, the transit time of the salt
through the four steam generators associated with that

loop increases, and the secondary-salt cold leg tempera-
ture decreases. To prevent freezing of the secondary salt
in the shell of the steam generator near the feedwater
inlet, the flow of steam through the tubes must be
decreased. A reduction in load by about 25% must take
place upon the loss of flow in a secondary-salt loop at a
rate sufficient to prevent excessively low coolant-salt
temperatures. The fuel-salt temperature in the primary
heat exchanger tends to increase upon loss of second-
ary-salt flow and thus does not approach the freezing
point.

If there is a loss of fuel-salt flow, the temperature of
the salt in the primary heat exchanger decreases to
undesirably low values. The freezing point of the
primary salt is approximately 930°F, and the tempera-
ture of the secondary salt entering the primary heat
exchanger at design point is 850°F. Analog simula-
tions' ' ¥ have shown that due to transit time of the
secondary salt in the piping from the steam generators
to the primary heat exchanger, a reduction in steam
flow in the stcam generators does not reflect rapidly
enough in the primary exchanger to prevent low
temperature of the fuel salt in the tubes. Loss of
primary flow in a loop must therefore be followed by a
reduction in secondary-salt flow, and, as discussed
above, a major reduction in secondary-salt flow requires
a redoction in steam flow through the four steam
generators associated with the particular loop.

In summary, loss of primary or secondary flow in a
loop requires that in the loop affected the reactor
system must be decoupled from the steam system to
prevent low temperatures from occurring in the salts. If
secondary flow is reduced, the associated steam flow
must be reduced, but the associated primary flow need
not be reduced. If primary flow is reduced, both the
associated secondary-salt and steam flows must be
reduced to prevent low salt temperatures. In any of
these situations the reactor power must be quickly
lowered in proportion to the net reduction in steam
load. Similarly, upon large or total loss of load, it may
be necessary to assist the control system by providing
fast power reduction and perhaps fast reduction of
secondary-salt flow rate to keep system temperatures
within acceptable bounds.

10.9 STARTUP, STANDBY, AND SHUTDOWN
PROCEDURES

E.C. Hise

10.9.1 General

This preliminary study of the startup, standby, and
shutdown procedures was carried only to the point of
indicating feasibility. Although they have not had the
benefit of close study or opfimization, the arrange-
ments do not appear more complicated or restrictive
than the systems now in use in large supercritical-
pressure steam stations. The procedures would lend
themselves to computerized program control, as is
presently the trend.

The freezing temperatures of the primary and sec-
ondary salts are such that the salt systems must be filled
and circulating isothermally at 1000°F before power
withdrawal can be initiated by decreasing the coolant-
salt temperature. To avoid freezing of the salt and to
prevent excessive temperature gradients, the minimum
feedwater temperature to the steam generators must
vary between 1000°F at zero load and 700°F in the 8
to 100% power range. In addition, the afterheat load in
the reactor system, which decays essentially as in-
dicated in Fig. 6.4, requires that the feedwater and heat
rejection systems remain in operation following shut-
down of the main steam system. Most of the special
systems and equipment needed to handle the startup
and shutdown conditions in an MSBR station are
therefore associated with the steam-power system. The
requirements impose some departure from the equiva-
lent systems used in conventional fossil-fired super-
critical-pressure steam plants and will require further
study.

The proposed general arrangement of the MSBR
steam system was described in Sect. 5, and the overall
steam system flowsheet was shown in Fig. 5.1. For
convenience, pertinent aspects of that flowsheet are

123

included in the startup, standby, and shutdown flow-
sheet, Fig. 10.2. (The letters used in the following
discussion refer to Fig. 10.2.)

Briefly, steam at 3500 psia and 1000°F is supplied by
16 steam generators SG. Superheat control is partially
by varying the coolant-salt circulation rate and by
vaporizing a small amount of 700°F water into the
outlet steam at the attemperator 4. Feedwater at
700°F is normally supplied by mixing steam with the
SS0°F feedwater leaving the top extraction heater TEH
in a mixing chamber M. The steam used for this
feedwater heating is the 867°F exit heating steam from
the reheat steam preheater RSP. The heated feedwater
is raised to about 3800 psia inlet steam generator
pressure by boiler feedwater pressure-booster pumps
PBP. The 552°F exhaust of the high-pressure turbine
HPT is first preheated to about 650°F in a heat
exchanger RSP supplied with 3600-psia, 1000°F steam
from the steam generator outlet, The reheat steam then
enters the reheaters RH, in which coolant salt is
circulated to raise the steam temperature to 1000°F.
Reheat temperature control is by varying the coolant-
salt flow rate. The feedwater system contains steam-
driven feedwater pumps BFP, conventional feedwater
heaters, condensers, full-flow demineralizers, de-
aerators, etc.

The equipment necessary for startup, hot standby,
and heat rejection is also included in the steam system.,
Briefly, this consists of an auxiliary startup boiler AB,
either oil or gas fired, which can deliver supercritical-
pressure steam at 1000°F, an associated auxiliary boiler
feed pump A-BFP, a desuperheater DSH, a steam dryer
SD, and various throttling and letdown valves, as will be
discussed below., A standby-power steam turbine-
generator S-7G of about 10 MW(e) capacity, as dis-
cussed in Sect. 11.], may also be considered in
conjunction with the startup and standby systems,

It may be noted in the flowsheet, Fig. 10.2, that the
boiler feed pump drive turbine BFP-T is supplied both
with extracted steam from the high-pressure turbine
and from the dryer SD in the standby system in order
to assure continued operation of the feed pumps when
the flow of steam to the main turbines is interrupted
for any reason. Steam for the dryer is obtained by
taking off a small portion of the steam generator outlet
steam at the boiler throttle valve BTV, reducing its
pressure to 1100 psia (860°F) through the boiler
extraction valve BE, and reheating it to about 950°F in
the steam dryer SD by means of heat exchange with
some of the 3600 psia, 1000°F prime steam. Steam
from the dryer also plays an important part in startup,
restart, and shutdown operations, as will be explained
below.
124

ORNL-OWGE 70-673T

840 P 1000 F
3600 P 00O F l
HOT fi::::::a::as | DEMINERALIZER
sy g
ANT SALT Le
coo ‘go AL BOOSTER
—_— N PUMP
~ - -
RH |MEMEATER @ A <l ®
COOLANT E
SALT g
850 ¢ 3 CONDENSER
(1.5 in Hg ABS) .j
wap FEEOWATER
WELL EXTRACTION STEAM HEATERS
3600 F BSO F FROM HPT
W &
PUMP
REHEAT STEAM PREHEATERS 1050 @
.. 950 F
Upoas B PUMP
BTV X TURBINE
aF P
e ¥ BOILER 2 |so
ATTEMPERATOR g OO P K0 Gt o Ay DEAERATOR
4 ARQF ~—= STANDBY
2 r"‘\ - POWER TURBINE (57
RiiER X DSH STEAM GENERATOR
RELIEF DESUPER-  DRYER <
VALVE HEATER -
COOLANT SALT {150 f 1WO0 P 656 F o
EXTRALTION STEAM DUHING HEAT REJECTION
?smm-up AND STANDBY THRATTLE VALVE
C i '} EXTRACTION STEAM
) (IB00P 3580 P ¥ 2507 Jven
700 F 14
ot M [ HIGH PRESSURE FEEDWATER MEATERS BFP
CDOLANT SALT 850 F PBP MIXER 1P FEEDWATER 1P BUOSTER
S 8F BOOSTER PUMP AB HEATERS PUMP
STEAM GENERATOR
AUX BOILER A~BFP
AUX BFFP
(200000 (/e ) MOTOR: DRIVE

Fig. 10,2, MSBR steam plant startup and shutdown system.

10.9.2 Startup Procedures

There are two startup procedures to be considered:
(1) cold startup, with all systems cold and empty, and
(2) hot restart from a hot standby condition. As in any
thermal power station, the ability to hold the system in
hot standby and to achieve quick starts from this
condition is desirable to avoid excessive outage times
for the plant.

10.9.2.1 Cold Start. A normal startup from the
cold-empty condition proceeds as follows: The primary
and secondary cell electric heaters are turned on, and
the primary and secondary circulation pumps are
started to circulate helium in the salt systems. When the
temperature of the secondary system reaches 850°F,
the loop is filled with coolant salt from the heated drain
tank, and sall circulation is started. When the primary
system reaches 1000°F, it is filled from the fuel-salt
drain tank, and salt circulation is commenced. Both salt
systems will continue to be circulated isothermally at
1000°F until power escalation is started. The primary-
and secondary-salt flow rates are at the levels required
for the zero-power mode.

The reactor is made critical at essentially zero power
using the methods discussed above. This operation
requires removal of safety rods and further addition of
reactivity by insertion of graphite control rods under

the surveillance of startup instrumentation and a flux
level control system. When the power reaches an
appropriate level, which is still below the sensible power
generating range, the automatic neutron flux level
controller is used to control the power.

Concurrently with the salt systems being electrically
heated, the steam system is warmed and brought to
operating conditions by means of an oil- or gas-fired
auxiliary boiler. Deaeration and demineralization of the
feedwater and warmup of piping, feedwater heaters,
turbines, etc., proceed in a conventional manner with
steam taken from this auxiliary boiler. To avoid
excessive thermal gradients in a steam generator, it must
be at nearly full operating conditions of 3600 psia and
1000°F before steam is admitted. As the auxiliary
boiler is being raised to this pressure, steam from it is
throttled through the boiler extraction valve BE and
through the desuperheater DSH, and is used for
feedwater heating, for warming and rolling the boiler
feed pump drive turbines BFP-T, and for warming the
high-pressure feedwater heaters. When the auxiliary
boiler reaches full pressure and temperature, circulation
can be started through the steam generator.

When the steam system is ready to take on load, the
set point of the flux controller is adjusted as required to
maintain the desired salt temperatures as the feedwater
flow is increased. The feedwater temperature to the
steam generator is reduced by tempering the feed steam
with 550°F water in the mixing chamber M. As the
steamn load is slowly increased the reactor power is
matched to the load, and salt temperatures are kept at
the desired level by manipulating the flux set point. (In
the 2 to 10% power range, temperature changes are
slow, and control should not be difficult.) When the
load reaches 800,000 1b/hr, or about 8 ta 10% of full
load, the reactor can be put in a temperature control
mode instead of a flux control mode after matching the
temperature set point with the existing outlet tempera-
ture. The load is held essentially constant until the
system comes to equilibrium, at which point the reactor
outlet temperature set point is adjusted to meet the
requirements for subsequent load-following control,
The boiler feedwater pressure-booster pumps PBP are
then started to raise the steam generator inlet pressure
to about 3800 psia, and the auxiliary boiler and its
feedwater pump can be taken off the line. The system is
now self-suppoiting at about 8% load.

At this point in the startup procedure, part of the
steam generator output is going to the mixer M via the
reheat steam preheater, and the remaining steam is
going through the boiler extraction valve BE to drive
the main boiler feed pumps, etc. The main turbines,
which have previously been warmed, can now be
gradually brought up to speed and temperature, first
using steam from the hot standby equipment and then
switching to steam taken directly from the steam
generators.

The load is next increased to about 20%, at which
time the steam temperature controller is activated. At
this power level the “normal” control system regulates
the reactor outlet temperature as a function of load,
and the steam temperature controller holds the stcam
temperature at 1000°F. To prevent undesirable tran-
sients as the control system is first activated, the various
system parameters and set poinis are adjusted to the
requirements of the existing power demand prior to
switching to fully automatic control,

More exact definition of the conditions at which the
various steps of the startup program are initiated, as
well as allowable rates of change of the variables, was
beyond the scope of the present study.

10.9.2.2 Hot Standby and Startup. On reduction of
the main turbine load and closure of the stop valve SV,
steam will be immediately let down through the boiler
extraction valve BE, through the desuperheater and
heat rejection valve HRTV, and then to the main
turbine condenser. Except for extreme situations of
sudden loss of turbine load, and possibly not then, the

125

boiler pressure-relief valves need not vent steam to the
atmosphere.

A portion of the steam from the steam generator can
be used to drive the boiler feed pump turbine BFP-T
and to continue circulation of feedwater to the steam
generators for heat removal and rejection to the turbine
condensers. Another portion of the steam will continue
to drive the standby steam turbine-generator o supply
standby power (if not available from the electric power
grid through the station service transformer) to drive
the salt circulation pumps, some of the main con-
densing water supply pumps, and the hot well, pressure
booster, and other pumps required to maintain the
feedwater system operative.

Afterheat from the reactor system will continue to be
transferred to the steam system and maintain it at
operating temperature for several hours, depending
upon the burden of fission products in the system. As
this heat source decays, the auxiliary boiler can be
started if it is desired to maintain the system in the hot
standby condition. The time required for restart from
this mode would be limited by the acceptable rate of
temperature rise in the main turbines, as in con-
ventional steam systems.

10.9.3 Normal Shutdown

The normal shutdown procedure is for the system
power to be reduced under control of the operating
circuits (until about 8% of full-load power is reached)
by gradually reducing the flow to the main turbines to
zero and al the same time transferring the generated
steam to the hot standby system through the boiler
extraction valve BE and thence to the turbine con-
denser. If it is desired to stay in the hot standby
condition the auxiliary boiler can be started; if not, the
main turbine can be allowed to cool, the rate being
controlled by admitting some steam from the steam
dryer SD through the turbine seals and warmup system.
Feedwater will continue to be supplied to as many of
the steam generators as required (probably one or two)
to remove reactor afterheat and to maintain the desired
salt temperature profiles. After about ten days of
afterheat removal (depending on the operating history
of the reactor) the fuel salt will be transferred to the
drain tank. The cell electric heaters will maintain the
cell temperature high enough for the coolant salt to
remain in the molten condition. With termination of all
steam generation the steam system can be allowed to
cool.
P,

11.1 AUXILIARY ELECTRIC POWER
E. S, Bettis

Even though the MSBR is designed on the basis that
the safety of the public will not be endangered even if
there were a complete loss of electric power, it is highly
desirable that a small amount of power be available to
operate the controls system and certain other com-
ponents to prevent possible damage to equipment in
particular emergency situations.

The MSBR will probably use an auxiliary power
source for instruments and controls the same as that
employed successfully at the MSRE. This was a system
of storage batteries kept charged by an ac-dc motor-
generator (M-G) set. Without the M-G set operative the
batteries can deliver 100 kW of power at 250 V for at
least an hour. In addition to freedom from interruption
of the power supply, use of the batteries also eliminates
concern for any possible transients in voltage, etc., that
could be induced if there were other connected
equipment. A static dc-ac inverter changes the power
from the batteries into the ac required for the instru-
ments and controls circuits.

In addition to the relatively small amount of auxiliary
power needed for instruments and controls, standby
power is also required for the salt circulation pumps,
freeze-valve coolant pumps, cell cooling systems, etc. A
delay of several minutes can be tolerated in restoring
these items to service, however. The total connected
load for this type of equipment cannot be precisely
estimated at this time, but even with ample allowances
for uncertainties, it should not exceed about 10 MW(e).

Several possible methods were considered for pro-
ducing the standby power. It was decided to use
auxiliary steam turbine-generators, although diesel-
driven generators and gas turbines were also likely
candidates. The steam turbines seem a logical choice
because an ample source of steam is always available,
either from the afterheat-removal system or from the
auxiliary startup boiler. As shown in the flowsheet, Fig.

126

Auxiliary Systems

10.2, the auxiliary steam turbines take their steam from
the steam dryer in the startup system. These units must
be kept at operating temperature at all times in any
case, since it is part of the heat-rejection system for
nuclear afterheat and would be required in event of a
main turbine trip and loss of plant load. The supply of
steam from the afterheat disposal system is sufficient to
drive the auxiliary turbines for several hours. Should
the MSBR be isolated from the power grid for a longer
period, the auxiliary startup boiler can be fired to
supply the necessary steam.

11.2 CELL ELECTRIC HEATING SYSTEMS
E. S. Bettis

All the cells containing fuel or coolant salts (except
the chemical processing cell) operate at ambient tem-
peratures of 1000 to 1100°F. Heat losses from the
equipment are sufficient to maintain most of the cells
at this temperature during normal operation of the
MSBR. During initial warmups, downtime, or possibly
at very low reactor power levels, electric space heaters
are used to heat the cells. The cells can be likened to
low-temperature electrically heated furnaces, with
thermal insulation in the walls to reduce heat losses.
The biological shielding is cooled to prevent the
concrete temperature from exceeding 150°F. The
heater element design is essentially the same as that
used successfully in the MSRE for over five years.

The heater units consist of two lengths of ¥%-in.-diam
X 0.035-in.-wall-thickness Inconel tubing about 20 ft
long with the two ends welded together at the bottom
to form a hairpin shape, as shown in Fig. 11.1. Each
unit is contained within a thimble of a similar hairpin
shape made from 2-in.-OD stainless steel tubing with
Lavite bushings spaced at 3-ft intervals to center the
heater within the thimble. The heaters are designed for
120-V, three-phase power from a solid-state-controlled
supply which limits the thimble surface to about
1200°F.
127

ORNL-DWG 70-11968

Bfi-0i

Ift-6in

20f1-0in

ol i

Fig. 11.1. Electric cell heating unit cluster.

Ll

The heater element electrical leads are copper rods
brazed to the top ends and extending about 10 ft
through the top shielding structure of the cells. The exit
cooling gas from the cell liner space passes through the
heater lead penetrations to cool the copper rods. Three
heaters are connected in series to reduce the number of
connector leads and penetrations required. A removable
flanged cover encloses each group of three heaters to

collect the exit cooling gas and return it to the
circulating system, The electrical leads pass through
gas-tight electrically insulated penetrations in these
cover boxes.

The heater thimbles are welded to the inner liner of
the cell and thus become part of the containment
system. With this arrangement the heater elements can
be withdrawn without disturbing the integrity of the
containment. A total of 592 thimbles are arranged
around the periphery of the reactor cell in such a way
as to avoid too close proximity to cell equipment.
There are eight symmetrical groupings of 74 heaters
each. Heaters in the drain tank and steam-generating
cells are similarly arranged. Some of the heaters in the
cells will be used as spares, thus making it possible to
postpone a shutdown of the reactor in event a heater
repair becomes necessary.

The cell heating loads and heater data are given in
Table 11.1.

Table 11.1. Cell heating loads and electric heater data

Reactor Steam Drain tank
cell cell? cell
Heat Joss at 1100° cell 413 195 122
temperature, kW
Cell contents heatup load, kWhr 86,000 5000  ~10,000
Heatup power, kW 413 195 122
Approximate heatup time, days 9 1 6
Heater length, ft 40 40 40
Kilowatts per heater 2.66 2.66 2,66
Number heaters required 312 147 93
Number installed 592 147 186

@Each of four.

11.3 RADIOACTIVE MATERIAL DISPOSAL
SYSTEM

E. S. Bettis

Although it is recognized that storage and disposal of
radioactive materials is subject to many regulations and
would affect siting considerations, the reference MSBR
design assumes that it will be possible to retain within
the shielded containment all radioactive debris accumu-
lated over the design lifetime of the plant. This waste
material would include solid fission products from the
chemical processing plant, spent cores taken from the
reactor vessel, failed pieces of equipment which could
not be salvaged, and other radioactive materials.
A waste pit provides the necessary storage space. The
pit is a circular cell about 72 ft in diameter and 30 ft
deep located directly beneath the reactor cell (see Fig.
13.3). Calculations of the heat production in the waste
materials give equilibrium values of between 100 and
800 kW, The cell is cooled Ly a circulaling gas,
probably nitrogen, which passes through the cell and
over a water-cooled coil, The circulating fans and the
heat exchangers are located in a shielded and sealed cell
immediately adjacent to the waste storage cell. The heat
exchanger has stop valves in the water system in event
of a break or leak in the tubes. It is estimated that even
if all the water in the coil were to leak into the cell and
be vaporized, there would be an insignificant rise in the
cell pressure., Redundancy could be provided in the
cooling system if required.

128

It may be practical to containerize the fission
products from the chemical processing system before
depositing them in the storage cell. Residue resulting
from decontamination of the crane bay and other areas
will also be packaged before being stored in the waste
cell.

No specific plans have been made for removal of
wastes from the storage pit after an MSBR station has
been permanently shut down for obsolescence or other
reasons. It may be permissible to pour concrete into the
waste pit to encapsulate the material. The MSBR design
could obviously be modified to accommodate shipment
of radioactive wastes to disposal sites, should this be
required.
12. Maintenance and Repair Systems

E. C. Hise

12.1 GENERAL

It is evident that a practical method of remote
maintenance and a method for replacing the core
graphite are essential for the success of the MSBR
conceptual design presented in this report. Since the
size and radioactivity level of some of the items of
MSBR equipment are greater than the present range of
maintenance experience, many of the procedures re-
main to be developed. To reach a reasonably valid
judgment as to the feasibility of the maintenance
arrangements, it is necessary to visualize each of the
major steps required.

The plan for maintenance of the MSBR follows the
technology developed for previous fluid-fuel reactors.
All the radioactive MSBR equipment is installed in
containment cells having the overhead shielding ar-
ranged in removable sections to permit access from the
top. The systems will be designed so that each piece of
equipment, its supports, electrical instrumentation,
process piping connections, etc., may be viewed from
above and be accessible when using remotely operated
tools. The usual procedure would be to remove and
replace a failed component rather than to make repairs
in place, since the latter would usually result in a longer
plant downtime, The defective unit would be trans-
ported in a shielded carrier to a hot cell within the
reactor complex for examination and be either repaired
or discarded to the waste storage cell.

Some of the MSBR items requiring maintenance will
be comparable in size and type with the equipment
used in the MSRE, for which there is a valuable
background of practical maintenance experience. The
design of the special tools and MSRE maintenance
procedures were described by Blumberg' ' 7 and in
MSR progress reports.?5+ A feasible method for
remotely cutting and welding radioactive piping is being
developed by Holz.''®

Since most of the cell areas cannot be reentered once
the reactor has generated neutrons, maintenance proce-
dures must be carefully planned, with much of the
special equipment and fixtures installed and tested as
the plant is constructed. The maintenance system must
therefore be an integral part of the plant design.

The investment required for the equipment needed
for major maintenance operations has been included as
a capital cost for an MSBR station. Relatively small and
routine maintenance operations are considered as a
plant operating cost. The expense of the materials and
special labor required for periodic replacement of the
core graphite is treated as a separate account (see Sect.
15 and Table D.15).

The MSBR maintenance requirements fit into four
general classes:

Class I — permanent equipment. This category con-
tains all those items which should last the design
lifetime of the plant and will normally require no
maintenance. Examples are the reactor vessel, the pump
vessels, primary heat exchanger shells, the fuel-salt drain
tank, thermal shielding, thermal insulation, connecting
process piping, etc. Although essentially no provisions
are included with the installation for maintenance of
these items, it would be possible to replace them using
specially prepared facilities and at the expense of a long
plant outage. (All of this equipment, however, does
have built-in provisions for in-service inspection.)

Class II — equipment allowing direct maintenance.
This group includes the items which probably can be
approached for direct maintenance once the coolant
salt has been drained and flushed and a decay period of
several days has elapsed. The steam generators, re-
heaters, coolant-salt pumps, and the equipment in the
heat-rejection cell fall into this class. In the unlikely
event that a component did become radioactive, its
removal would be treated as a class IIl or IV item,
discussed below. Once the source of activity was

129
removed from the cell, cleanup and component replace-
ment could proceed in the normal fashion using direct
maintenance.

Class IIl — equipment requiring semidirect mainte-
nance. Much of the equipment in the off-gas and

. .
m " cinee  anlla asenle
chemical processing cells, such

-~ —

biowers,
valves, processing vessels, filters, etc., will become
radioactive. In general, these items are of relatively
small size and are comparable with MSRE equipment
size. The in-cell maintenance methods for this class of
equipment will, however, require appropriate changes in
the shielding, etc., to accommodate MSBR radiation
levels, which may be a factor of 10 or more higher than
experienced in the MSRE.

Class IV — large equipment requiring remote mainte-
nance. This group includes items which are clearly
beyond present experience because of a combination of
size, radiation level, afterheat removal, and disposal
considerations. Examples are the pump rotary element,
the primary heat exchanger bundle, etc. The principal
maintenance operation falling into this classification is
replacement of the reactor core moderator assembly,
Since this operation must be repeated several times
during the lifetime of the plant, the procedures can be
planned in considerable detail.

12.2 SEMIDIRECT MAINTENANCE PROCEDURES

To perform maintenance on class Il items, and those
in class II if the activity level requires it, the roof
section, or plug, immediately above the component is
removed and set aside. A work shield similar to that
shown in Fig. 12.1 is then placed over the opening. The
work shield would have viewing ports and lights,
openings for insertion of periscopes, extension tools,
and other maintenance equipment. Movement of the
slides and eccentrics in the shield can place any of the
openings in the shield over the desired point. The
mechanical operations of disconnecting and reconnect-
ing components are done with extension tools inserted
through the work shield. A failed component is drawn
through the work shield into a shielded carrier for
transport to a hot cell for repair or disposal.

12.3 REMOTE MAINTENANCE PROCEDURES

Replacement of the reactor core assembly is one of
the more difficult maintenance operations both because
of the size of the equipment and the intensity of the
radioactivity encountered. Special maintenance equip-
ment will be required, the major item being a 20-ft-
diam, 40-ft-high shielded transport cask for the reactor
core assembly. As shown in Fig. 12.2, the cask is an

130

integral part of a polar crane which can be rotated tu
cover all points in the reactor building. The cask moves
laterally (but not vertically) and has a 240-ton-capacity
remotely controlled hoisting mechanism on top to draw
the core assembly up into the cask. The carbon steel
walls of the cask are aboui 2 in. thick, which is
sufficient to reduce the radiation level on contact with
the outside of the cask to about 1000 R/hr after a
ten-day decay period for the core. (The activity level on
contact with the outside wall of the reactor building
would be less than 100 mR/hr.) After this decay time
the estimated heat generation in the core assembly is
about 0,25 MW, as shown in Fig. 3.25. Conservative
estimates indicate that this amount of heat can be
safely dissipated through the cask wall and that no
special cooling system for the cask will be required. The
cask is provided with an adjustable sealing ring and shield
at the bottom to provide a tight connection with the cell
closure transition pieces described below. The cask can
be closed at the bottom with a two-leaf gate valve, or
shutter,

As shown in Fig. 12.2, a domed maintenance contain-
ment vessel is permanently installed over the top of the
reactor cell. It is relatively thin walled and is designed
primarily to contain airborne contaminants during
maintenance operations. It is provided with access ports
over the fuel-salt pumps and heat exchangers and has a
central 24-ft-diam cover which can be removed to
provide access to the shield plugs covering the reactor
vessel. This top opening in the maintenance vessel has
an inner extension in the form of a cylinder with a
four-leaf gate valve at the bottom, termed the reactor
vessel maintenance closure in Fig, 12.2, which extends
to the top elevation of the roof plugs. The cylinder
serves as a transition piece between the reactor vessel
and the transport cask to provide positive containment
during the core hoisting operation. It is equipped with a
high-capacity exhaust fan to assure an inward move-
ment of air through the opening. The gate valve
prevents convective circulation of gases from the
reactor cell while the reactor vessel is open.

A reactor work shield will also be required. It has the
same dimensions as the roof plug covering the reactor
vessel and is installed in its place to provide viewing
ports and tool access for engaging the moderator lifting
rods and other semiremote maintenance operations.

Transition pieces are also provided for temporarily
connecting the transport cask to the spent equipment
cells and to the new core replacement cell to prevent
escape of particulates into the high-bay area.

A 150-ton conventional hoist, shown in Fig. 12.2,
also travels on the polar crane to handle work shields.
131

PHOTO 63913A
N

Tt e

: b . !

o

b
~

FILLER MODULE

Fig. 12.1. Portable work shield for MSBR reactor cell.
132

j——22 ft3in 1N

ORNL—-DWG 69—12692A

~—HOIST (4 TOTAL)

W HIGH BAY CONTAINMENT

44 ft

(134 f+ DIAM)

r CASK
AUXILIARY SUPPORT
LR CARRIAGE .. — TRANSPORT CASK
R
_—POLAR CRANE R
£ i
| _—— TRANSPORT CASK CLOSURE
CONTROL ROD~- -— -
. h\\\ ' /
. _——ADJUSTABLE SHIELD ol
£ _—— REACTOR VESSEL \
N MAINTENANCE o
W2 CLOSURE } s
19§+ 6 in A — MAINTENANCE |
15 H3in CONTAINMENT
.I-z?jl_ ‘ = s | —'] ‘ - ]
- e A -( T
= VENT SYSTEM-|* W d I
—-. ;
-1 2
g fennd ' ]
' ~REACTOR b
VESSEL LID v

~— MODERATOR ASSEMBLY

Fig. 12.2. MSBR reactor core assembly transport cask and maintenance system.

transition pieces, etc. This hoist, as well as those on the
transport cask, and other equipment such as the polar
crane, the reactor vessel maintenance closure, surveil-
lance television, etc., can be controlled from the
maintenance control room. This room is a protected
area with shielded windows overlooking the high bay, as
indicated in Fig. 13.4.

The functions of the equipment can best be explained
by the following brief description of the steps used in
replacing a core moderator assembly,

During the reactor cooldown period, transition pieces
are set up over the new core replacement cell and over
the spent core storage cell (see Fig. 13.5). At the end of

about ten days the central cover in the maintenance
containment vessel is set aside. The high-volume ex-
haust system from the maintenance vessel assures a
controlled movement of air in the working zone.
Through direct and semiremote means the control rod
drive mechanism is disconnected at the elevation of the
top of the shield plug, and the mechanism is drawn up
into a cask, sealed, and stored in the high-bay area
awaiting reinstallation. The control rod tube opening
into the reactor vessel is closed with a blind flange. The
holddown bolts for the reactor vessel top head are
removed, and the shield plug is prepared for lift. The
auxiliary hoist is engaged with the shield plug, and the

hoist is initiated to assure that it is clear. At this
juncture the maintenance crew vacates the high-bay
area,

From the maintenance control room the reactor
vessel shield plug is lifted and set aside. After the
reactor work shield is installed in its place, the
maintenance crew can return to the high bay. Using
semiremote methods through the work shield, the
moderator assembly lifting rod ports are opened and
the lifting rods are engaged (see Fig. 3.7). The 150-ton
auxiliary hoist is engaged with the work shield and
prepared for lift, The high-bay area is again vacated.

By operating the hoist from the control room, the
work shield is removed and set aside. The transport cask
is positioned over the reactor vessel, and the adjustable
shield is closed to provide a good seal with the
maintenance vessel. The four cask hoists are engaged to
the eight lifting rods, and the core assembly is carefully
hoisted into the transport cask. The valves at the
bottom of the cask and at the top of the cell are then
closed, the adjustable shield at the bottom of the cell is
released, and the cask is moved into position over the
spent storage cell.

The cask is engaged with the transition piece over the
spent core storage cell, the lower valve in the cask is
opened, and the moderator assembly is lowered into the
cell. The assembly is supported by the top head flange
in the same manner as it was installed in the reactor
cell. The cask valve is closed, and the cask is moved to
one side to permit the auxiliary hoist to place a shield
plug over the spent core storage cell and to place the
work shield over the reactor vessel.

After the transport cask has been decontaminated,
the reactor vessel work shield is reinstalled, and the
high-bay area is again made safe for occupancy, the
maintenance crew can return to inspect the reactor
vessel. Optical and ultrasonic equipment is operated
through the work shield to inspect vessel welds, etc.,
and to assure that the vessel is ready for installation of a
new moderator assembly.

After again clearing the high bay of personnel, the
auxiliary hoist is used to set aside the work shield. The
reactor vessel maintenance valve is closed to maintain
containment as the shield is lifted. The new moderator
assembly, previously made ready and standing by in the
new core replacement cell, is then hoisted into the
transport cask and moved into position above the
reactor vessel. After sealing the cask to the maintenance
closure, the maintenance valve is opened, and the new
core is carefully lowered into place inside the reactor
vessel. About a 2-in. radial clearance has been provided
for the assembly, and it is not necessary to observe any

133

rotational orientation of the moderator with respect to
the vessel. The auxiliary hoist is then used to replace
the work shield.

Personnel can then return to the high bay to perform
the semiremote operations of disengaging the lifting
rods and resealing the lifting rod ports in the top head
of the vessel. Operating from the maintenance control
room again, the work shield is removed, and the
permanent reactor shield plug is installed. Personnel can
then seal the vessel closure by direct approach and also
install the control rod drives. The system can then be
leak tested and prepared for operation.

12.4 GRAPHITE DISPOSAL AND ALTERNATE
REACTOR VESSEL HEAD RECLAMATION

An MSBR reactor core assembly is estimated to have
a useful full-power life of about four years. During this
operating period the spent core assembly would be
disposed of, and the alternate reactor vessel head, with
its attached reflector graphite and upper cylinder
extension, would be prepared for reuse, The spent core
storage cell would also be cleared to receive the next
core assembly, and a new core would be prepared in the
core replacement cell.

The spent core storage cell is equipped with viewing
windows, manipulators, and tooling for dismantling the
assembly. The graphite moderator sticks are removed
and broken into short lengths and deposited in the
waste storage cell beneath the reactor, as mentioned in
Sect. 13.6. The Hastelloy N support plate for the
graphite will also be cut into smaller pieces and stored
in the waste cell.

After an extended decay time¢ the top head and its
attached graphite reflector, which will be reused along
with the head, are decontaminated as much as possible
by wiping and vacuuming,

A new shop-assembled core is brought into the
reactor building through the air lock shown in Fig, 13.5
and is set into the new core replacement cell. The
alternate top head for the reactor vessel is then brought
from the spent core storage cell by means of the
transport cask. Using semiremote maintenance proce-
dures through a work shield, the lifting rods are
installed and the reactor vessel closure seal rings on the
head are replaced. The assembly is now ready for
installation when needed.

The spent core storage cell is then decontaminated as
much as possible and cleared for the next maintenance
operation.
12.5 DECONTAMINATION

On the basis of past experience with the MSRE, few
decontamination problems are likely to arise. The
contamination can be almost entirely restricted to the
reactor equipment cells. The tools are bagged on
withdrawai from the cell and, along with the transport
casks, are sent to decontamination. MSRE experience
has been that parliculale contaminalion is readily
removed by scrubbing with high-pressure water jets
alone or with the aid of detergents. Occasionally an
inhibited acid may be required.

134

The large transport cask will become contaminated
after it is used to move the reactor core assembly to the
storage cell. It must be decontaminated to a lower
radiation level before the maintenance crew can enter
the high-bay area. It is cleaned in place on the polar
crane by mounting g catch pan benmeath ii, and
high-pressure pumps are used to circulate a decontami-
nating fluid through nozzles which can be manipulated
to clean all portions of the interior.
13. Buildings and Containment

E. S. Bettis

13.1 GENERAL

Plan and elevation layout drawings for the station are
shown in Figs. 13.1 and 13.2. The principal structures
are the cylindrical reactor building, the steam-generator
bay, the steam piping and feedwater heater bay, and the
turbine-generator bay. The reactor and steam-generator
facilities are located on one reinforced concrete pad and
the remaining structures on another. With this arrange-
ment relative displacements due to seismic disturbances
would not threaten the integrity of the containment,
since no piping or connections containing radioactive
materials would cross the boundary between the pads,

The plant site is briefly discussed in Sect. 14.

H. L. Watts

H. M. Poly

13.2 REACTOR BUILDING

One of the primary functions of the cylindrical
reactor building is to provide containment and biologi-
cal shielding during the maintenance operation of
removing and replacing the reactor core assembly.
During normal operation the reactor cell is the primary
containment.

The cylindrical portion of the reactor building is
shown in the elevation drawing in Fig. 13.3. Plan views
at the three major levels are shown in Fig. 13.4 (crane
bay), Fig. 13.5 (upper level), and Fig. 13.6 (lower
level), The building is 189 ft high and 134 ft in
diameter. Excavation for the reactor building will be to

DANL-DWG 69-104924

P e e WA T O s — e
| ——— ] L —— PO 112 # Om - O L T E—
’ - AT — :]
3
SHOPS
OFFICES l— ~l J-El ]
—t e
FEEOWATER
HEATER BAY
AM
sTEam | 28711
On
- _— - -
STEAM , =
V o
_'\‘\‘ - . —
A
STEAM Bs
= =1t =
: L,
1T UNLOAD-
STEAM x ING BAY
TURBINE
ROOM

Fig. 13.1. Overall plan view of MSBR power station.

135
ORNL DWG B9-109884

EL 18311 Oin
EL 154 ft Oin.
N S S N R I P Z .
= ),
W EL 124 ft Omn
| =" e N ) R Y Y A O
! fi L 1 R R
e = g A

™ Tl —— )

]

=
3 i1 |
; 3 ~ o

oy EL 38 ff Oin. @
KEsg s a1 S L. -~ > - -~ ‘b_ \ )
' . - d— ——
| Rk
' [ b e

EL Off Oin} | i [
R ‘.o'.— LA 5 N e ey l‘ — z, pa Ra2d) D SRt A T A 9T Iy APUULGIR NS %

Fig. 13.2. Sectional elevation of MSBR power station.

the depth required for firm support of the monolithic
concrete pad upon which it rests. Finished grade level
thus depends upon particular site conditions and would
preferably be with about two-thirds of the building
showing above ground. The grade level shown in Fig.
13.2 corresponds to the AEC typical site condition
having the top of the limestone formation about 8 ft
below grade.

The reactor cell is located on the first level of the
reactor building, as shown in Figs. 13.3 and 13.5. The
cell is about 72 ft in inside diameter and 30 ft deep
with about 8 ft of concrete biological shielding on the
sides and top, the latter consisting of two layers of
removable roof plugs which permit access for in-
stallation and maintenance of equipment. The double
containment and other construction features of the
reactor cell are described in more detail in Sect. 13.3.

The first level of the reactor building also contains
cells for processing the fuel salt and for off-gas
handling, instrumentation, and storage of spent reactor
cores and heat exchangers. The lower level has a large
shielded and sealed storage cell for permanent storage
of spent graphite, discarded equipment, and other
radioactive waste from the plant, as shown in Fig. 13.6.
A means for depositing radioactive material into the
storage cell is indicated in Fig. 13.3. The volume of the
cell is based on a reasonable assumption of the amount

of material that would be accumulated over the 30-year
life of the MSBR station.

The lower level also provides cells for the primary
drain tank, miscellaneous auxiliary equipment and work
areas, and hot cells equipped with remote manipulators
for examination and repair of radioactive equipment,
Space is also included for the lower section of the
60-ft-deep off-gas and chemical processing cells. All the
other cells are approximately 30 ft deep and have
biological shielding with controlled atmospheres where
required.

The building is constructed of a 3-ft thickness of
ordinary concrete covering a % -in.-thick carbon steel
shell, or liner. The liner acts as a sealing membrane to
permit the building to meet specifications of less than
0.1% leakage per 24 hr. All piping and penetrations are
sealed, and an air lock is provided in the upper level for
moving in new reactor core assemblies (see Fig. 13.5).
During routine operation the building is maintained at
slightly below atmospheric pressure by a controlled
ventilation system discharging through filters and up
the stack. This is an extra measure of protection in
addition to that provided by the primary system and
the double containment of the reactor cell. Operating
personnel would have access to the building at all times
except during certain phases of the maintenance oper-
ations, such as when the spent reactor core is being
137

ORNL-DOWG 69-10489A

AURILIARY )
CRANE ~__ |}

HIGH BAY
CONTAINMENT

~TRANSPORT CASK

CRANE BAY
LEVEL
— -

CRANE BAY

SPENT HEAT

EXCHANGER CELL—-} |- | )

UPPER LEVEL _ |/ 1 CEL

o
REACTOR CELL ut: ol

2=
/

STEAM CELL

FEEDWATER
HEATER BAY

LOWER LEVEL
pr———

WASTE STORAGE CELL

T 4-.—4\'”‘5'

COOLANT SALT
ORAIN TANK CELL

AUXILIARY
1 EQUIPMENT CELL

Fig. 13.3. Sectional elevation through reactor plant building.

drawn up into the transport cask. During these periods
the remotely controlled equipment can be viewed
through shielded windows in the building wall at the
crane bay level, as indicated in Fig. 13.4. (Maintenance
procedures are described in Sect. 12.)

In addition to providing missile protection, the
building serves as sealed containment during mainte-
nance operations and as biological shielding. The 3-ft
thickness of concrete covering the entire structure,
together with the shielding of the transport cask, results
in a reading of less than 100 mR/hr on outside contact
with the reactor building wall as the core assembly is
being removed. Although the building wall thickness
was not optimized, values below 3 ft would require a
corresponding increase in the cask shielding used during
maintenance and increase the weight to near the
maximum load desired for the polar crane.

The concrete shell provides tornado protection, the
building having been designed on the basis of a
300-mph wind with a storm-caused 3-psi negative
pressure differential. It is also designed to withstand
missiles weighing 2500 lb, 15 in. in diameter, and
traveling at 150 mph. The assumed seismic design
conditions are more stringent than those specified for
the reference site (see Sect. 14), having been taken as '
g horizontal and Y, g vertical.

A polar crane is used to service the equipment within
the cylindrical building. The bridge spans the building
and can be rotated to cover essentially all areas. Two
cranes are mounted on the bridge; one is a conventional
hoist of 150 tons capacity, and the other is unique in
that the 20-ft-diam 40-ft-high transport cask is an
integral part of the crane, as indicated in Fig. 12.2. The
cask is fixed as to vertical elevation but can move
138

ORNL-DWG 69-10486A

ACCEEE TC
STEAM CELLS

STEAM CELL BAY

{4 TOTAL) i
/ LL— a = | :
 pe————k == ot .;—*I—" -
; T E:
% L :-“
i b ACCESS TO 3
J REAcEgz Tr»&wg&#mce % ; INSTRUMENTATION CELL ;
\I‘.{ \ \ .
= £ \ 2§ VIEWING WiNDOWS |
& 4 \ 3 i
v SHOPS ‘
, /’ ' \ ¥ 17
(| . i CONTROL ROOM e
i ) /5 0
ACCESS TO i /. ‘;4
1V CONTROL ROD i e 5
iy STORAGE CELL # - — ACCESS TO NEW &
” 1" CORE REPLACEMENT |
. / 5 CELL ,
1 ) w“
2 ACCESS TO i
7|k FREEZE VALVE CELL |
71 : LTy )
\ = l‘:’., .
O M T
" // \\ NG ~
—F AT | \ Y
£y | STACKAREA | /
o | / //
& g7 ~ \ L o
’~ ~ N - ¢
2
A __.// 4
— ~. //
- 4
~
ACCESS TO ot o
SPENT CORE

STORAGE CELL

ACCESS TO WORK AREA

OF F-GAS CELL

ACCESS TO
SPENT HEAT EXCHANGER
STORAGE CELL

Fig. 13.4, Plan view of reactor plant at crane bay elevation.

laterally from above the reactor cell to positions over
the spent core storage cell and the core replacement
pickup point. The hoisting mechanism for lifting
equipment into the transport cask and the other
maintenance procedures are described in Sect. 12.

13.3 REACTOR CELL

The reactor cell provides primary containment for the
reactor, the four primary heat exchangers, the four
fuel-salt circulation pumps, and the interconnecting salt

piping. In addition to leak-tightness meeting the speci-
fications for a containment system, the cell walls
provide a minimum thickness of 8 ft of concrete for
biological shielding. Missile protection is provided by
the domed concrete structure of the reactor building, as
mentioned above. Protection against seismic disturb-
ances is afforded by the monolithic concrete pad upon
which the reactor building rests, as previously discussed,
and by the methods used to mount the equipment, to
be described subsequently.
139

ORNL-DWG 89-10491A

2301t Oin
~ 4811 Oin = f=—30ft Oin—| |=—30ft Oin ~=| |=— 30ft Oin.—=f |=*—30ftOCin —e| f=——481f Oin ——mf
D e = - P = ¥ e Y e T~ r e e 'fiy- T o AR T 5
il 8
f d
& STEAM CELLS ff( ' .
4611 Oin NO. 1 Nno.2 | NO.3 ! NO. 4 t
: 0
S % .‘,Q, x
g b
. i - = | Jil
. & REACTOR CORE
FCONTAINMENT /) AssemelY ARea
(134t QinID) —\ REACTOR CELL WA <
. \id \ X\~ INSTRUMENTA- ,
192 ft Oin X\  TION CELL }
y§ O \ |
87fioin | W » ; o\
/ NEW CORE __ | ;
AUXlLIARYF ogoummsm ] REPLACEMENT CELL —f~ 7 /-NR LOCK
' )i \ T e
| cremicaL PROCESSING [T cuemicar N Aficetis 76 —
| PROCESSING N\ FREEZE VALVE CELL
W : CELL o 1oy BN
gl WORK AREA T
| CONTROL ROO S S——
STORAGE 1" L3~ — —
CELL~" = 5 . .i
Tm - ! i
r (!
\ T : o | OFF-GAS CELL
\ \ "-, VT e V_Q £z
)\ CHEMICAL PROCESSING '\ '\ ! A &8 DRAIN TANK AND /.
HEAT REMOVAL N ' - ¥ " OF F-GAS HEAT RE-
SYSTEM AREA B % i MOVAL SYSTEM /i
2 :'-, ™ | - ' 2 = AREA '
~7 > T letR < /351 Qin.
L ~ —— SPENT CORE ¥
: TANK CELL \ STORAGE CELL
SPENT HEAT EXCHANGER
STORAGE CELL

Fig. 13.5. Plan view of reactor plant at reactor cell elevation,

The atmosphere of the reactor cell (probably nitrogen
with 3 to 5% oxygen) will normally be operated at
about 13 psia and between 1000 and 1100°F. Under
assumed design basis accident situations the cell pres-
sure could rise above atmospheric, however, and the cell
has been designed for S0 psia. During normal operation
the cell atmosphere will become contaminated by
neutron activation and by tritium. In postulated ac-
cidents involving loss of fuel salt from the circulating
system, the cell atmosphere would, of course, become
heavily contaminated. In meeting the shielding, pres-
sure-retention, and leak-tightness requirements, the cell
wall construction must provide both thermal insulation
and gamma shielding to protect the concrete structures

from excessive temperatures. The maximum allowable
temperature for the concrete was taken as | 50°F.

The reactor cell is about 72 ft [D X 30 ft deep and is
located within the reactor building, as shown in Figs.
13.3 and 13.4. The arrangement of equipment in the
cell is as indicated in Figs. 13.7 and 13.8.

The cell wall consists of two concentric carbon steel
shells, both 2 in. thick and separated by a 6-in.-wide
annular space, as indicated in Fig. 13.10 and listed in
Table 13.1. The same type of double wall construction
is also provided in the roof plugs and in the floor
structure. The total thickness of 4 in. of steel supplies
the necessary gamma shielding and the strength to
withstand the 50-psig design pressure. Some of the
140

ORNL-DWG 69-10490A

RS e D T =-_f,;--;_.w}z‘»m..:;.m@‘i’"-.(:—_V“W:d'.\', N ) S B S N o S —

\ 3 I

_{l Y &

) q 11— COOLANT SALT DRAIN TANKS |
S (Y / i ND
2 i | | AUXILIARY EQUIPMENT CELLS i
; ! i [

h i
S TGN
e O

Riider ey,

R T ot
CHERPRINEE

oGyt s )
Lema [

T

CHEMICAL
PROCESSING CELLS

e T e
TELRD W

e

WASTE STORAGE CELL

——————
Yo TR T

WORK AREA —

HOT CELLSH

AUXILIARY
EQUIPMENT CELL

..:',
74 ‘
.'__" :{‘]
5
compoue»irs STORAGE
ASSEMBLY AREA
AUXILIARY ’i
EQUIPMENT CELL i
. L— PRIMARY SALT I
g 1 DRAIN TANK CELL
} ¥

AUXILIARY
EQUIPMENT CELL

AUXILIARY EQUIPMENT CELL

Fig. 13.6. Plan view of reactor plant at waste storage cell elevation.

pressure loading of the inner shell is transmitted to the
outer wall by spacers. A minimum of 8 ft of concrete is
provided on the outside for biological shielding.

An inert gas, probably nitrogen, will be circulated
through the space between the inner and outer shells to
remove the heat due to gamma absorptions and the heat
conducted from the cell interior. The circulating gas
will normally operate at a pressure higher than the
ambient cell pressure to assure that any leakage would
be inward. Heat is removed from the circulating gas
stream by water-cooled coils sealed within a compart-

ment that is an extension of the outer wall of the cell.
Both this gas and the cell atmosphere are provided with
cleanup and disposal systems.

The inner and outer shells will probably operate at
sufficiently different temperatures to require accommo-
dation of relative movement. The outer vessel is
therefore an integral part of the concrete structure,
while the inner one is hung from the top of the cell but
with much of the weight carried by helical coil springs
at the bottom, as shown in Fig. 13.9. The differential
expansion of the shells is also accommodated at the
ORNL-DWG 7011969

/ HEAT EXCHANGER oo RS e
4 ".'.._ - J pe
/ A '
.'. v _ -‘. P

//
/. /lf '. .,
NPy
/ @ O S
v
e - PRIMARY
.y SALT PUMP L
. L3
o' REACTOR 0 - ot o - PI R A I B 5 0N S0 AT e n oL
m - '.. -tk '_;_.vq’ ‘_': " '.'OT-',"-
\ . s A e R T R e =y :
\ T
e / STEAM GENERATOR oy
¥ y

Fig. 13.7. Plan view of reactor and steam cells.
CRNL-DWG 70-11870

CONTAINMENT ——. P

CONTROL x .
ROD DRIVE PRIMARY .3 SECONDARY
SALT PUMP y SALT PUMP

~ — -
| . " e v N . B o 4
. i - . . AZEERO A Ml ST G A S S ) pR R E TR NS L S, v Rl
o L Wl Uy B¢ A B, » . S - SN SO .
- . . . 3 . .
- . ¥ - B - v .
Jq * ‘ 3 . . - 4 A\l > . .
‘ . oy O ‘e . B T T R <O 0 A O s N PR X R AN AN N T RTINS IAL
v . v 4 " 2 A Y Ry £ . .
- . » “B v y
v . K .
. . O 4 . AL . M . . i -

~STEAM GENERATOR
|

" e o t. e ":E
&g : STEAM
gy 4, g 3 \ L™\
NI { 4| ~ PIPING
» § | t \ -
4 r & i gn . H 4
I, U p— — ——— e o - e . ey + . .
by - = b P - r- - - 3 Jd B L"".T
£ [B.A aE: = =
PR TR ‘ 11 3 S
= L - o REACTOR 1| & &
o 1 -
< o %t IR g
. r 1 ‘. .
23 3| | wwew)
: 4 1 2
\' :
LJ .‘ ;. o',
A ¢

HEAT EXCHANGER -

Fig. 13.8. Cross-sectional elevation of reactor and steam cells.
143

Table 13.1, Summary of cell wall construction features

Cell wall constmcu'onb

lé in, CS; 36 in, concrete
Y6 in. SS, 9 in. TI, 2 in. CS, 6 in. AS, 1 in, CS, min 8 ft concrete shielding

for inhabited areas

Cell Heaters®
Reactor confinement building None
Reactor containment cell A
Fuel-salt drain tank cell A
Freeze-valve cell B
Spent equipment cells None
Waste storage cell None
Chemical processing cell A
Off-gas cell B
Steam generator cells A
Coolant-salt drain cell B
Instruments and controls cells None
Hot cells for repair and inspection None

Y6 in. 88, 9 in. TI, % in. CS, 6 in. AS, % in. CS, coner.

Y & in. 88,9 in, TI, % in. CS, 6 in. AS, % in. CS, coner.

Y6 in. S8, 9 in. TI, % in. CS, 6 in. AS, % in. CS, concr,

% in. CS, coner,

%16 in. 88, 9 in. TI, % in. CS, 6 in. AS, % in. CS, concr.

l/,5 in. 8§, coner.

%6 in. 88, 9 in. TI, 1 in, CS (ribbed), 6 in. AS, % in. CS, concr.
‘/15 in. 88, concr.

Concrete

l/“; in. 8§, concr.

PHeaters: A = heated cell; B = trace heating of equipment.

bApplies also to roof and floor structure, except floor may not have 8 ft of concrete in all cases. Listed as going from interior to
exterior of cell. Floors have "-in. SS and walls % 4-in. SS liners. S8 = stainless steel, CS = carbon steel, TI = thermal insulation
(form of firebrick), concr. = ordinary concrete, and AS = air space.

pipe seals, as shown in Fig. 13.10. The coolant-salt
piping is the principal penetration through the cell wall.

A layer of thermal insulation, not yet selected but
probably a rigid block type, is provided on the inside
surface of the reactor cell. A thin stainless steel liner
protects the insulation and serves as an effective radiant
heat reflector to lower the heat losses through the wall
structure. Although not hermetically sealed, the liner
presents a smooth surface for the inside of the cell.

The reactor, heat exchangers, pumps, and salt piping
are all suspended from the roof of the reactor cell. This
arrangement allows relative thermal expansion of the
components, provides better seismic protection than
pedestal-type mounts, and also makes it possible to
locate the sealing flange for the reactor vessel in a lower
temperature region. The primary heat exchangers are
suspended by gimbal mounts at about midelevation of
the units. This arrangement permits the differential
expansion between the inlet and outlet salt piping to be
accommodated by rotation of the heat exchangers and
thus avoids excessive stresses at any of the components
in the system. (Piping stresses are discussed in Sect.
3.6.)

The reactor cell is heated by hairpin-type Inconel
electric resistance heating units inserted in thimbles
located around the periphery of the cell, as described in
Sect. 11.2. The heater elements can therefore be
replaced without disturbing the integrity of the contain-
ment. The circulating inert gas used to cool the double

walls of the cell is also arranged to cool the heater leads.

As may be seen in Fig. 13.8, the fuel-salt pumps have
their drive motors mounted above the cell roof plugs in
hermetically sealed covers which are, in effect, part of
the outer wall of the reactor cell. The control rod drives
are canned in a similar fashion. This location for the
drive equipment permits easier access for inspection and
maintenance. All the roof-mounted equipment is
covered by a 72-ft-diam dome of %-in.-thick carbon
steel, which provides additional leak protection during
normal operation of the reactor. The dome also is
principal containment during maintenance of the drive
equipment, as discussed in Sect. 12.

A stainless steel catch pan in the bottom of the
reactor cell would collect any spilled salt in the unlikely
event of a leak in the fuel- or coolant-salt systems inside
the reactor cell. This pan is pitched toward a drain
which is connected to the primary-salt drain tank
through two valves in series. The upstream valve is a
special type having a disk punctured by a solenoid-
actuated plunger controlled by a thermal switch. In the
event that hot salt reaches this valve via the catch pan,
the valve would open and permit the spilled salt to flow
by gravity into the drain tank. The valve would be
arranged for replacement using remotely operated
tooling. The second valve is a mechanical bellows-sealed
type that is normally open but can be closed to isolate
the drain tank contents when the first valve is open or is
to be replaced. This catch pan arrangement permits
144

ORNL-DWG 70-44974

{=——40 ft-0in. DIAM

CATCH PAN

OSSN \5 THERMAL INSULATION
\‘/ ~

OIRRHKIHK

0703000 % 0%
. 9. » ———PRIMARY CONTAINMENT
NN \\\ AND THERMAL SHIELD

GAS COOLING

7777,
» « & 00 0p # ]

HARDENED STEEL WAY/

m

7 VZA///V/?"/- SECONDARY CONTAINMENT

|+, ANTI-FRICTION WAY BEARING

!

A

‘. - =
e ; ’
™ f/ "
’ |
v (A ’
¢ A
5 A
: j / "/ :
e &% ‘ 4
Rt /,/////,',//

L
<r'-. ‘ " . P,i‘ o - . "v' .
[ BEAM

Fig. 13.9. Lower support for reactor cell containment vessel.

more rapid cleanup of a salt spill and, in event of a
major loss of fuel salt such as postulated for the
maximum credible accident, is a feasible method of
taking care of the afterheat in the fuel salt.

The roof of the reactor cell consists of removable
plugs arranged in two layers and with stepped joints, as
best shown in Fig. 12.2. The total thickness is 8 ft, and
with few exceptions each layer is 4 ft thick. The plugs
rest on structural steel supports and have a seal pan to
form a leak-tight structure. As previously mentioned, a
cooling flow of inert gas passes between the two heavy
steel plates used for gamma shielding.

13.4 PRIMARY DRAIN TANK CELL

The primary drain tank cell houses the 14-ft-diam,
22-ft-high fuel-salt drain tank. The cell is approximately
22 X 22 X 30 ft deep and is located on the lower level
of the reactor building, as shown in Figs, 13.5 and 6.3.
The requirements for this cell are very similar to those
of the reactor cell, and, in fact, the two cells are
interconnected by the duct through which the fuel-salt
drain line passes. The cells thus operate with the same
ambient atmosphere and essentially at the same pres-
sure and temperature, Gamma shielding is not required
145

ORNL-DWE 70-11972
-~ THERMAL INSULATION
\ / jo ml—Gin

STAINLESS STEEL
LINER ——_

SECONDARY COOLANT

-

T N T T R O T N e S S e W e
o o I L W o y W R e o
bl X e e PP PN,

O RN R ATET AN N AR ROROR T SR TR RN KR XY ALY X X K »
o NN S N O e e KN KN N AN

-+ REACTOR CELL . ! L X
-( CONTAINMENT > 3% el STEAM GENERATOR
o ! ] r’# ’ ! CELL CONTAINMENT
. v, l=—BUILDING ’ A
i Il CONTAINMENT AIR
f COOLING
REACTOR CELL COOLING STEAM GENERATOR CELL

Fig. 13.10. Coolant-salt pipe penetration between reactor cell and steam cell.

to protect the concrete structure of the drain tank cell,  Roof plugs are provided for access to the valves. The
however, and the double walls consist of ',-in.-thick  cell wall construction is essentially the same as that
carbon steel plate. An inert gas is circulated between  used in the drain tank cell. The reactor cell catch pan
these plates for cooling of the wall structure. (The gas  drains into a pan in the freeze-valve cell, and this pan in
stream is an extension of the reactor cell wall cooling  turn drains into the previously described valves leading
system.) Thermal insulation and a stainless steel liner  to the fuel-salt drain tank.

are used on the inside surface, as in the reactor cell.

Removable roof plugs provide access to the drain tank 13.6 SPENT REACTOR CORE AND HEAT

for maintenance through the new core replacement cell, EXCHANGER CELLS

as indicated in Fig. 6.3. The cell floor contains
water-cooled coils to carry off the afterheat in the fuel
salt in event of a major spill.

A cell is provided in the upper level of the reactor
building adjacent to the reactor cell for storage and
dismantling of reactor core assemblies, as shown in Figs.
13.3 and 13.4. The top access opening is shown in Fig.

13.5 FREEZE-VALVE CELL 134, These drawings also show the similar cell for
handling heat exchangers and other radioactive equip-

The freeze valve on the fuel-salt drain line and the ment which has been removed from the system and
valves for the reactor cell catch pan are located in this requires disposal. After a suitable decay period in the
cell. The cell space is directly connected to the reactor cells, the equipment is cut up as required and dropped
and drain tank cell volumes, so that they all operate  through chutes into the hot storage, or waste, cell
with the same atmosphere and essentially at the same located beneath the reactor cell. During the storage
temperature and pressure. The floor area of the  period sufficient fission products will be present on the
freeze-valve cell has the shape of a right triangle (see  equipment to require some cooling, since heat losses
Fig. 13.6) with legs about 18 ft long. The cell is from the cell are low. Both the enclosures therefore
approximately 15 ft deep and is located between the have double walls and use a common inert-gas cooling
reactor cell and the drain tank cell and at about  system which operates in a closed circuit much in the
midelevation between the two, as best seen in Fig. 6.3. same manner as the reactor cell wall cooling system.
A work area is provided adjacent to the above cells
for operation of the remotely controlled equipment
used in the dismantling of the radioactive components,
as indicated in Fig. 13.4, Shielded windows overlooking
the two cells provide visual observation of the proce-
dures. These windows are protected From heat during
the decay period by movable shields.

13,7 WASTE STORAGE CELL

As menfioned above, this waste storage cell is
designed to permanently store waste equipment from
the plant over its useful lifetime, including spent
graphite from the core and radioactive wastes from the
chemical processing plant. It is about the same size as
the reactor cell, 72 ft in diameter and 30 ft deep, and is
located below grade on the lower level directly beneath
the reactor cell. Estimates of the heat generation in the
waste vary over a wide range depending upon the
assumptions used, but the maximum will probably fall
within the 100-t0-600-kW range. A closed-circuit inert-
gas cooling system, similar to those previously de-
scribed, will be used to cool the cell.

13.8 CHEMICAL PROCESSING CELL

A relatively large shielded area with 60-f1 cell height
has been set aside in the reactor building for the
fuelsalt processing equipment, as indicated in Figs.
13,5 and 13.6. This cell will be heated as a furnace and
will employ coolers and thermal insulation as required
for individual control elements, etc. The cells will be
heated to the desired operating temperature by resist-
ance heaters, as described in Sect. 11.2. Remote
maintenance facilities, cell integrity, etc., will be similar
to other cells containing highly radioactive materials.

13.9 OFF-GAS SYSTEM CELL

The cell for treating the off-gas is similar to the
chemical processing cell described above. The cell
houses the charcoal adsorber beds and other equipment
needed for treatment of the radioactive gases taken
from the primary circulating system.

13.10 MISCELLANEOUS REACTOR BUILDING
CELLS

In addition to the above-mentioned cells, the reactor
building contains hot cells for examination, analysis,
and repair of radivactive equipmen! and malterials, cells
for storage of control rods, storage of new reactor core
assemblies, work areas, and a relatively large cell set

146

aside for instrumentation and controls equipment. The
locations of these cells are shown in Figs. 13.3—13.6.

13.11 STEAM-GENERATOR CELLS AND
SERVICE ARFEAS

There are four steam-generating cells in the reactor
building, each 30 X 48 X 30 ft deep. The cells are at the
same elevation as the primary heat exchangers in the
reactor cell, and each contains a coolant-salt circulation
pump, four steam generators, two reheaters, and inter-
connecting coolant-salt and steam system piping. The
cells are sealed and provided with biological shielding
because of the induced activity in the coolant salt and
the remote possibility that fuel salt could enter the
steam cell via the coolantsalt circuit. Tritium might
also find its way into the cell. Since the steam cells will
be heated to about 1000°F to ensure that the coolant
salt remains above its liquidus temperature, thermal
insulation is provided at the walls, and a double wall
with a circulated inert-gas cooling system, such as
employed in the reactor building cells, is used to
protect the concrete from excessive temperatures.

A principal consideration in the conceptual design of
the steam cells was selection of the design pressure. A
major possible source of pressure buildup is the
emergency relief of the steam system into the cell via
the rupture disks provided in the coolant-salt circuits.
(In event of a major leakage of steam into the coolant
salt these disks would prevent a pressure buildup on the
primary heat exchanger tubes.) To curtail the amount
of steam that could expand into the steam cell by this
route, quick-acting stop valves are provided on the
steamn generator unift in each cell so that the loss of
steam can be restricted to little more than that
contained in one steam generator, On this basis, a
50-psig design pressure was assumed for the steam cells.

The wall construction is similar to that used in the
reactor cell. The inner wall transmits a portion of the
pressure loading through spacers to the outer wall.
Provisions are made for differential thermal expansion
of the two steel shells.

A cell for the coolant-salt drain tank is located on the
lower level directly beneath the steam cells. This drain
tank will utilize heater equipment on the tank and
obviate the need for the furnace concept of cell heating.

The reactor building also includes several service
areas, many of which can be conventional building
construction. These include the control rooms, shops,
equipment assembly spaces, instrumentation rooms,
storage spaces, and, at the base of the stack, a cell for
the drain tank and off-gas heat-removal equipment.
13.12 FEEDWATER HEATER AND TURBINE
BUILDINGS

The steam system equipment requires greater building
space than does the reactor system. As shown in Figs.
13.1 and 13.2, there are three buildings, or bays,
associated with the turbine plant: (1) the feedwater
heater and steam piping bay, 112 X 257 X 154 ft high;
(2) the turbine-generator building, 133 X 257 X 124 ft
high; and (3) an unloading and equipment setdown
area, about 50 X 257 X 75 ft high.

The buildings were not studied in any detail and no
optimization studies were made, since the structures

147

will follow conventional power station practice. The
layout dimensions for the tandem-compounded
1000-MW(e) turbine-generator are not exact, but the
building dimensions are probably representative. A large
building is shown for the feedwater heater space since
this area also included manifolding and large thermal
expansion loops for both the throttle and reheat steam
lines.

It is visualized that these buildings will be of steel
frame construction, with steel roof trusses, precast
concrete roof slabs, concrete fleors with steel gratings
as required, and insulated aluminum or steel panel
walls.
14. Site Description

The site assumed for the MSBR station is the AEC
standard.’'? Briefly, this site consists of grass-covered
level terrain adjacent to a river which has adequate
cooling water conditions to maintain an average 1% in.
Hg abs back pressure for the turbine. The ground
elevation is about 15 ft above the mean river level. A

RIVER FLOW
Ap—

limestone formation about 30 ft thick has its top about
8 ft below grade and has a bearing capacity of 18,000
psf.

The general layout of the site is shown in Fig. 14.1.
Intake and discharge structures for cooling water, a
deep well, a water purification plant, and a water

ORNL-DWG 70-11973

e

"'m,m,
PUMPS -~ '.?.;, r';{/‘,/

— ) ’ NORMAL
] ~ _ RIVER LEVEL
0
INTAK
= & & T
. 0
/ ELEVATION
15 ()
20

o,,

 WATER PURIF ICATION

TO oowoeussas
© DEEP WELL
‘\ "\ WATER
( )STORAGE
\ TRANSFORMERS
FEEDWATER PUMP \ SHOPS —
SANITATION \\ /
DIS - -
i M
— | swITcH-
i 0| e
Ay EATER ]
BAY ]
/ C]
/nmo

Fig. 14.1. Plot plan for 1000-MW(e) MSBR power station.

148
storage tank are provided. The electric switchyard is
adjacent to the plant, and a railway spur serves for
transportation of heavy equipment. An oil tank is
shown for storage of fuel, although natural gas is a more
likely candidate for fueling the startup boiler. The usual
services are provided, including a waste-treatment plant
for the sanitary discharge.

The standard site assumes the electrical distribution
system to be single-source transmission and would be
subject to occasional outages. An emergency power
source is therefore required in the plant.

The site is assumed to have a sufficient frequency of
tornado occurrences to require class I structure design.

149

Seismic disturbances in the area have ranged 4 to 6 on
the Mercalli scale (equivalent to about 0.007 to 0.07 g
horizontal ground acceleration), and the site has been
designated as zone 1, that is, an area which is normally
below the threshold of damage.

The site location is satisfactory with respect to
population centers, meteorological conditions, fre-
quency and intensity of earthquakes, heat discharge,
and other environmental factors, so that no special
design conditions or costs are imposed other than those
“normally” expected to meet licensing requirements.
15. Cost Estimates for the MSBR Station

15.1 CAPTIAL COST ESTIMATE

Roy C. Robertson M. L. Myers
H. I. Bowers

A capital cost estimate for the reference MSBR
station is given in Table 15.1. Sources of the data are
explained in the footnotes to the table, and the details
of the estimates are included in Appendix D. To give a
frame of reference for the MSBR estimates, the costs
are compared with those for a PWR.

The capitalization costs for the two reactor types are
not greatly different. In a broad sense this can be
explained by the fact that only about one-third of the
total cost is for reactor equipment, the remainder being
for the heat-power syslem, general facilities, and in-
direct costs, which are expenses that are somewhat
similar for all thermal power plants. Variations in
reactor equipment costs are not of sufficient magnitude
to cause striking differences in the overall capital
requirement because there are rough similarities in costs
of vessels, shielding, etc., and many of the differences
that do exist are offsetting.

Insofar as possible the MSBR and PWR cost estimates
were put on the same basis. In both estimates the cost
of the fuel-processing plant is included in the fuel cost
rather than in the plant capital cost. Both estimates use
the accounts recommended in NUS-531 (ref. 119), are
based on the Janvary 1970 value of the dollar, and
include indirect costs of about 35%. Private ownership
of the plants is assumed, and interest (at 8%) during a
five-year construction peniod is included. Neither esti-
mate, however, considers escalation of costs during the
construction period.

The Hastelloy N equipment in the MSBR is assumed
to have a fabricated cost of $8 to $38 per pound,
depending upon the complexity (see Table D.4). The
reflector graphite is estimated to cost $9 per pound and
the extruded core elements $11 per pound (see Table
D.5).

It is important to note that the MSBR construction
cost estimates are not for a first-of-akind plant but
assume that the station is of a proven design for an

150

established molten-salt reactor industry in which de-
velopment costs have been largely absorbed and in
which manufacture of materials, plant construction,
and licensing are routine. As recommended in NUS-531
(ref. 119), however, recognition was taken of the fact
that the MSBR cost estimate is based on conceptual
designs rather than on actual construction experience,
and a 15% contingency allowance was applied to
reactor materials. A contingency factor of only 3% was
used in the corresponding portion of the PWR estimate.
As indicated in Table 15,1, this difference in con-
tingency factors applied to the reactor materials adds
about $8 million to the total MSBR cost estimate after
indirect costs are included.

One of the distinguishing features of the MSBR
station is the use of initial steam conditions of 1000°F
and 3500 psia, with reheat to 1000°F. As shown in
account 231, Table 15.1, a turbine-generator for these
conditions has a relatively low first cost compared with
the turbine-generator for a PWR. Good utilization of
the available heat in the MSBR is reflected in the
relatively low steam mass flow rates and amount of heat
transfer surface needed. Although no credit was taken
for it in the MSBR cost estimate, this factor could also
influence siting and environmental control costs in that
the heat rejected to the MSBR condensing water is only
about one-half that for the PWR.

The alternate reactor vessel head assembly used to
facilitate replacement of the core graphite in the MSBR
is included in the first cost of the plant. The estimate
also includes the special maintenance equipment used
for the replacement operation. The MSBR does not
consider a safeguards cooling system (account 223,
Table 15.1) as such but does require a drain tank with
afterheat-removal capability, as included in account
225, Table D.1. In several instances, such as the off-gas
cooling system, cell heating and cooling systems, etc.,
the conceptual design work was not sufficiently de-
tailed to serve as a basis for a cost study, and the values
used in Table D.1 are more in the nature of an
allowance than an estimate.
151

Table 15.1. Summary of 1000-MW(e) MSBR station
construction costs and comparison with PWR station
costs

Expressed in millions of dollars and based on
January 1970 costs

Account No. item MSBR?  PWR?
20 Land 0.6 0.6
21 Structures and site facilities 28.8 25.6
22 Reactor plant equipment

221 Reactor equipment 18.0 17.8
222 Main heat transfer systems 25.2 29.2
223 Safeguards cooling system 4.1
224 Liquid waste treatment and disposal 0.7 0.7
225 Nuclear fuel storage 4.2 13
226 Other reactor systems and equipment 9.8 0.5
227 Instruments and controls 4.0 ol
Contingencies and spare parts 9.0 2.9
Total account 22 70.9 61.6
23 Turbine plant equipment

231 Turbine-generator 20.8 32.7
232 Condensing water system 2.0 3.1
233 Condensers " 4P 4.7
234 Feedwater heating system (A 6.1
235 Other turbine-plant equipment 6.2 39
236 Turbine instruments and controls 0.5 0.7
Contingencies and spare parts 2.2 2.5
Total account 23 41.6 53.7
24 Electric plant equipment 8.0 8.0
25 Miscellaneous plant equipment 2.0 2.0

26 Special materials 1.0
Total |direct construction cost 152.3 150.9
91-94 Indirect costs 50.3 49.2
Total capital investment 202.6 200.7

@Details of the MSBR cost estimate are given in Appendix D.

DPWR costs were taken from studies made in connection with the capital cost
computer program being developed at ORNL for the AEC under the Studies and
Evaluation Program (report to be published). Costs were escalated from a
mid-1967 basis to Janvary 1970. Some accounts were adjusted to reflect
increased costs due to design changes dictated by more stringent safety
requirements, as discussed in a United Engineers report (ref, 120).

require plant outages in addition to those accommo-
dated by the plant factor. The capital cost of the fuel
processing equipment for the MSBR is not known with
certainty at this time due to the preliminary nature of
the conceptual designs for the equipment and the use of
relatively large amounts of molybdenum as a construc-

15.2 POWER PRODUCTION COST

The estimated cost to produce electric power in the
reference design MSBR station is shown in Table 15.2.
The table is based on 80% plant factor, January 1970
conditions, and fixed charges of 13.7% on the station

capital cost and 13.2% on the fuel inventory. (Other
assumptions are given in the footnotes to Table 15.2.)

The cost for periodic core graphite replacement in the
MSBR is included as a separate production cost in Table
15.2. It is assumed that the core maintenance does not

tion material, for which there is little background of
cost experience. The effect of the chemical plant
capitalization on the fuel cycle and total power
production costs is indicated in Fig. 15.1. The MSBR
fuel-cycle cost shown in Tables 15.2 and D.2 is based
on an assumed expenditure of $13.5 million (including
indirect costs) for the chemical plant equipment.

Two other uncertainties entering into the MSBR cost
estimates are the cost of graphite and the life of the

1apie 15.Z. Esumated power production
cost (mills/kWhr) in the MSBR station®

Fixed charges on total plant capital investment at 13.7%? 4.0

Cost of periodically replacing graphite® 0.2
Fuel cycle cost? 0.8
Operating cost? 0.3

Total 5.3

@Based on investor-owned plant and 80% plant factor.

PBased on capital costs shown in Table D.1 and fixed charges
of 13.7% on depreciating equipment, as listed in Table D.14,
and 12.8% on land, as recommended in NUS-531 (ref. 119).

“The graphite replacement cost is shown in Table D.15.

dMSBR fuel cost, as shown in Table D.2, is based on 13.2%
fixed charges on inventory capitalization, on the 1970 value of
the dollar, and a total cost for fuel processing equipment of
$13.5 million.

€Estimated operating costs are shown in Table D.16. These
costs are based on the recommendations in NUS-531 (ref, 119)
and agree reasonably well with those reported by Susskind and
Raseman (ref. 121).

8 ORNL—DWG 70—-11974

| TOTAL POWER

PRODUCTION COST

mills/kwhr
D

2 FUEL CYCLE}ST/

0 20 40 60

MSBR FUEL—PROCESSING PLANT COST
(million $)

0

Fig. 15.1, Effect of MSBR fuel-processing plant capital cost
on fuel-cycle and power production costs (based on fuel at
$13/g).

152

reactor core before it would require replacement. The
effects of these two factors on the cost to produce
electric power are shown in Figs. 15.2 and 15.3.

Power production costs for the MSBR were based on
the present “standard” fuel cost for 233U of $13 per
2357 of $11.20 per giaw, and a conespund-
ing cost for 23?Pu of $9.30 per gram. A comparison of
MSBR production costs with those of other reactor
types should take into account the changed price
structure of nuclear fuels that will undoubtedly exist by
the time molten-salt reactor power stations are con-
structed in quantity, since these changes in prices and
fuel resources could have a significant effect on the
molten-salt reactor economics. The next 30 years could
witness significant changes in the sizes of plants, in
light-water fuel-cycle costs,122:123 swings in the price
of plutonium, and use of cross-progeny fueling of
reactors.! 2% Also, the higher market value of electric
power will be a feedback into fuel diffusion and
separation plant operating costs and will change the
relative costs of fissile fuels. Analysis of these com-
plexities is beyond the scope of this report. It can be
stated here only that the estimated power-generating
costs for the molten-salt reactor appear competitive and
that the concept gives promise of making important
future savings in the nation’s fuel resources.

P
Braiii, 101

ORNL—-DWG 70—11975

230 . 5.8
| -
| E
s | mills/kwhr 1 2
§ 220 } L 56 E
£ | 7 %
- | A S
2 |~ 3
O 210 I N S
- $40 ta
= = |
o 7 =)
a 7. S
3 e | o
- 2
-5 I REFERENCE e
= [~ DESIGN z
| &
l
190 L 5.0
5 0 5 20

AVERAGE GRAPHITE COST ($/Ib)

Fig. 15.2. Effect of average graphite cost on total MSBR
plant cost and power production cost (based on fuel-cycle cost
of 0,8 mill/kWhr and four-year graphite life).
TOTAL POWER PRODUCTION COST (mills/kwhr)

6.0

o
®

o
o

4
»

o
()

5.0

ORNL-DWG 70— 11976

AVERAGE GRAPHITE
COST OF $20/Ib

\\\

AVERAGE GRAPHITE
COST OF $10/1b

8 16 24
GRAPHITE LIFE (years)

32

153

Fig. 15.3. Effect of graphite life on total MSBR power
production cost (based on fuel-cycle cost of 0.8 mill/kWhr).
16.

Uncertainties and Alternatives, and Their Effects

on Feasibility and Performance

E. S. Bettis

16.1 GENERAL

In making this conceptual study it was necessary to
base some of the judgments on preliminary designs, test
results, properties of materials, and other design infor-
mation that will require further study and verification.
While these judgments were made conservatively and it
is reasonable to expect that some aspects will perform
even better than anticipated, a primary concern is the
effect on MSBR feasibility if one or more of the design
uncertainties prove to be very difficult or expensive to
resolve or if the behavior falls significantly short of
expectations.

The major uncertainties as now known are in the
areas of tritium confinement, fuelsalt processing,
graphite and Hastelloy N behavior, suitability of the
coolant salt, maintenance procedures, and behavior of
the off-gas particulates, This section discusses the
impact of these and other uncertainties on MSBR
feasibility in relation to safety, nuclear performance,
dependability, and economics of power generation. The
order of discussion is by systems rather than by degree
of uncertainty.

16.2 MATERIALS
16.2.1 Fuel Salt

As discussed in Sect. 3.2.1, the composition of the
MSBR fuel salt was chosen on the basis of neutron cross
sections, viscosity, chemical stability, and liquidus
temperature. There is little uncertainty with regard to
its phase behavior, most of its physical properties, its
behavior under irradiation, and its interactions with the
container and moderator materials. Less well known are
the effects of the oxidation-reduction state of the salt
on its surface tension and on the behavior of the noble

P. N. Haubenreich

154

Roy C. Robertson

metal fission products. Significant limitations to use of
the salt are imposed by its rather high liquidus
temperature (930°F), the limited solubility of uranium
oxide (about 40 ppm of the oxide ion), and the
restricted choice of container materials. The problem
that looms largest at the present is the production of
relatively large amounts of tritium by neutron inter-
action with the lithium, as will be discussed in Sect.
16.4.

Some variations in the composition of the fuel salt are
possible and may prove desirable to circumvent or to
mitigate some of the above-mentioned limitations. The
UF,; and ThF, concentrations can be varied as required
for criticality and optimization of the breeding per-
formance, The continuous processing of the fuel salt is
expecied to keep the oxide concentration low and to
make a low UO, solubility acceptable. The oxide
tolerance of the salt can be increased by the addition of
ZrF; (as was done in the MSRE), although at the
expense of parasitic absorption of neutrons in the
zirconium and complication of the chemical processing.
The constraints of a high liquidus temperature and the
problem of tritium cannot be mitigated, however. If the
molten-salt reactor is to breed with thermal neutrons,
cross sections limit the choice of diluent salt constit-
uents to the fluorides of beryllium and lithium (with
very low ®Li content).”® In the LiF-BeF,-ThF, system
(Fig. 3.5a), liquidus temperatures much below that of
the reference MSBR salt cannot be attained without
reducing the ThF,; concentration to the extent that
breeding performance is seriously impaired. The tritium
production in a molten-salt reactor could be cut to little
more than the fission yield if an NaF-ZrF,-ThF,-UF,
fuel salt were used, but neutron absorptions in the
sodium and zirconium would preclude breeding. In
summary, if the molten-salt reactor is to breed, there is
no reasonable alternative to fuel salt of the approximate
composition chosen for the reference study. The
limitations attending its use must therefore be accom-
modated in the design.

The market price of ”Li has a limited effect on the
total fuel-cycle cost. For example, if the price of
99.99% "Li as lithium hydroxide monohydrate were
doubled from the $120 per kilogram assumed in the
reference design, the MSBR fuel-cycle cost would he
increased from about 0.76 to 0.82 mill/kWhr.

16.2.2 Secondary Fluid

As stated in Sect. 3.2.2, the factors determining the
choice of the fluid for the secondary system are
chemical stability, susceptibility to radiation damage,
compatibility with materials of construction, heat
transfer and fluid flow properties, and cost. The fluid
chosen for the reference design, sodium fluoroborate,
offers advantages over other fluids in some of these
areas and on the whole promises to be an acceptable
material to use. There are some problems associated
with it, however, and some remaining uncertainties.
These are discussed below, followed by a discussion of
alternative fluids and the influence their use would have
on the design and performance of the MSBR.

Loop tests have shown that if water can be excluded,
the sodium fluoroborate is quite compatible with
Hastelloy N, with corrosion rates of only about 0.2
mil/year at MSBR temperatures. While it is possible to
limit the water intrusion into test loops to very small
amounts, it is not certain to what limits it will be
practical to restrict entry of water by leakage from the
steam generators. The corrosion rate to be expected in
an operating MSBR is thus somewhat uncertain. Tests
in which steam was deliberately added to fluoroborate
systems showed corrosion of Hastelloy N at a rate
above 20 mils/year for a week or so after the
addition.’' The effect of continuous injection of water
into a fluoroborate system has not been studied, but it
appears that very little continuous leakage can be
permitted in an MSBR. Whether it will be practical to
guarantee a sufficiently low leakage rate remains to be
determined.

The reaction between water and fluoroborate is not
violent and should contribute little if anything to the
wastage of metal by a high-velocity jet of water from a
leak in a steam generator. There has been no experi-
ment of this sort with fluoroborate and water, how-
ever, so the requirements for immediate response to a
steam leak cannot be specified realistically at the
present time.

155

Processing is likely to be required to hold the
corrosion products and other undesirable contaminants
to low concentrations in the salt. The requirements for
processing have not been established, but no major
technical difficulties are expected to be encountered in
developing a purification process.

The consequences of mixing sodium fluoroborate
with the MSBR fuel salt (as through a leak in a primary
heat exchanger) have not been considered in detail.
Wastage and enhanced corrosion are not likely to be
serious, but the amounts of inleakage must be limited
for other reasons. Boron trifluoride gas is likely to be
evolved as the fluoroborate salt mixes with the fuel salt,
and, depending upon the extent of mixing, phases with
high melting temperatures may be formed. Although
the high-cross-section boron could be sparged from
the fuel salt as BF; gas, the sodium, unless chemically
removed, would remain in the fuel salt and diminish the
breeding performance. The sodium from about 100 ft?
of coolant salt would reduce the breeding ratio from
1.063 to 1.056.

The cover gas for fluoroborate must be the proper
mixture of BF; and inert gas to prevent changes in the
NaF-NaBF,; composition. The off-gas from flucroborate
loops has been found to contain various condensables
which require special handling. These problems have
been dealt with in a practical manner in development
tests, but the gas systems for fluoroborate loops tend to
be somewhat more complicated than if some other salts
were used for heat transport.

If the results of further tests of fluoroborate should
indicate that its use in the MSBR would be impractical,
the most assured alternative is the 2LiF-BeF,; mixture
that was used in the MSRE. Its use as the secondary salt
in the MSBR would eliminate problems of chemical
compatibility with the fuel salt. (Separated "Li would
have to be used, however, because mixing would
otherwise require expensive isotopic purification of the
lithium in the fuel.) The corrosion situation would be
alleviated, possibly easing the restrictions on moisture
contamination and widening the possibilities for con-
tainer materials. Constraints and penalties would be
imposed, however, because of the higher melting point
and much greater cost of 7LiF-BeF, relative to NaF-
NaBF,. The liquidus temperature of LiF-BeF, (66-34
mole %) is 858°F, compared with about 725°F for
NaBF,-NaF (92-8 mole %). This would complicate the
design by requiring a higher degree of feedwater heating
and/or special design of the steam generators. Equip-
ment costs and plant thermal efficiency would be
adversely affected, but the greatest penalty would be in
inventory charges. If the volume of coolant salt were
the same (8400 ft*) the "LiF-BeF, inventory would
cost $13 million compared with $0.5 million for
fluoroborate, This difference amounts to ~0.3
mill/kWhr in power costs.

Another candidate for the secondary fluid is a
mixture of potassium and zirconium fiuorides of the
composition KF-ZrF,; (58-42 mole %). This mixture has
received little attention to date because its 750°F
liquidus temperature is higher than that of sodium
fluoroborate. It has a low vapor pressure, reasonably
good heat transfer properties, and is relatively inex-
pensive (about $1 per pound). The effects of mixing
with fuel salt and with water are unexplored.

Other alternative coolants are considered inferior or
impractical for various reasons. Nitrate-nitrite mixtures
(Hitec, for example) would be cheap, probably would
block tritium transfer to the steam system, and would
permit design simplifications because of their relatively
low melting points (around 300°F). Their stability and
corrosion behavior above about 1000°F are not well
known, however. The most serious drawback to their
use as a secondary salt is that the nitrate-nitrites would
precipitate UO, if they leaked into the primary system
and possibly would react violently with the graphite.

Alkali metals are undesirable because they react with
both fuel salt and steam. Metal coolants such as lead or
bismuth undergo no violent reactions, but they are not
compatible with Hastelloy N or other nickel-base alloys.
Several binary chloride systems have eutectics melting
below 700°F, but the more stable nonvolatile chlorides
are those containing lithium, which would be expensive
if 7Li were used. High-pressure gas (possibly containing
moisture to trap tritium) has some advantages as a
secondary coolant, but would open the possibility of
excessively pressurizing the fuel system, and the poorer
heat transfer with gas would substantially increase the
inventory of fuel salt in the primary heat exchangers.

16.2.3 Hastelloy N

Although additional work is needed on the use of
Hastelloy N for the container material for the fuel and
coolant salts, the remaining uncertainties are not
sufficient to jeopardize the feasibility of the MSBR.

Hastelloy N suffers embrittlement in a neutron
environment, and the damage increases with the total
fluence and operating temperature. The approach used
in this study has been to limit the temperature and the
neutron exposure of the more critical portions of the
reactor vessel. Since there are to date no approved code
cases for irradiated Hastelloy N upon which to base a
design criterion, the considered judgment is that the

156

irradiation should be limited to the extent that the
creep ductility will not be less than 5%. The standard
alloy of Hastelloy N does not meet this requirement,
The advances described in Sect. 3.2.4 for obtaining a
modified Hastelloy N with adequate resistance to
radiation embrittiement (through use of additives, such
as titanium, hafnium, and niobium) appear very prom-
ising, but further testing is needed to select the best
composition. Large heats must be obtained to show
that the favorable properties are retained in commercial
materials, and the modified alloy must be subjected to
enough testing to have it approved for pressure vessel
use by the ASME.

In the event that the embrittlement problem imposes
more severe limitations than now expected, the design
can be revised to make more use of the 1050°F inlet
salt to the reactor to cool the higher-temperature
portions of the vessel, such as the outlet nozzles. A
further recourse would be to reduce the outlet salt
temperature from the reactor to 1200—1250°F. Re-
ducing the outlet temperature would require a higher
circulation rate and larger inventory of salt in the
primary system but would not necessarily lower the
steam temperature and the thermal efficiency of the
cycle, as discussed in Sect. 16.7. The effects are not
great enough to threaten the feasibility of the MSBR
concept.

In this study the allowable design stress of standard
Hastelloy N was taken to be 3500 psi at 1300°F, a
stress that has received ASME code approval. The
standard alloy consistently shows better strength char-
acteristics than those upon which the code case was
approved, and the additives increase the strength of the
modified Hastelloy N. What adjustments will be made
in the code-approved allowable design stress for Hastel-
loy N are not certain, but they may permit higher
stresses and thinner metal sections in the reactor vessel.
As mentioned above, this would help to lower the
estimated maximum metal temperature and ameliorate
the radiation damage problem.

The modified alloy is expected to be as resistant to
corrosion by fluoride salts as standard Hastelloy N, but
the behavior must be demonstrated in tests with fuel
and coolant salts under simulated reactor operating
conditions.

Hastelloy N is specified as the material of con-
struction for the steam generators in the reference
design, and, as mentioned in Sect. 3.2.4.2, both the
standard and modified alloys have demonstrated good
resistance to corrosion by supercritical-pressure steam
at 1000°F in tests made in the TVA Bull Run steam
plant. The data were obtained with unstressed speci-
mens, however, Stressed samples are being tested at Bull
Run, and these results will be important in assessing the
compatibility of Hastelloy N with steam. If the material
proves unsatisfactory for service in water and steam, the
probable solution would be 1o use tubes of Incoloy 800
clad with nickel on the salt side and to clad the water
side of the vessel heads and tube sheets with Incoloy or
Inconel.

16,2.4 Graphite

At the present time industry does not have the
facilities for manufacturing Lhe large-sized pieces of the
special grade of graphite needed for a 1000-MW(e)
MSBR. Although there is some confidence that core
elements of the desired length (about 20 ft) can be
extruded, failure to meet this objective would require
that the elements be assembled from shorter sections.
This would add to the cost and would be inconvenient.

The important uncertainties with regard to MSBR
graphite are gas permeability, usable life, and the cost
of the installed material. The gas permeability affects
both the breeding performance and the power produc-
tion cost; the useful life and the graphite price primarily
affect the production cost alone. In general, these
aspects are examples of uncertainties where future
development is likely to lead to improved situations
rather than worse ones, but, to pursue the objectives of
this section, the consequences of unfavorable develop-
ments will be reviewed.

16.2.4.1 Gas permeability. With the turbulent flow
assumed through the reactor core, the graphite must
have a gas permeability in the order of 107® cm? /sec to
keep the xenon poison fraction down to the 0.5% used
as a “‘target” in the reference MSBR design and as a
basis for the performance estimates. This resistance to
gas diffusion can be achieved only by sealing the
graphite. Small pieces have been successfully sealed to
these standards, and methods for treating the MSBR
core elements can probably be devised, but nevertheless
sealing of the large pieces remains to be demonstrated.

While sealing the graphite to minimize xenon absorp-
tion is desirable, it is not essential to the MSBR
concept. Figure A.2 shows the calculated effects of
coating thickness and permeability on xenon poisoning
when used in conjunction with a reasonably effective
gas sparging system. Even with unsealed graphite
(helium permeability 107° cm?/sec) the calculated
poison fraction is less than 2%. Allowing the xenon
poisoning to increase from the reference value of 0.5 to
2.0% is estimated to reduce the breeding ratio of the
MSBR from 1.063 to 1.045. Recent measurements
indicate that the mass transfer coefficients used in the

157

calculations are conservative and that the effects may
not be this great.

As indicated in Fig. A.2, a 5-mil surface layer on the
graphite having a permeability of 107 cm? /sec for the
coating is enough to permit the sparging system to hold
the xenon poison level to the target value of 0.5%. This
degree of sealing has been achieved with pyrolytic
carbon, as discussed in Sect. 3.2.3. The serviceability of
sealed graphite and the cost of the sealing are yet to be
resolved. Plugging of the graphite pores by a vacuum-
pulse gas impregnation process produces a tight surface,
but under neutron irradiation the permeability increases
very rapidly, and dimensional changes are apparently
accelerated above the rates obtained with unsealed
graphite. Deposition of pyrolytic carbon on specimens
in a fluidized-bed furnace gave coatings 3 to 5 mils
thick with permeabilities of <107 cm? /sec. Irradiation
tests of these specimens are encouraging, but the
coatings are relatively easy to damage by handling.

If the target xenon poison fraction cannol be attained
and a longer doubling time must be accepted in any
event, consideration can be given to designing the
reactor for laminar flow in the core. The power density
must be reduced considerably, and this increases the
doubling time because of the larger core volume, but
the breeding gain is not as dependent upon sealing the
graphite. The lower power density would increase the
graphite life and reduce the frequency of graphite
replacement, although this factor may have limited
importance, as discussed below.

16.2.4.2 Useful life of graphite, The lifetime of the
graphite is limited by the requirement that it be
impermeable to the fuel salt. As explained in Sect.
3.2.3, this requirement is readily met when the graphite
is new, but there is an uncertainty as to how long the
graphite will remain impermeable under fast-neutron
irradiation. In the absence of conclusive measurements,
the useful life of the graphite in the MSBR has been
defined as the point at which the most highly irradiated
graphite in the core expands past its original density.
This appears to be conservative in that the graphite
probably remains impermeable to salt to somewhat
beyond this point, An additional conservatism in the
reference design was the assumption that the MSBR
graphite would last no longer than commercial grades
currently available, Improved graphites with consider-
ably longer life could result from the development now
in progress, although probably not to the point of
lasting the 30-year life of a plant at the proposed power
density.

Replacement of the core graphite entails not only the
periodic expense for new graphite but also the capital
cost of a reactor design which permits core replace-
ment, the maintenance equipment, and the expenses
attendant to handling the highly radioactive core
material, Once the investment is made in the special
provisions for graphite replacement, however, the elec-
tric power production cost is not very sensitive to the
replacement interval required. As shown in Fig. 15.3,
there would be only modest savings if the graphite were
good for 8, or even 16, years instead of the 4 years
assumed in the reference design, It should be noted,
however, that these costs assume that the core graphite
can be replaced in a time that can be accommodated in
the 0.8 plant factor. If an outage of many months is
required for graphite replacement, the power cost
would of course be more sensitive to the graphite life.

16.2.4.3 Graphite cost. The costs shown in Fig. 15.3
and those given elsewhere in this report are based on an
installed cost of graphite of $9 to §11 per pound. Some
estimators believe that large-scale production of graph-
ite would bring this price down, but others think it is
too low, particularly if special measures to seal the
graphite against Xenon prove to be expensive. Figure
15.2 shows the effect of the graphite price on the
power production cost, based on a four-year replace-
ment interval. If the graphite proved to cost, say, $20
per pound, the increase in the power cost is about 0.2
mill/kWhr.

16.3 SYSTEMS AND COMPONENTS
16.3.1 Reactor

The conceptual design of the reactor core and vessel
was carried only to the point of indicating feasibility
and performance. A more detailed study would un-
doubtedly disclose some problem areas not yet delin-
eated, The basic arrangement appears sound, however,
and it seems certain that an acceptlable design can be
made for a molten-salt reactor core and vessel. Perhaps
the largest uncertainties are in the procedures for
replacing the core graphite, They will be discussed
separately in Sect. 16.8.

Some of the aspects of the reactor design that will
require particular attention before arriving at a final
design are:

1. A detailed analysis must be made of the temper-
ature and stress distributions, particularly in the high-
temperature regions. As discussed in Sect. 16.2, some
adjustments may be necessary to keep radiation damage
in the graphite and Hastelloy N to within tolerable
limits, The outlet nozzles on the vessel have not yet
been analyzed in detail for stresses.

158

2. The core hydrodynamics needs to be studied,
using models, to check the flow distribution and to
eliminate any tendencies that may exist for flow-
induced vibrations.

3. The methods suggested in the conceptual design
for sccommecdating dimensicnal changes in the graphite
will require more detailed design.

4. The exact number of control and safety rods needs
to be determined. The drive mechanisms for the rods
have not been studied in detail, but since a fast-
scramming action is not necessary, the requirements do
not appear to be stringent. Dimensional changes that
occur in the control rod graphite can undoubtedly be
accommodated, but the expected life of the rods and
the means for replacement need to be studied in more
detail.

5. The methods proposed in the conceptual study for
mounting the reactor vessel and making the top closure
will require more detailed design. The earthquake
resistance of the reactor support system was indicated
to be satisfactory in preliminary studies, but a more
comprehensive analysis is needed.

16.3.2 Primary Heat Exchangers

Although not a serious factoy in determining the
feasibility of the MSBR concept, an uncertainty in the
primary heat exchanger design presented in this report
is the use of special tubing in certain portions to
enhance the heat transfer, The enhancement consists in
indenting a shallow spiral groove in the tube wall. Tests
with water indicated that the groove improves the heat
transfer coefficient on the inside by a factor of about 2
and on the outside by a factor of about 1.3. These and
other heat transfer data need to be confirmed with
circulating salt, however, The tubes do not appear to be
weakened by the grooving process, but more informa-
tion is needed, particularly with regard to the effect on
collapsing strength. Tubing manufacturers have indi-
cated a capability for producing the tubing at a
reasonable cost.

The penalty for using plain tubes rather than en-
hanced tubes would be a need for more heat transfer
surface and an increase of about 5% in the total fuel-salt
inventory of the primary system. Although this would
lengthen the doubling time, the feasibility of the MSBR
is not contingent upon preventing this small increase.

16.3.3 Salt Circulation Pumps

The salt circulation pumps used in the MSRE and in
test loops have performed well, and the manufacturers
believe that they can be extrapolated to the capacities
needed in an MSBR with few development difficulties.
The larger size can probably use an overhung shaft and
impeller to eliminate the need for a lower bearing
operating in the salt, but this remains to be demon-
strated. If the lower bearing is required, salt bearing
development work already accomplished at ORNL
appears promising. A disadvantage of use of the
salt-lubricated bearing is that the pumps could not be
operated to circulate gas during warmup of the system
before it is [illed with salt. In this event the startup
equipment and procedures would have to be revised.

16.3.4 Drain Tank

The primary drain tank approaches the reactor vessel
in complexity and cost, yet in this conceptual study
relatively little effort could be devoted to its optimiza-
tion. The design of the drain tank is strongly influenced
by the drain system flowsheet. The proposed method of
cooling the drain tank head and walls by a continuous
salt overflow from the pump bowls, use of jet pumps to
return the salt to the primary system, and employing
the drain tank for holdup and decay of off-gases are all
aspects of a drain system which represents but one of
many possible arrangements. Study of the drain system
flowsheet is continuing at ORNL, and some revisions
may be necessary, particularly with regard to the
continuous salt letdown and pump-back arrangement.
The modifications are not likely to increase the
complexity and cost, however.

For the drain tank design proposed in this report, it
will be necessary to evaluate the performance of the jet
pumps and possibly to substitute centrifugal pumps;
investigate the radiant heat transfer aspects; carefully
consider the behavior of noble metal fission product
particles brought down with the off-gases; demonstrate
the reliability of the cooling system; and provide the
required means for inspection and maintenance. As
indicated in Sect. 6,4, a NaK cooling system for the
drain tank may be superior to the proposed salt cooling
system. Other improvements are likely to result from
more detailed study of the design.

16.3.5 Fuel-Salt Drain Valve

The reference MSBR design proposes that the “‘valve”
which provides positive shutoff to hold the fuel salt in
the primary circulation system, yet which can be
opened fairly quickly to allow the salt to flow into the
drain tank, be of the freeze type used successfully in
the MSRE. The MSRE “valve™ consisted of a flattened
section of the 2-in. drain line provided with external
heaters and coolers. It is to be noted, however, that a

159

single drain line for the MSBR would be 6 in. in
diameter, and since the ability to freeze a pipe decreases
rapidly with size, it poses a markedly different problem.

The direction the development of an MSBR freeze
valve will take is not known at this time. One possibility
is that it will have the appearance of a small shell-and-
tube heat exchanger with the salt flowing through the
tubes, A mechanical-type valve with the seat chilled to
provide positive shutoff may also be considered.
Development of a suitable positive shutoff device
appears generally within present technology and is not a
major uncertainty in the MSBR design.

16.3.6 Gaseous Fission Product Removal System

Fission product gases will be purged from the
circulating fuel salt by introducing helium bubbles in a
side stream and subsequently stripping the gas from the
system. The bubble generator and bubble separator,
described in Sects. 3.9.2 and 3.9.3, have been tested on a
small scale in water, and the concept appears to involve
few major uncertainties. Development of larger size
equipment and testing in salt will be required, however.

16.3.7 Off-Gas System

An off-gas system is proposed for cleaning up the
helium purge gas so that it can be recycled, for holding
up the xenon and krypton to allow decay, for gathering
the fission product particulates, and for trapping the
tritium, Means will have to be provided for disposal of
the collected radioactive materials, Although the MSRE
provided considerable background of experience, addi-
tional development will be needed for the components
in the MSBR off-gas system. The charcoal traps, helium
compressors, particle traps, etc., must be effectively
cooled to remove decay heat. All areas appear amenable
to further study and development, however.

The conceptual design proposes that the radioactive
gaseous wastes from an MSBR be collected in gas
cylinders for long-term storage and decay. Whether the
bottles are stored at the MSBR plant site or at other
sites, approved equipment and procedures must be
developed for handling them.

16.3.8 Steam Generators

Although there is no specific operating experience
with a once-through salt-heated steam generator of the
type proposed for the MSBR, experience with similar
heat transport fluids and with steam generators de-
veloped for other reactor types leads to the conclusion
that design of the MSBR units is within present
technology. A plan for industrial study and develop-
ment of a molten-salt steam generator has been initiated
by ORNL.

If the coolant salt accidentally mixes with the steam,
there are no exothermic reactions, although a blowout
disk will be provided tc relieve pressure buildup in the
coolant-salt circuits.

The lowest allowable feedwater temperature for the
steam generator remains to be determined experi-
mentally. The 700°F value assumed in this design study
probably can be lowered without causing excessive
freezing of coolant salt in the steam generators,

The steam generator tubing must be compatible with
the high-pressure, high-temperature steam on the inside
of the tubes and with the coolant salt on the outside.
As discussed in Sect. 16.2.3, the compatibility of
Hastelloy N with sodium fluoroborate salt is excellent,
provided that water is excluded from the secondary
system. The compatibility of the metal with steam also
appears excellent, but testing is not yet complete. In
the unlikely event that the results are unfavorable,
duplex tubing having a proven steam-side material, such
as Incoloy 800,*' could be used.

16.3.9 Instrumentation and Controls

Section 10.5 outlined the development problems
associated with the components in the instrumentation
and controls system that must be located in the high
ambient temperatures of the reactor and drain tank
cells. Wiring, connectors, and cell wall penetrations will
require special treatment, and the nuclear detectors
were mentioned as particular problems. While the
specific measures to be taken are uncertain in many
instances, none are judged too severe for reasonable
solution, A “fall back™ position for many of the
components is to install them in cooled compartments
within the reactor cell.

The stability of the control system during transients
and the procedures for startup, standby, and shutdown
have received only preliminary study. While the need
for detailed investigation is apparent in many areas,
none have been singled out to date as presenting a
major problem.

If reheat is employed, as proposed in the reference
design, the coolant flow will need to be proportioned
between the steam generators and the reheaters to
achieve the required exit steam temperatures. Valves for
salt service have received relatively little development.
Since the requirement is for proportioning rather than
positive shutoff, however, development of a mechan-
ical-type valve such as those already in use on salt loops

160

appears to be within present technology. A fluidic-type
valve may have promise. If valves prove impractical,
separate variable-speed coolant-salt pumps can be used.

16.3.10 Piping and Equipment Supports

The piping flexibility analysis for the reference design
was made on the basis that the reactor vessel is
anchored and that the heat exchangers and pumps can
move with the only restraint being the vertical hangers.
However, the flexibility of the system must be con-
trolled during an earthquake or after an accidental
break to prevent whipping or other excessive movement
of the piping. Conventional hydraulic dashpots used to
dampen rapid movements would not be usable because
of the high temperature in the reactor cell. Dashpots
will need to be developed which use gases, molten salts,
or pellet beds as the working medium, or cooling
systems for the conventional dashpots will need to be
devised. The manufacturers of this type of equipment
have not been consulted to date because this detail of
the design has not appeared to be one of the major
uncertainties.

The conceptual design calls for the major equipment
to be suspended from the cell roof structure. The
supports have not been designed, but the uncertainties
do not appear to be major ones. A detailed seismic
analysis needs to be made of the entire reactor plant.

16.3.11 Cell Construction

The cell wall construction proposed in the reference
design represents just one possible arrangement for
satisfying the requirements of protecting the concrete
biological shielding from excessive temperature and
radiation damage while at the same time providing
thermal insulation and a double-walled containment
that can be leak tested and monitored. Subsequent
studies have indicated that the reference design may be
overcautious in this respect. Use of electric resistance
heating elements for bringing the cells up to the high
operating temperatures also may not be the most
efficient arrangement. In general, these design aspects
represent optimization questions rather than major
uncertainties.

16.4 TRITIUM CONFINEMENT

Tritium production and distribution in the MSBR
were discussed in Sect. 3.3.7. There is little uncertainty
in the calculated rate of production of 2400 Ci/day, an
amount that is far more than could be permitted to
escape to the plant surroundings. It is not clear at this
time, however, just how much would escape from the
reference design MSBR, how much the release rate must
be reduced to be tolerable, and what is the best way to
modify the systems to effect the reduction.

Even in the reference design, which contains no
special provisions for tritium confinement, the esti-
mated concentration in the condenser cooling water
leaving the plant would be below the current MPC for
release to uncontrolled areas (see Sect. 3.3.7). It will
certainly be required, however, that the release rate be
reduced as far as practicable. Added to the “minimum
practicable” criterion will be the compelling require-
ment that the tritium release from an MSBR not be so
great as to offset other advantages that the concept may
have. This latter requirement probably means that the
tritium release rate from an MSBR be less than 1% of
the production rate.

There are several ways currently under consideration
for holding the tritium release rate to below the value
calculated for the reference design. Until the results of
various measurements and tests now under way become
available, however, a decision as to what special tritium
confinement modifications should be incorporated in
the MSBR cannot be made. Some of the measures being
studied are discussed below,

Gas sparging of the fuel salt reduces the amount of
tritium diffusing into the coolant salt. The sparging is
probably more effective than was described in Sect.
3.7.7 because conseryatively high values for the tritium
solubility were assumed in the calculations. Increasing
the helium sparging rate and reducing the U** to U™
ratio would take out more tritium with the primary
system off-gas. Lowering the U* to U* ratio, however,
would tend to increase corrosion, although perhaps not
seriously. In any event, taking these measures in the
primary system may not reduce the tritium release rate
as much as will be required.

It appears that injection of 1 to 10 cc/sec of HF into
the coolant salt would be quite effective in reducing the
amount of tritium that could transfer into the steam
system. The major uncertainty is the fraction of
hydrogen fluoride (or tritium fluoride) that would react
with the metal wall. The loss of the metal would be
tolerable, but the reaction could release atomic tritium
that would diffuse through the wall. If the fraction of
tritium fluoride which reacts with the metal walls is
small, most of the tritium could be taken out by the
coolant-salt off-gas system.

Reaction of tritium with trace constituents in the
coolant salt is being explored. Consideration has also
been given to changing the heat transport fluid to one
that would positively trap tritium. As explained in Sect.

161

16.2.2, however, no other liquid is now known that
would do this and also be compatible with the fuel salt,
Gas coolants that would trap tritium have disadvantages
that discourage their use.

In principle, the most straightforward way of re-
ducing tritium transfer to the steam would be to use
heat exchanger tubes less permeable to tritium. Few
metals that can be considered, however, are much less
permeable than Hastelloy N, with perhaps the excep-
tion of tungsten and molybdenum. Although use of
tubes coated with these metals would introduce tech-
nical difficulties and higher costs, perhaps they should
not be dismissed out of hand. The same might be said
of glass coatings. An oxide layer on the steam generator
tubes would increase their resistance to tritium penetra-
tion, but additional data are needed to evaluate the
effectiveness of such a coating.

A sealed steam system has been considered, but
tritium would concentrate in it, and the leakage would
have to be held to extremely low levels. This method
appears unattractively complicated and expensive.

One method of blocking tritium transport to the
steam would be to interpose another circulating heat
transport loop between the secondary salt and the
steamn generators. This additional system would use a
fluid, such as Hitec, that would positively trap the
tritium. Hitec is a commercially available, widely used
heat transfer salt with the composition KNO;-NaNO,-
NaNO; (44-49-7 mole %) that would chemically react
with the tritium. (If the additional loop is used, an
interesting possibility is to use ’LiF-BeF, as the
secondary salt to transport heat from the primary heat
exchanger to the Hitec, although, as mentioned previ-
ously, the relatively high cost of 7Li would have to be
taken into consideration.) The Hitec would be circu-
lated through the steam generators and reheaters. The
cost of the extra salt system would be partially offset
by the fact that the Hitec would allow use of less
expensive materials in the steam equipment, and its
relatively low liquidus temperature of 288°F would
eliminate the need to preheat the feedwater to 700°F
and the reheat steam to 650°F, One uncertainty,
however, is the maximum temperature at which the
Hitec can be operated. It might be necessary to drop
the steam temperature to the turbine to 900°F if the
Hitec system were used to solve the tritium problem.

In summary, several different methods for reducing
the estimated tritium release from the reference design
MSBR are currently receiving serious study, and there is
reason to expect that acceptable rates can be attained
without serious economic penalty. Certainly, use of an
additional heat transport loop would practically elimi-
nate diffusion of the tritium into the steam system.
16.5 CHEMICAL PROCESSING SYSTEM

An essential requirement for breeding with thermal
neutrons in a molten-salt reactor is the rapid processing
of the fluid fuel to remove fission products and
protactinium. Xenon and Krypton can be remioved by a
physical separation process (as described in Sect. 3.9),
but the isolation of protactinium and the removal of
rare earths require that the fuel salt be chemically
processed, Neutron losses to rare earths are acceptably
low if their removal cycle is on the order of a month or
s0, but the cycle time for protactinium isolation needs
to be on the order of a few days (see Fig. 16.1).

The chemical processing must be (1) fundamentally
sound, (2) practical, and (3) economical if the MSBR is
to be successful. On the first point there is little room
for doubt. There are several chemical processes having
equilibria and rates which are well known and favorable
for MSBR application. The ones proposed in this MSBR
reference design include fluorination, hydrofluori-
nation, and various exchange reactions between fuel salt
and liquid bismuth and between bismuth and other
salts, such as lithium chloride. There are sufficient data
at hand to assure that these processes are chemically
sound.

There is less assurance of the practicability of the
continuous processing system described in Sect. 8, Most
of the operations involved have to date been carried out
only in small-scale experiments. Development of com-
ponents and instrumentation is in the earliest stages.
Although the results to date have disclosed no insur-
mountable obstacles, several problem areas have been
identified and are discussed below.

1 Pa PROCESSING CYCLE (doys)

=0 L4 a

T-
5 (©33pal/z(®33) : D37

'

N BREEDING RATIC

LOSE

o2

10"
FRAGTION OF 23%pa REMOVED PER DAY

Fig. 16.1. Effect of 233 Pa capture vs processing rate and fuel
specific power.

162

The most basic problem is that of materials for
equipment which is exposed to both bismuth and salt,
As explained in Sect. 8, molybdenum has quite satis-
factory corrosion resistance and appears to be the best
overall choice despite the unusual problems of designing
and fabricating joints in this metal., Development has
progressed far enough to give reasonable assurance that
these problems can be overcome and the required
equipment can be built. The fabrication costs for
molybdenum systems are still uncertain but are sure to
be high. Thus there is an economic incentive for
eliminating the need for molybdenum, either by chang-
ing to another process or by developing an alternative
material (possibly graphite).

The use of bismuth in the salt processing requires
dependable measures to prevent accidental gross or
chronic small carryover of bismuth in the salt returning
to the reactor. A cleanup device for removing bismuth
exists only in concept. Information is needed both on
the performance of such a device and on the tolerance
limits for bismuth in the fuel salt. If dependable,
adequate cleanup of the returning salt should prove to
be impracticable, it would be necessary to make
substantial changes from the process described in this
report,

Corrosion protection in the fluorinator requires a
layer of frozen salt on the wall. Small-scale develap-
ment tests indicate that the requisite frozen layer can
be established, although reliable control of the thick-
ness may be difficult. Therefore occasional loss of the
frozen wall must be anticipated. If the vessel is made of
nickel, formation of an adherent NiF, layer would be
expected to limit corrosion, so that occasional failure of
the frozen-wall protection (on the order of once a week
to once a month) could be tolerated. Therefore the
frozen-wall fluorinator should be practical to build and
operate.

The varying, sometimes intense, sources of decay heat
due to the concentrated protactinium and fission
products in the processing plant will require carefully
designed cooling systems. In the reference design,
however, the radioactive materials are always in solu-
tion, so there is little or no chance of local hot spots
due to heat-generating sediments. Design of a satis-
factory cooling system should therefore be feasible.

The performance of the MSBR as a breeder is
sensitive to uranium losses in the chemical processing
plant. Although there has been no pilot plant operation
to measure losses in a system like this, some reasonable
judgment is possible. The probable losses are not
directly related to the throughput of salt or uranium,
since nowhere in the process does there appear to be
the potential for gradual, irrecoverable buildup of a
significant fraction of the uranium passing through.
Instead, one must consider the various materials leaving
the plant and estimate how much uranium (or protac-
tininm) might be carried out with them. The flowsheet
(Fig. 2.4) shows three small discard streams: salt from
the Pa decay system, Bi-Li carrying the divalent rare
earths, and Bi-Li carrying the trivalent rare earths. The
amount of uranium in the Bi-Li discard streams should
be negligible, but if this were not the case, the uranium
could be recovered rather simply by hydrofluorinating
the Bi-Li in the presence of salt from the Pa decay
system. The salt discarded from the Pa system will be
fluorinated to recover uranium in a batch operation
almost identical to that carried out successfully in the
MSRE.'?5 The MSRE experience indicates that the
MSBR losses in the discarded salt should be on the
order of 0.2 kg of U per year. Probably more
significant, and certainly more difficult to predict, are
the amounts of uranium that will be discarded in other
ways, such as the replacement of NaF absorbers,
bismuth cleanup elements, salt filters, and miscel-
laneous pieces of equipment.

The chemical processing plant is designed to con-
tinuously treat a side stream of the fuel salt and return
it to the reactor circulating loop. The chemical plant
and the reactor plant are essentially independent, so
that malfunctions in one would not necessarily affect
the other, Chemical plant operation can be interrupted
for several days with only minor effects on reactivity
and nuclear performance, but if there were a prolonged
shutdown of the processing plant, neutron losses to
protactinium would cause the production of 2?3U to
fall below the consumption rate. The reactor would still
perform as a high-gain converter, which, if need be,
could be kept running for several years without
chemical processing by adding fissile material (as UFg
or PuF3) through simple equipment that must be
provided for this contingency. Specific information on
the dependability of the processing system will not be
available until pilot plants are operated.

Although the reactivity effects of perturbations in the
chemical plant operation are ecasily manageable, it is
essential that they be understood. In the MSBR it will
be necessary to distinguish any truly anomalous effects
that might occur from the effects of changing concen-
trations of neutron poisons and fissile material in the
circulating fuel due to on-line processing. The results of
a complete interruption of salt flow between the
reactor and the chemical processing plant are simple (in
principle) and could be calculated by an on-line
computer. The possibility of variations in the compo-

163

sition of the salt continuously flowing back into the
reactor requires that concentrations and inventories in
the processing plant be measured. The chemical analysis
procedures now in use are accurate but are slow.
On-line analytical techniques that can provide direct
inputs to a computer are needed. If these prove very
difficult to develop, an alternative would be to ease the
analytical demands by interposing parallel holdup tanks
between the processing plant and the reactor, so that
batches of processed salt could be sampled and ana-
lyzed before being pumped back into the fuel system.

With regard to the third requisite for the chemical
plant, that it be economical, the capital and operating
costs for the MSBR chemical processing have not been
estimated for the currently proposed system. The
concept described in this report was adopted because it
promised to be less expensive to construct than
previous concepts, but as of this writing, detailed
flowsheets and equipment concepts upon which to base
cost estimates have not been completed. The cost
uncertainties are therefore quite large. Conceivably, the
costs could be high enough to make breeding in a
molten-salt breeder reactor (as described in this report)
economically unattractive, In this case it would be
possible to produce lower-cost power by operating the
reactor as a high-gain converter with a much simpler
chemical processing system. The ultimate goal of an
economical breeder could be realized later when a
lower-cost processing system became available.

There are alternative processes that may have tech-
nical or economic advantages, but only preliminary
investigations of basic feasibility have been made.
Perhaps the foremost of these at present is the oxide
precipitation process, which exploits the differences in
oxidation potential required to form Pa, O, UO,, and
other oxides. This process would hopefully have lower
capital and operating costs than the fluorination—
reductive-extraction system described in this report.

16.6 FISSION PRODUCT BEHAVIOR

How the fission product particulates will distribute
themselves in an MSBR is still not known with
certainty. Accumulations of the products are of pri-
mary concern because of the possibility of localized
high temperature due to decay heat. Portions of the
reference design thought to be likely deposition sites
have been provided with special cooling systems. After
the fission product distribution has been determined
with more certainty, possibly by operation of a
prototype MSBR, the cooling systems in future designs
will be modified as required.
The distribution of fission products is also of interest
because the noble metals would have an effect on the
breeding ratio if they concentrated in the core graphite.
Fission product behavior was studied in the MSRE in
some detail, ' however, and although it was found that
the nohle metals depogited on surfaces, it was 2also
evident that they deposited more heavily on the
Hastelloy than on the graphite. If the examined
specimens were representative of the MSRE core, about
7% of the ?5Nb and from 2 to 5% of the other noble
metals were on the core graphite. If it is assumed that
10% of the nuclides with potentially the greatest
poisoning effect remain in the core region of the MSBR,
the effect is still not as great as the credit which could
be taken for the burnout of ' ° Binitially present in the
graphite (see Perry and Bauman, ref. 10, pp. 208—-219).
Thus, even though it was assumed in the MSBR
performance estimates that no noble metals were
deposited on the graphite, by taking no credit for boron
burnout, the estimates are conservative.

16.7 STEAM CONDITIONS IN THE
THERMAL-POWER CYCLE

Mention has been made elsewhere in this section of
lowering the top temperatures in various systems to
mitigate uncertainties regarding some of the material
properties. Lowering the temperatures does not neces-
sarily mean that the steam temperature in the heat-
power system must also be reduced. Since this is a
possible effect, however, there is interest in what the
impact would be on MSBR performance.

If the steam system conditions were modified from
the reference design conditions of 3500 psia
1000°F/1000°F to 3500 psia 900°F/900°F,* the ther-
mal efficiency of the cycle would be reduced from
about 44.4 to 42.0%.

For a thermal efficiency of 42% the thermal capacity
of the reactor plant would have to be about 2400
MW(t) rather than the 2250 MW(t) used in the
conceptual study. If one assumes that capital costs and
fuel costs are directly proportional to the thermal
capacity, the estimated power production cost with the
lower efficiency is about 5.7 mills/kWhr as compared
with 5.4 mills/kWhr with 44.4% efficiency. In a large

*Contrary to what would be expected in conventional cycles,
the 700°F feedwater requirement in the MSBR justifies the use
of supercritical-pressure steam even at 900°F. If a 2400-psig
900°F/900°F cycle were used and a Loeffler cycle were
employed 1o obtain the 700°F feedwater, the efficiency of the
cycle would be about 39,5%.

164

molten-salt reactor, however, the capitalization and fuel
costs would not increase linearly with capacity, and the
effect of lowering the top temperature by 100°F is not
likely to be an overriding consideration.

If an MSBR station must use wet natural-draft cooling
towers for the condenser cooling water supply rather
than the once-through freshwater source assumed in the
reference design, the back pressure on the turbine
would be increased to about 2% in. Hg abs and the heat
rate raised to about 7800 Btu/kWhr. The capital cost of
the MSBR station would be increased by about $5
million, and the power production cost would increase
by about 0.13 mill/kWhr, as explained in Table D.17.
These incremental increases due to use of a cooling
tower are substantially less than the impact of use of
towers in the lower-efficiency light-water nuclear sta-
tions, also as shown in Table D.17,

16.8 MAINTENANCE EQUIPMENT AND
PROCEDURES

The MSRE provided valuable experience in the use of
remotely operated tools and viewing equipment for
maintaining a molten-salt reactor. The MSBR require-
ments for maintenance and inspection (other than core
graphite replacement) can probably be met within the
bounds of reasonable development.

Replacement of the reactor core graphite, however,
involves the handling of a large and intensely radio-
active piece of equipment. Although the frequency with
which the maintenance is required will encourage
development of detailed procedures and special tools
and equipment, the magnitude of the task and the
potential hazards involved should not be minimized. As
with fuel handling in a solid-fuel reactor, it is an
undesirable feature that the owner of a molten-salt
breeder reactor may have to accept.

A feasibility study was made (see Sect. 13) of the
maintenance equipment and procedures needed for an
MSBR. The major uncertainties are whether the $4.5
million allowance included in the cost estimate for
maintenance equipment is adequate and whether the
required plant downtime for graphite replacement can
be accommodated within the 80% plant factor. With
regard to the latter, the four-year useful life of the
graphite in the MSBR reference design roughly corre-
sponds to the required interval between major steam
turbine overhauls, and there is reason to believe that the
graphite replacement could be accomplished without
adding significantly to the downtime now experienced
in most plants.
16.9 SAFETY STUDIES

A comprehensive safety study has not been made of
an MSBR power station. The conceptual design is
believed to be conservative in th= containment provided
for radioactive materials during normal operation, but

165

detailed safety studies may disclose structural or opera-
tional features that will dictate design modifications,
These changes are not expected to pose particularly
difficult technical problems, but they could add to the
capital cost.
References

1. MSR Program Semiann. Progr. Rept. Aug. 31,
1969, ORNL-4449,

2. MSR Program Semiann. Progr. Rept. Aug. 31,
1968, ORNL-4344.

3. MSR Program Semiann. Progr. Rept. Feb. 29,
1968, ORNL-4254.

4. Paul R, Kasten, E. S. Bettis, and Roy C. Robert-
son, Design Studies of 1000-Mw(e) Molten-Salt Breeder
Reactors, ORNL-3996 (August 1966).

5. MSR Program Semiann. Progr. Rept. Aug. 31,
1966, ORNL-4037.

6. MSR Program Semiann. Progr. Rept. Feb. 28,
1967, ORNL-4119.

7. MSR Program Semiann. Progr. Rept. Aug. 31,
1967, ORNL-4191.

8. Roy C. Robertson et al.,, Two-Fluid Molten-Sult
Breeder Reactor Design Study, ORNL-4528 (August
1970).

9. MSR Program Semiann. Progr. Rept. Feb. 28,
1969, ORNL-4396.

10. W. R. Grimes, “Molten-Salt Reactor Chemistry,”
Nucl. Appl. Technol. 8(2), 13755 (February 1970).

11. MSR Program Semiann. Progr. Rept. Feb. 28,
1970, ORNL-4548.

12. H. A. McLain, MSBR Primary Salt Pressure
Drops, ORNL internal correspondence MSR-69-34
(Apr. 21, 1969).

13. H. E. McCoy et al.,, “New Developments in
Materials for Molten-Salt Reactors,” Nucl Appl
Technol. 8(2), 156—69 (February 1970).

14. R. E. Thoma et al., “Phase Equilibria in the
System BeF,-ThF, and in LiF-BeF,-ThF;,” J. Phys.
Chem. 64, 865 (1960).

15. H. F. McDuffie et al., Assessment of Molten Salts
as Intermediate Coolants for LMFBR's, ORNL-TM-
2696 (Sept. 3, 1969).

16. R. W. Henson, A. J. Perks, and J. H. W. Simmons,
Lattice Parameter and Dimensional Changes in Graphite
Irradiated between 300 and 1350°C, AERE-R-5489
(1967).

166

17. J. W. Helm, “Long Term lrradiation Effects on
Graphite,” paper MI-77, 8th Biennial Conference on
Carbon, Buffalo, New York (June 1967).

18. P. R, Kasten et al., Graphite Behavior and [ts
Effect on MSBR Performance, ORNL-TM-2136
(February 1969).

19. W. P. Eatherly et al., Technical Analysis and
Program Proposal: Graphite for Molten Salt Reactors,
ORNL internal correspondence CF-68-11-18 (Novem-
ber 1968).

20. D. Scott and W. P. Eatherly, “Graphite and
Xenon Behavior and Their Influence on Molten-Salt
Reactor Design,” Nucl. Appl. 8(2), 179—89 (February
1970).

21. J. Chang, W. P. Eatherly, and J. W. Prados, MSR
Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL-
4396, pp. 229-31.

22. D. V. Kiplinger and R. L. Beatty, MSR Program
Semiann. Progr. Rept. Aug. 31, 1968, ORNL-4344, pp.
230--31.

23. GCR Program Semiann. Progr. Rept. June 1967,
ORNL-4170, pp. 203-4.

24. J. H. W, Simmons, A Relation between Thermal
Expansion and Dimensional Change for Polycrystalline
Graphite, AERE-R-3883 (1961).

25. J. C. Bokros and A. S. Schwartz, “A Model to
Describe Neutron-Induced Dimensional Changes in
Pyrolytic Carbon,” Carbon 5, 481 (1967).

26. W. D. Manly et al., “Metallurgical Problems in
Molten Fluoride Systems,” Prog. Nucl. Energy, Ser. 4,
2, 16479 (1960).

27. R. W. Swindeman, The Mechanical Properties of
INOR-8, ORNL-2780 (January 1961).

28, I. T. Venard, Tensile and Creep Properties of
INOR-8 for the Molten-Salt Reactor Experiment,
ORNL-TM-1017 (February 1965).

29. H. E. McCoy and J, R, Weir, Materials Develop-
ment for Molten-Salt Breeder Reactors, ORNL-TM-
1854 (June 1967).
30. Interpretations of ASME Boiler and Pressure
Vessel Code, Case 1315.3, Am. Soc. of Mech. Engrs.,
New York (Apr. 25, 1968).

31. Interpretations of ASME Boiler and Pressure
Vessel Code, Case 1345.1, Am. Soc. of Mech. Engrs.,
New York (Mar. 14, 1966).

32. H. E. McCoy, Jr., An Evaluation of the Malten-
Salt Reactor Experiment Hastelloy-N Surveillance Spec-
imens - First Group, ORNL-TM-1997 (November
1967).

33. H. E. McCoy, Ir., An Evaluation of the Molten-
Salt Reactor Experiment Hastelloy-N Surveillance Spec-
imens — Second Group, ORNL-TM-2359 (February
1969).

34. H. E. McCoy, Ir., An Evaluation of the Molten-
Salt Reactor Experiment Hastelloy-N Surveillance Spec-
imens — Third Group, ORNL-TM-2647 (1969).

35. 1. O. Stiegler and E. E. Bloom, “The Effects of
Large Fast-Neutron Fluences on the Structure of
Stainless Steel,”” J. Nuel. Mater. (to be published).

36. H. E. McCoy, Jr., and J. R. Weir, Jr., “Stress-
Rupture Properties of Irradiated and Unirradiated
Hastelloy-N Tubes,” Nucl Appl. 4(2), 96—104 (Feb-
ruary 1968).

37. H. E. McCoy, Jr., Influence of Titanium, Zirco-
nium, and Hafnium Additions on the Resistance of
Modified Hastelloy N to lrradiation Damage — Phase I,
ORNL-TM-3064 (January 1971).

38. C. E. Sessions and T. S. Lundy, Diffusion of
Titanium -in Modified Hastelloy-N, ORNL-TM-2392
(January 1969).

39. J. H. Devan, Effect of Alloying Additions on
Corrosion Behavior of Nickel-Molybdenum Alloys in
Fused Fluoride Mixtures, ORNL-TM-2021, vol. 1 (May
1969).

40. J. W. Koger and A. P. Litman, Compatibility of
Fused Sodium Fluoroborates and BF'y Gas with Hastel-
loy-N Alloys, ORNL-TM-2978 (June 1970).

41. MSR Program Semiann. Progr. Rept. Aug. 31,
1970, ORNL-4622.

42. N. M. Greene and C. W. Craven, Jr., XSDRN: A
Discrete Ordinates Spectral Averaging Code, ORNL-
TM-2500 (July 1969).

43. P. N. Haubenreich, Tritium in the MSRE: Calcu-
lated Production Rates and Observed Amounts, ORNL
internal correspondence CF-70-2-7 (Feb. 4, 1970).

44. R. B. Briggs, Calculation of the Tritium Distribu-
tion in the MSRE, ORNL internal correspondence
CF-70-7-13 (August 1970).

45. H. T. Kerr and A. M. Perry, Tritium Production
in MSBR's, ORNL internal correspondence MSR-
69-116 (Dec. 3, 1969).

167

46. USAEC, “Standards for Protection Against Radi-
ation,” Title 10 — Code of Federal Regulations, part 20
(1970).

47. W. R. Bush, Review of Chalk River Experience
with Tritiated Heavy Water, AECL-2756 (July 1967).

48. V. P. Bond, “Evaluation of Potential Hazards
from Tritium Water,” paper IAEA SM 146/13, IAEA
Symposium on Environmental Aspects of Nuclear
Power Stations, New York (Aug. 10—14, 1970).

49. W. Eifler and R. Nijsing, Fundamental Studies of
Fluid Flow and Heat Transfer in Fuel Element Geom-
etries. Il. Experimental Investigation of Velocity Dis-
tributions in a Triangular Array of Parallel Rods,
EURATOM report EUR 2193.6, Joint Nuclear Re-
search Center, Ispra Establishment, Italy (1965).

50. R. R. Liguori and J. W. Stephenson, The Heating
Program, Astra, Inc., Raleigh, N.C. (January 1961).

51. S. J. Chang, C. E. Pugh, and S. E. Moore,
Viscoelastic Analysis of Graphite under Neutron Irradi-
ation and Temperature Distribution, ORNL-TM-2407
(October 1969).

52, B. Cox, Preliminary Heat Transfer Results with a
Molten Salt Containing LiF-BeF,-ThF,-UF,; Flowing
inside a Smooth Horvizontal Tube, ORNL internal
correspondence CF-69-9-44 (September 1969).

53, D. Burgreen et al., “Vibration of Rod Induced by
Water in Parallel Flow,” Trans. ASME (July 1958).

54. Y. N. Chen, “Flow-Induced Vibration and Noise
in Tube-Bank Heat Exchanger Due to von Karman
Streets,” ASME paper 67-VIBR-48 (January 1967).

55. A. Kalnins, Static, Free Vibration, Instability
Analysis of Thin, Elastic Shells of Revolution, AFF-
DL-TR-68-144, Air Force Flight Dynamics Laboratory,
Air Force Systems Command, Wright-Patterson Air
Force Base (March 1969).

56. ASME Boiler and Pressure Vessel Code, Section
IIl, Nuclear Vessels, Am. Soc. of Mech. Engrs., New
York, 1968 ed.

57. ASME' Boiler and Pressure Vessel Code, Section
VIII, Pressure Vessels, Division 1, Am. Soc. of Mech.
Engrs., New York, 1968 ed.

58. Interpretations of ASME Boiler and Pressure
Vessel Code, Case 1331.4, Am. Soc. of Mech. Engrs.,
New York, Aug. 15, 1967.

59. Piping Flexibility Analysis Program MEL-21, San
Francisco Bay Naval Shipyard, Mare Island Division,
Vallejo, Calif.

60. ASME Nuclear Power Piping, USAS B31.7, Am.
Soc. of Mech. Engrs., New York, tentative, 1969.

61. H. 1. Sexton, Seismic Review of the Molten-Salt
Breeder Reactor Design Concept, H. J. Sexton and
Associates, Engineers, San Francisco, Calif.
62. C. E. Bettis et al., Design Study of a Heat
Exchanger System for One MSBR Concept, ORNL-TM-
1545 (September 1967).

63. C. E. Bettis et al.,, Computer Programs for the
Design of MSBR Hear Exchangers, ORNL-TM-2815
(1970).

64. C. H. Gabbard, Reactor Power Measurement and
Heat Transfer Performance in the Molten-Salt Reactor
Experiment, ORNL-TM-3002 (May 1970).

65. B. Cox, Preliminary Heat Transfer Results with a
Molten Salt Mixture Containing LiF-BeF,-ThF,-UF,
Flowing inside a Smooth, Horizontal Tube, ORNL
internal correspondence CF-69-9-44 (Sept. 25, 1969).

66. E. N. Sieder and G, E. Tate, “Heat Transfer and
Pressure Drop of Liquids in Tubes,” Ind. Eng. Chem.
28(12), 142935 (1936).

67. H. W. Hoffman and S. L. Cohen, Fused Salt Heat
Transfer — Part III: Forced Convection Heat Transfer in
Circular Tubes Containing the Salt Mixture NaNQ,-
NaNO,-KNO, , ORNL-2433 (March 1960).

68. H. A. McLain, Revised Correlations for the
MSBR Primary Salt Heat Transfer Coefficient, ORNL
internal correspondence MSR-69-89 (Sept. 24, 1969).

69. O. P. Bergelin, G. A. Brown, and A. P. Colburn,
“Heat Transfer and Fluid Friction During Flow across
Bank of Tubes — V: A Study of a Cylindrical Baffled
Exchanger without Internal Leakage,” Trans. ASME 76,
841-50(1954).

70. D. A. Donahue, “Heat Transfer and Pressure
Drop in Heat Exchangers,” Ind. Eng. Chem, 41(11),
2499-2511 (November 1949).

71. C, G. Lawson, R, J. Kedl, and R. E. McDonald,
“Enhanced Heat Transfer Tube for Horizontal Con-
denser with Possible Application in Nuclear Power Plant
Design,” Am. Nucl. Soc. Trans. 9(2), 565—66 (1966).

72. H. A. McLain, Revised Primary Salt Heat Trans-
Jfer Coefficient for MSBR Primary Heat Exchanger
Design, ORNL internal correspondence MSR-67-70
(July 31, 1969).

73. O. P. Bergelin, K. J. Bell, and M. D. Leighton,
“Heat Transfer and Fluid Friction During Flow across
Banks of Tubes — VI: The Effect of Internal Leakage
within Segmentally Baffled Exchangers,” Trans. ASME
80, 53-60(1958).

74. P. G. Smith, Development of Fuel- and Coolant-
Salt Centrifugal Pumps for the Molten-Salt Reactor
Experiment, ORNL-TM-2987 (October 1970).

75. 1. A. Hafford, Development of the Pipeline Gas
Separator, ORNL-1602 (February 1954).

76, H. S, Swenson, C, R. Kakarala, and J. A. Carver,
“Heat Transfer to Supercritical Water in Smooth-Bore

168

Tubes,” Trans. ASME, Ser C: J. Heat Transfer 87(4),
477—84 (November 1965).

77. Chemical Engineer's Handbook, 3d ed., McGraw-
Hill, New York, 1950.

78. J. H. Keenan and F, G. Keyes, Thermodynamic
Properties of Steain, Wiley, New Yoik, 19306.

79. E. S. Nowak and R. J. Grosh, An Investigation of
Certains Thermodynamic and Transport Properties of
Water and Water Vapor in the Critical Region, ANL-
6064 (October 1959).

80. K. Goldmann, S. L. Israel, and D. J. Nolan, Final
Status Report: Performance Evaluation of Heat Ex-
changers for Sodium-Cooled Reactors, UNC-5236,
United Nuclear Corporation, Elmsford, N.Y. (June
1969).

81. L. S, Tong, Boiling Heat Transfer and Two-Phase
Flow, Wiley, New York, 1965.

82. L. Y. Krasyakova and B. N. Glusker, “Hydraulic
Study of Three-Pass Panels with Bottom Inlet Headers
for Once-Through Boilers,” Teploenergetika 12(8),
17-23 (1965) (UDC 532:621.181.91.001.5).

83. E. R. Quandt, “Analysis and Measurement of
Flow Oscillations,” Chem. Eng. Prog. Symp. Ser.
57(32)(1961).

84. L. M. Shotkin, “Stability Considerations in Two-
Phase Flow,” Nucl. Sci. Eng. 28, 31724 (1967).

85. J. F. Myers and L. E. Wood, “Enhancing Ac-
curacy of Rupture Disks,” Chem. Eng. (Nov. 8, 1965).

86. The Bull Run Steam Plant, Tenn. Valley Author-
ity, Knoxville, Tenn., Tech. Report No. 38 (November
1967).

87. Roy C. Raobertson, MSBR Steam System Per-
formance Calculations, ORNL internal correspondence
MSR-66-18 (July 5, 1966).

88, Roy C. Robertson, Survey of Steam Conditions
for 1000 Mw(e) MSBR, ORNL internal correspondence
CF-69-10-34 (October 1969).

89. I. M. Keyfitz, 1000 Mwe Fast Breeder Reactor
Follow-On Study, Moisture Separation or Steam Reheat
vs Sodium Reheat Plant Cycle, Technical and Economic
Evaluation, WAPD-2000-20, reissued January 1969,

90, B. E. Short, “Flow Geometry and Heat Ex-
changer Performance,” Chem. Eng. Progr. 61(7), 6370
(July 1965).

91. Theodore Rockwell IIl, ed., Reacror Shielding
Design Manual, Van Nostrand, Princeton, N.J., 1956,

92. W. K. Furlong, “Afterheat Removal in Molten-
Salt Reactors,” Am. Soc. of Mech. Engrs., Winter
Annual Meeting, paper ASME-69-WA/NE-19 (Novem:-
ber 1969).

93. O. T. Zimmerman and Irvin Lavine, “Air-Cooled
Heat Exchangers,” Cost Eng. (July 1960).
94, MSR Program Semiann. Progr. Rep:. July 31,
1964, ORNL-3708.

95. Roy C, Robertson, MSRE Design and Operations
Report, Part I, Description of Reactor Design, ORNL-
TM-728 (January 1965).

96. W. E. Browning and C. C. Bolta, Measuremeni
and Analysis of the Holdup of Gas Mixtures by
Charcoal Adsorption Traps, ORNL-2116 (July 1956).

97. R. D. Ackley and W. E, Browning, Jr., Equilib-
rium Adsorption of Kr and Xe on Activated Carbon and
Linde Molecular Sieves, ORNL internal correspondence
CF-61-2-32 (Feb. 14, 1961).

98. W. D. Burch et al., Xenon Control in Fluid Fuel
Reactors, ORNL internal correspondence CF-60-2-2
(July 1960).

99. L. E. McNeese, Engineering Development Studies
for Molten Salt Breeder Reactor Processing No. 2,
ORNL-TM-3137 (in publication).

100. L. E. McNeese, Engineering Development
Studies for Molten Salt Breeder Reactor Processing No.
5, ORNL-TM-3140 (in publication),

101. L. E. McNeese, Engineering Development
Studies for Molten Salt Breeder Reactor Processing No.
3, ORNL-TM-3138 (in publication).

102. L. E. McNeese, Engineering Development
Studies for Molten Salt Breeder Reactor Processing No.,
4, ORNL-TM-3139 (in publication).

103. J. S. Watson and L. E. McNeese, Unit Opera-
tions Quarterly Progress Report July—-September 1968,
ORNL-4366, pp. 5798 (April 1970).

104. W. H. Sides, MSBR Control Studies, ORNL-
TM-2489 (June 2, 1969).

105. H. F. Bauman, Control Requirements for the
MSBR, ORNL internal correspondence MSR-68-107
(July 18, 1968).

106. J. R. Tallackson, MSRE Design and Operations
Report, Part I1I-A: Nuclear and Process Instrumentation,
ORNL-TM-729 (February 1968).

107. 1. B. Ruble and S. H. Hanauer, “A High-
Temperature Fission Chamber,” Proceedings of Nucl.
Eng. and Sci. Conference (Mar. 17, 1958).

169

108. J. C. Robinson and D. N. Fry, Determination of

the Void Fraction in MSRE Using Small Induced
Pressure Perturbations, ORNL-TM-2318 (Feb. 6, 1969).

109. D. N. Fry et al., Measurement of Helium Void
Fraction in MSRE Fuel Salt Using Neutron-Noise
Analysis, ORNL-TM-2315 (Aug. 27, 1968).

110. G. D. Robbins, Electrical Conductivity of Mol-
ten Fluorides, a Review, ORNL-TM-2180,

111. J. R. Tallackson, R. L. Moore, and S. J. Ditto,
Instrumentation and Controls Development for Mol-
ten-Salt Breeder Reactors, ORNL-TM-1856 (May 22,
1967).

112, LMFBR Program Plan, Vol. 4, Instruments and
Controls, WASH-1104 (August 1968),

113. R. L. Moore, Further Discussion of Instrumen-
tation and Controls Development Needed for an MSBR,
ORNL-TM-3303 (to be published).

114. R. B. Briggs, Allowable Rates of Load Change
and Temperature Change for a Large MSBR, ORNL
internal correspondence MSR-70-7 (Jan. 22, 1970).

115. W. H. Sides, Jr., Control Studies of a 1000
Mw(e) MSBR, ORNL-TM-2927 (May 18, 1970).

116. Robert Blumberg and E. C. Hise, MSRE Design
and Operations Report, Part X, Maintenance Equip-
ment and Procedures, ORNL-TM-910 (June 1968).

117. Robert Blumberg, Maintenance Development
for Molten-Salt Breeder Reactors, ORNL-TM-1859
(June 30, 1967).

118, Peter P. Holz, Feasibility Study of Remote
Cutting and Welding for Nuclear Plant Maintenance,
ORNL-TM-2712 (November 1969).

119. Nuclear Utilities Services Corporation, Rock-
ville, Md., Guide for Economic Evaluation of Nuclear
Reactor Plant Designs, prepared for AEC and ORNL,
NUS-531 (TID-4500) (January 1969).

120. United Engineers and Constructors, Inc., Phila-
delphia, Trends Affecting the Rise in Costs of Light
Water Nuclear Energy Plants for Utility Electric Genera-
tion, Summary Report (February 1970). (No number
assigned to date.)

121. H. Susskind and C. J. Raseman, Power Plant
Operating and Maintenance Costs, BNL-50235 (T-572)
(April 1970),

122. Advanced Converter Task Force, AEC Division
RD&T, Evaluation of Advanced Converter Reactors,
Table 6.14, WASH-1087 (April 1969).

123. Jackson and Moreland, and S. M. Stoller Asso-
ciates, Current Status and Future Technical and Eco-
nomic Potential of Light Water Reactors, WASH-1082
(March 1968).

124. L. W. Lang, “Utility Incentives for Implement-
ing Crossed-Progeny Fueling,” Nucl. Appl. Technol. 9,
242 (August 1970),

125. R. B. Lindauer, Processing of the MSRE Flush
and Fuel Salts, ORNL-TM-2578 (August 1969).

126. R. J. Kedl and A. Houtzeel, Development of a
Model for Computing ' ** Xe Migration in the MSRE,
ORNL-4069 (June 1967).

127. R. J. Kedl, 4 Model for Compuring the Migra-
tion of Very Short Lived Noble Guses into MSRE
Graphite, ORNL-TM-1810 (July 1967).

128. F. N. Peebles, Removal of Xenon-135 from
Circulating Fuel Salt of the MSBR by Mass Transfer to
Helium Bubbles, ORNL-TM-2245 (July 1968).

129. L. G. Hauser, “Evaluate Your Cost of Cooling
Steam Turbines,” Elec. Light Power, January 1971, pp.
32-34.
Appendix A

Theory of Noble-Gas Migration

R.J. Kedl

INTRODUCTION

Noble gases, particularly xenon, have an extremely
low solubility in fuel salt, The amount that does
dissolve forms a true solution; that is, there is no
chemical interaction between the noble gas and salt,
This being the case, one would expect xenon and
krypton to migrate from the fuel salt where they are
born to various sinks in accordance with the laws of
mass transfer, This implies that the mass transfer
coefficient controls the migration rate. The sinks will be
comprised of any salt-gas interfaces available to xenon
and krypton, such as circulating bubbles, the voids in
graphite, and the gas space in the pump bowl. Other
sinks are decay and burnup. An analytical model was
developed for the MSRE based on this concept, as
reported by Kedl and Houtzeel.' 2® Another analytical
model, complementary to the above model and specifi-
cally applicable to the very short-ived noble gases, was
reported by Kedl,'?” and it agrees well with data from
the MSRE, The more general model checked out fairly
well under some operating conditions but not so well
under others. With argon as the cover gas, measured and
computed '?**Xe poison fractions are in substantial
agreement over all ranges of circulating bubble void
fraction. With helium as the cover gas the agreement is
good at high void fractions, but at low void fractions
the measured value is considerably less than the
calculated value. The analytical model would predict
very little difference, if any, with helium or argon as the
cover gas. This discrepancy seems to be associated with
the difference in solubility of helium and argon and its
interaction in some way with bubble mechanics. Never-
theless, the above analytical model will be used for
MSBR design calculations; if in error, the design should
be conservative as far as '3%Xe is concerned. As the
model is improved, these calculations will be updated.

170

THEORY

The steady-state analytical model involves a rate
balance on the noble gas in fuel salt and a fuel loop
with the characteristics of a well-stirred pot:
generation rate = decay rate in salt +

burnup rate in salt + migration rate to graphite +

migration rate to circulating bubbles,
where

migration rate to graphite = decay rate in graphite +

burnup rate in graphite
and

migration rate to circulating bubbles = decay rate in
bubbles + burnup rate in bubbles +
stripping rate of bubbles.

A typical migration term can also be represented as
follows:

migration rate to bubbles = 24(C — C;) ,

where

h = mass transfer coefficient,
A = total bubble surface area,

C = concentration of xenon isotope dissolved in bulk
salt,

C; = concentration of xenon isotope in salt of bubble
interface.
These equations are explained in detail and each term
is evaluated (for the MSRE) in ORNL4069.'%® The
mass transfer coefficients have been evaluated from
standard relationships for heat transfer coefficients and
use of the heat-transfer—mass-transfer analogy. In the
first equation, shown above, the term on the left of the
equality sign is a constant at a given power level. All
terms on the right side of the equality sign are functions
of the '3°Xe concentration dissolved in salt, The
concentration therefore may be solved for. Knowing
this, the rate terms and the '?%Xe poisoning due to
xenon in the salt, bubbles, and graphite can be
computed. The wvalues for '?SXe poisoning are
presented in this report in terms of the *“poison
fraction,” which is defined as the number of neutrons
absorbed by '?5Xe compared with the total number of
neutrons (fast and thermal) absorbed by 232U. The
reactor parameters used here are listed in Table A.1.
The values may not be exactly the same as those used
elsewhere in this report, but they are sufficiently close
and no great error is involved,

Very early in the MSBR conceptual design, it was
decided that bubbles would be injected into the fuel
loop at the core discharge and removed at the core
inlet. The objective was to keep the core nominally free
of bubbles and thus avoid any effects that they might

Table A.1. Reactor parameters used in noble-gas migration

calculations?
Reactor power, MW(t) 2250
Salt volume in fuel loop, £t 1416
Total fuel-salt fow rate, cfm 7710
Totul volume of core zones | and 11, [ P 1851
Total volume of annulus and plenums, ft 502
Graphite surface area in core zones | and I1, rn? 24 800
Graphite surface area in annulus and plenums, t° 706
Average salt fraction in zones [ and I, % 16
Salt fraction in annulus and plenums, % 100
AverEFc tht_:;mal-neutmn flux, neutrons 4.0 X 10'4
sec -~ cm
In zones I and 11 4.0x 10"
In annulus 3.0x 102
Average fast-neutron flux, neutrons
sec” 1 om 2
In zones [ and 11 6.3% 10'%
In annulus 20x 10"
Total 233U absorption cross section, b
For thermal neutrons 263.1
For fast neutrons 32.5
Effective 222U concentration in fuel salt, 841X 1075
atomsb ) em”
Henry's law constant for xenon in fuel salt, 275 % 107

moles of xenon per atmosphere per cubic
centimeter

@These parameters may not in all cases be exactly equal to
those used in the MSBR reference design, but the differences
would have small effect on the overall conceptual design.

171

have on reactivity. This was a considerable problem
because bubble generators and separators are normally
fairly high-pressure-drop components, and of course the
main fuel pump would have to generate this head. A
change in ground rules then allowed up to 1% bubbles
by volume of salt in the core, This greatly simplified the
problem because it permitted the bubbles to circulate
many times around the fuel loop and let them approach
saturation, The volumetric flow rate of helium in the
off-gas system is considerably reduced, and the bubble
generation and removal components can be put in a side
stream rather than in the main loop piping. The
question now is how many times bubbles can be
circulated around the fuel loop before they are almost
saturated with '3%Xe. Calculations pertinent to this
question were made, and the results are shown in Fig.
A.1. Two things apparent from this figure are: (1)
Bubbles can be recirculated about 20 times around the
fuel loop before the back pressure of '**Xe in the
bubble starts to significantly reduce the stripping

ORNL—DWG 68— {3037R

- —T
o
l G
L2 B T <
aal
w —
? %
- B z e
BE —
2 = Mg
S = @°
= 5 i 28
2] o @
= & w2
(T8 o O
w
= 4 ' 9 udJ E
@ R 5=
g . \ ) W w3
| < 2
QU c w <
| & U
1l \ 5 n=
2 = ———Y 9430
l%¥ ‘ J— —{ 04 18,900
\ : s ' 0.6 28,300
. } 08 37,700
i 1.0 47,200
o | | ¢,_
0 10 20 30 40 100

PERCENT OF BUBBLES STRIPPED
FROM LOOP PER CYCLE

Fig. A.1. Xenon-135 poison fraction as a function of percent
bubbles stripped from fuel loop per cycle.

Parameters:

1000 MW(e)

Unsealed graphite

Bubble diameter = 0.020 in.

Bubble mass transfer coefficient — 2.0 ft/hr

Graphite permeability to He at room temperature =103
em?/sec

Graphite void available to xenon = 10%
efficiency. (This is the basis for the 10% recycle around
the fuel pump specified in the reference design.) (2)
Even with a 1% average volume of bubbles in the fuel
loop and with a graphite permeability of about 107°
em? sec, the target value of a '35 Xe poison fraction of
0.5% is not quite attained. Average loop void fractions
as high as 1% are undesirable, because at these
concentrations small bubbles tend to coalesce. It may
be noted that if the average loop void fraction is 1%,
the maximum void fraction at the pump suction will be
a few times greater because of the pressure gradients in
the fuel-salt loop.

Since it is desirable to keep the average loop void
fraction well below 1%, another avenue to attack the
135Xe problem must be found. The most obvious one
is to use a graphite with a much lower permeability, but
this grade could be expensive and difficult to obtain. It
was therefore decided (o investigate the effect of a very
thin coating of low-permeability carbon (chemically
deposited) on the surface of higher-permeability bulk
graphite. Figure A.2 shows the results of this calcula-
tion, The parameters were chosen to yield a high poison
fraction (approximately 1.9%). With these parameters
the calculations were repeated to obtain the effects of
the permeability and thickness of the sealed layer on

ORNL —RWE an - 13038R
DIFFUSION COEFFICIENT OF Xe IN SURFACE COATING AT 1200°F “lz/lvi
(GPERMEABILITY OF Me AT ROOM TEMPERATURE WITH UNITS OF ém®/sec)
VOID FRACTION IN GRAPHITE AVAILABLE TO XENON ")
3 — e

——— E— T,_

0l {

@

® @

» :
we* a0
332

! |

354 ROISON FRAGTION (72

4 ]
10 1%
COATING THICKNESS lin)

o =

20

(et

Fig. A.2. Xenon-135 poison fraction as a function of graphite
sealing parameters.

Parameters:

1000 MW(e)

Bulk graphite permeability ~10~% cm”/sec

Average void fraction of bubbles in fuel loop = 0.2%
Bubble diameter = 0.020 in.

10% bubbles stripped per pass

172

the poison fraction. In this calculation it was assumed
that the void fraction in graphite available to xenon
decreased by one order of magnitude when the per-
meability decreased by two orders of magnitude.

It can be seen that the larget poison [raction of 0.5%
ie readily obtainable with
fraction of only 0.2% if coatings with permeabilities of
107® cm? /sec and only a few mils thick can be altained.
Work on the coating of graphite has been promising, as
discussed by Eatherly'® and in Sect. 3.2.3 of this
report. For the purpose of the MSBR design, it was
assumed that coated graphite will be available with
permeabilities of about 107® ¢m?/sec and a few mils
thick. The bubble generator and separator will therefore
be designed on the basis of 0.2% average void in the fuel
loop, of bubbles 0.020 in. in diameter, and a recycle
flow rate around the pump of 10%,

- e pesladtmiss wnSd
an QVOTagEC Citduating vouu

Migration calculations have been made for all other
fission product noble gases, of which there are over 30
kryptons and xenons. The results are shown in Table
A.2, and the flux terms are defined in Fig. A.3. The gas
migration parameters used to generate Table A.2 were
chosen to yield an equivalent ' ?* Xe poison fraction of
0.56%. The fluxes would be about the same for any
reasonable combination of parameters that yield the
same poison fraction. The decay constants and yields
listed in the table are not necessarily equal to the
accepted values in the literature but were chosen either
because of some peculiarity of the computer code or to
make some aspect of the design conservative, Note that
in the case of long-ived noble gases, the flux into the
bubbles is less than the flux out of the reactor. This is
because the longlived noble gases are recycled back
through the reactor as shown in the figure. In the case
of short-lived noble gases, the flux into the bubbles is
greater than the flux out of the reactor. This reflects
some decay of gases during their residence time in the
bubble but before the bubble is stripped from the fuel
salt. For very short-lived noble gases, the well-stirred
pot model is not applicable as pointed out earlier;
nevertheless, the computed results have been included
and are probably adequate for preliminary designs. The
very short-lived noble gases are not in themselves
significant in the reactor design.

With the above tabulated fluxes of noble gases into
the graphite, their contribution to afterheat can be
computed. Figure A.4 shows the results for an equiva-
lent '35 Xe poison fraction of 0.56%. It is assumed that
the noble-gas flux into graphite is constant and con-
tinues for two years with the reactor at power. The
total amount of noble gases and their daughters
accumulated in the graphite after this period of time
173

Table A.2. Noble-gas migration in the MSBR

135 e Poison Fraction = 0.56%

Cl::::ayn ; Cumylative Thermal Noble-Gas
Noble-Gas y Yield Cross Noble-Gas Noble-Gas Noble-Gas Flux to
fsatope Usedin o 233y gection  Flux to Fliix to Flux Out Helium
Calcul:uon (fraction) (barns)  Graphite Bubbles of Reactor Cleanup
(hr™") (atoms/hr) (atoms/hr) (atoms/hr) System
(atoms/hr)
82y 101075 0.003 45.0 1.24 x 10" 211 x10*' 1.05 X 10?* 2.11 % 10*!
83gy L.OX10™° 00114 0 285 % 10'7 3.00x10*' 1.50x10%** 3.00 x 10*
B4gr 1.0%x10°° 00110 0.160 281 x10'7 290 x10*' 1.45x10** 2.90 % 10*!
85kr 735 X10°% 00249 0096 4.67x10"7 6.56 X 10*" 3.28 x 10** 6.56 x 10*'
g 1.0X1075 00328 0.060 8.27x10'7 864 % 10*" 4.32X10** 8.64 X 10*’
87 ks 0.547 0.0450  500.0 7.15 % 10" 110X 10?111 x 10*? 8.35 x 10'?
oy < 0.247 0.0570 0 8.81 x 10°° 140 x10%?* 1.70x10*? 7.72 x 10?°
89kr 13.0 0.0623 0 9.49 X 10° 1.14 x 10*? 8.16x10** 0
0k, 75.6 0.0555 0 5.66 X 10%° 474 x 10*" 1.43x10%' 0
tdd 1) 249,0 0.0410 0 228 X 10°° 1.44x10*' 1.67%x10%° o
2y 832.0 0.0296 0 6.96 x 10'? 353x10%° 1.33x10'? o
93Ke 1230.0 0.0142 0 2.43% 10" 1.17x10%° 3.02x 10 0
ke 24960 0.0062 0 5.82x 10" 257 x10" 332x10'7 0O
SKkr 24900 0,0019 0 .78 x 10'® 7.87x10'® 1.02x10'7 o
120% 1.0 X107  0.0020 1.50 1.30x10'7 527 x10%° 263x10%" 5.26 x 10*°
128ye 1.0X107°  0.0002 250 1.36%x10'% s5.00%x10'? 250x10%° 4.99 x 10'?
129%¢e  10Xx10~°  0.0210 0 1.16 X 10'® 553 x10?" 2,76 x 10%? 5.52 x 10*"
130%e  1.0x10°  0.0010 250 7.38x10%° 271 x10*° 1.35x 10%' 2.71 % 10*°
I3lge 1.0X10™° 00385 1200 316 X 10"°  1.01 x 10** 5.05 X 10** 1.01 x 10?2
18820 1.0X107° 00548 020  3.10x10'® 144 x10%2? 720X 10%** 1.44 X 10%?
133%e  548%x107° 0.0648 1900 8.68 x 10%° 1.62 X 10®? 4.21x 10** 6.47 X 10*'
13%e  1.0Xx10™° 00683 0.20 3.87x10'® 1.80x10%** 898 x10** 1.79 X 10*?
135xe 0.0753 00616 1.05x 10% 1.93x 10*' 1.41x10%*?* 1.43x10** 7.68 % 10"
136 e 1.0 X107  0.0700 0.15 3.94 X 10'®  1.84 x 10*°? 9.20 x 10*°? 1.84 x 10*?
g 9.90 0.0716 0 2.04 X 10°'  1.3¢ x 10** 1.03x10*? 0
138y, 2446 0.0663 0 202X 10! 145 X 10*? 1.35%10** 0
e, 60.85 0.0493 0 9.01 X 102 475 x 10*" 1.66 X 10%" 0
140%:  156.0 0.0352 0 3.87x 10%° 180x 10*' 3.12x10%° 0
41xe 12500 0.0180 0 3.73x 10'° 146 x10%° 371%x10'® o
192y 1660.0 0.0163 0 2.60 % 10"° 1.00x10%° 194x10'® o
193%e 24900 0.0017 0 1.84 X 10'® 6.96 x 10'®* 9.01x10'% 0
144%e  2490,0 0.0001 0 7.66 X 10'® 290x 10'7 376 x 10*% 0

can then be computed. The computation assumes that
all daughters remain in the graphite. For simplicity, it
also assumes straight chain decay and no branching
decay loops.

There are several areas in the theory and application
of noble-gas migration where development is needed.
The most necessary and potentially fruitful area is to
explain the reason for the lower than expected poison
fraction in the MSRE at low void fractions with helium
as the cover gas. As noted in the beginning of this
Appendix, when argon is the cover gas, measured and

computed *?*5Xe poison fractions are in substantial
agreement over all ranges of circulating bubble void
fraction. With helium as the cover gas, the agreement is
good at high void fractions, but at low void fractions
the measured poison fraction is considerably less than
the calculated value. The analytical model would
predict very little difference, if any, between the two
cover gases. The discrepancy seems to be associated
with the difference in solubility of helium and argon
and the relationship between solubility and bubble
mechanics. To illustrate, suppose helium bubbles 0.020
ORNL-DWG 69-5486

|
VOLUME
HOL DUP
FLUX OUT
OF TOTAL DOLAY
REACTOR } Kr -6 hr
SRR Xe-48 hr
. 47-hr Xe
LINE HOLDUP
N CHARCOAL
REACTOR FUEL LOOF | BED
(WELL-STIRRED POT)
FLUX TO BUBBLES
———(=) BUBBLE
FLUX TO GRAPHITE
» /:":/;Z//"J
.‘G/B’ep./“l‘rEJ
S /
b 7 Z
0.8 OF FLOW
0.2 OF FLOW
PURE FLUX TO
He HELIUM
MAKEUP  GLEANUP
SYSTEM

Fig. A.3. Flow diagram to define terms used in Table A.2.

in. in diameter are injected at the MSBR pump suction.
Further, assume that the fuel salt is saturated with
helium at the pressure and temperature of the pump
suction and that the bubbles go through the pump
where the pressure is raised to over 200 psi. If the
bubbles are allowed 10 equilibrate with the salt, they
will completely dissolve and disappear. In the case of

5 ORNL-DWG 69~ 5487
10— TR rtmm—rflvmnrrfl:m“
- 411 CONTINUOUS TIME OF REACTOR AT POWER=2years |4
| NOBLE GAS MIGRATION PARAMETERS SUCH THAT

- A 135xs POISON FRACTION =(0.56 T
- TIME=0
= “ 4 — 3

— J

— ’7; i

N
S
3 ! . Nl
(6,94 days| J
ot s P j Lt 1| | @A days)
109 10! 107 104 109
TIME AFTER REACTOR SHUTDOWN (min)

i
Ladl

- A

37
T
i
F4uli
-t

— — -
I L

HEAT GENERATION RATZ IN GRAPHITE (kw)
l
1
\
1

Fig. A4. Afterheat contribution by noble gases and their
daughters adsorbed by the graphite in the MSBR core.

argon, the solubility is sufficiently low that the bubble
will not disappear but will only compress in size. Of
course, the bubbles are subjected to these high pressures
only for a few seconds, so the dissolution process must
be quite rapid. (The entire loop cycle time is only about
11 sec.) If the helium does dissolve completely, at some
location near the pump suction the bubbles will rapidly
nucleate and the gas come back out of solution.

A questionable parameter in noble-gas migration
calculations is the mass transfer coefficient to bubbles
suspended in a turbulent fluid. A literature survey and
analysis was made of this parameter by Peebles.'*® He
concluded that the mass transfer coefficient will fall in
the range of 2 to 13 ft/hr depending on whether the
bubble has a rigid or mobile interface respectively. A
program is currently under way to determine this
parameter for turbulent flow in a glycerol-water and
helium bubble system. A mass transfer coefficient of
20 ft/hr has been used in the design calculations
because the small helium bubbles in molten salt are
expected to have a rigid interface. Use of this coef-
ficient also tends to make the ' *°Xe poisoning calcula-
tions conservative.
Appendix B

Neutron Physics

Table B.1. Concentrations and neutron absorptions Table B.2. Neutron flux by energy group at the midplane
in fission products at equilibrium for the single-fluid of the single-fluid MSBR reference design
MSBR reference design
Effective processing cycle times are given in Table 3.7-B Energy a 0 neu trl;l::cm'“ sec")
Nuclide® (:fg:;c:")‘i’l‘“c‘;“.. ) Absorptionb Group Lower(::;;x Bdary Center of core ch:n:,;:fn
1498m 45% 1077 6.5% 1073 1 8.2x 10° 1.22 0.20
143Ng 24% 1077 15% 107 . %A 13: 241 L4g
15 - . 9 .
; 4;3"‘ 3K 10:: 1% 10_; 4 4.8% 10 2.41 0.46
ot o B 15510 5 1.9% 10° 2.02 0.29
'53gy 57% 107 7.5% 107 6 7.7% 107 0.61 0.07
155py 2.7% 1077 7.0%x 107* 7 1.8 x 107! 3.25 0.21
148p . 1.3x 1077 54x%x 107 8 6.0 X 10:; 3.33 0.14
154, 1.7 % 107% 49% 1079 9 4,7 X 10 1.06 0,04
145Nd 2,1% 1077 42x 107"
143py 1.0x 1077 23x 107
Yos 1,7 % 107° 2.2x% 107
205y 8.5% 107¢ 21x 107
150m 46x 107® 1.2x 107
H1lp, 2.0x 1077 58%x 107°
13784 51% 1077 49% 107°
s 1 1.7% 1077 3.3%x 1078
152gm 22% 1078 29% 1073
6" 2.6x 1077 29%107°
g - 4.6 %1077 28% 107°
1408, 1.0x 1077 24x% 1075
Others 5.1% 107
Total 152%x 1073

“In decreasing order of neutron absorption.
bNeutrons absorbed per neutron absorbed in fissile material,

175
Appendix C

Equivalent Units for English Engineering and International Systems®

: ineeri ipli ternatio
Quswity Engmhu:;?;n i ort:li:psl;e l:t:?ls sy:t‘;mr?;l) ur:ilts“
Area ft2 0.0929 m?
Density b/ft> 16.027 kg/m>
F_orceb Ib 44481 N
Heat load Btu/hr 0.293 W
Heat rate® BtukW ™! hr™ 292.8 X 107° JW™! sec™!
Heat transfer coefficient Btuhr™' 12 °p)! 5.67 wm™Z K™
Length ft 0.3048 m
“Mass’ flow rate Ib/hr 126 X 107° kg/sec
Power hp 745.7 W
Pressure psi 6894.6 N/m?
Quantity of heat® Btu 1054.7 J
Specific heat (heat capacity) Btulb™' (°F)”! 4187 Tkt ex)7!
4 : 3 .5
Specific speed (pumps) M 2,027 x 1072 graa/sboicinr Jpe)™
(££)0:75 (m)0:75
Stress psi 6894.6 N/m?*
Thermal conductivity Btu hr ' (°F) ) £t 1,731 wm™ k)™
Thermal expansion per°F 0.555 per °K
Velocity (linear) fps 0.3048 m/sec
Velocity (angular) pm 0,1047 radians/sec
Viscosity? bhr™ £t 0.4138 X 107 Nsec™! m™?
Volume it 0.02832 m>
Volume flow gpm 63.1% 107¢ m?/sec
Weight equivalent Ib 0,4536 kg
Work ft-Ib 1.351 J

@Table S,1 is expressed in English engineering units as commonly used in MSR literature and in the meter-kilogram-second (MKS)

gystem, which closely follows the International System (SI),

bIN=10° dynes=1kgmsec? =1 J/m.
©17=2.778 X 10™ W-hr = 0.0009482 Btu.

lesecm'z=lkgsec-' mw Y.

176
Appendix D

Cost Estimates for the MSBR Station

Roy C. Robertson
M. L. Myers

Table D.1. Estimated construction cost for MSBR power station?
Based on January 1970 costs

Account Cost (thousands of dollars)
[tem
No. Materials Labor? Total
20 Land€ (see account 94) 590
21 Structures and site facilities
211 Site improvements 500 565 1,065
212 Reactor building
212.1 Basic structure? 3,358 3,358 6,716
Special materials (see Table D.3)
Stainless steel liner at $1.20/1b 334 143 477
Carbon steel at §0.60/1b 1,850 1,240 3,090
Insulation at §10/ft? 321 137 458
212.2 Building services 325 175 500
212.3 Containment structures at $2/Ib 1,900 1,900 3,800
Subtotal for account 212 8,088 6,953 15,041
213 Turbine building® 2,200 1,800 4,000
214 Intake and discharge structures 540 360 900
218A Feedwater heater bayf 1,720 1,410 3,130
218B Loading and set-down bay/ 590 480 1,070
218C Offices, control rooms, ete. 450 300 750
218D Warehouses and miscellaneous 36 24 60
Subtotal for acoount 218 2,796 2,214 5,010
219 Heat rejection stack® 320 480 800
Subtotal for account 21 14,444 12,372 26,816
Contingency: 5% materials, 10% labor 722 1,237 1,959
Spare parts: % 76 76
Total for account 21 15,242 13,609 28,851
22 Reactor plant equipment
221 Reactor equipment”
221.1 Reactor vessel 9,100 400 9,500
221.2 Control rods 1,000 100 1,100
221.3 Graphite (see Table D,5Y 7,200 200 7.400
Subtotal for account 221 17,300 700 18,000
222 Main heat transfer system
222.11 Fuel-salt pumps 3,100 200 3,300
222.12 Primary system salt piping 300 129 429
222.13 Primary heat exchangers (see Table D.6) 7,100 200 7,300
222.31 Coolant-salt pumps 4,200 200 4,400
222.32 Secondary system salt piping 1,330 570 1,900
222.33 Steam generators (see Table D,7) 5,790 480 6,270
Reheaters (see Table D.8) 1468 200 1,668
Subtotal for account 222 23,288 1,979 25,267

177
178

Table D.1 (continued)

Account Cost (thousands of dollars)
Item
No. Materials Labor? Total
224 Radioactive waste treatment and disposal
224,1 Liquid waste 45 15 60
224.2 Off-gas system 350 150 500
224.3 Solid waste disposal (not fission products) 75 25 100
Subtotal for account 224 470 190 660
225 Nuclear fue] storage
2254 Primary drain tank (see Table D.9) 2,680 300 2,980
Fuel-salt storage tank (see Table D.10) 643 70 713
Salt transfer pump and jets 480 20 500
Subtotal for account 225 3,803 390 4,193
226 Other reactor equipment
226.1 Inert gas systems 280 120 400
226.2 Auxiliary boiler¥ 2,550 450 3,000
Cell heating systems’ 200 130 330
226.3 Coolant-salt drain tanks (see Table D.11) 765 35 800
2264 Coolant-salt handling 20 - 25
226.5 Coolant-salt purification system 125 25 150
226.6 Leak-detection system 150 100 250
226.7 Cell cooling system 150 150 300
226.8 Maintenance equipment (see Table D,12) 3,600 900 4,500
Subtotal for account 226 7,840 1,915 9,755
227 Instruments and controls 3,200 800 4,000
Subtotal for account 22 55,901 5,974 61,875
Contingency: 15% materials, 10% labor™ 8,385 597
Spare parts: 1.5%" 102
Total for account 22 64,388 6,571 70,959
23 Turbine plant equipment
231.1 Turbine-generator® 19,361 1,000 20,361
231.2 Foundations 225 225 450
Subtotal account 231 19,586 1,225 20,811
2323 Condensing water system 1,100 900 2,000
233 Condensers 1,500 700 2,200
234 Feedwater heating system
234.1 Regenerative feedwater heaters 1,800 100 1,900
2342 Condensate pumps 180 20 200
Boiler feed pumps 1,890 210 2,100
2343 Piping and miscellaneous
Feedwater and condensate 900 900 1,800
Extraction steam 375 37S 750
Drains and vents 125 125 250
Mixing chambers 72 8 80
Pressure-booster pumps 585 65 650
Subtotal account 234 5,927 1,803 7,730
179

Table D.1 (continued)

Account Cost (thousands of dollars)
No Item : 3
. Materials Labor Total
235 Other turbine plant equipment
235.1 Main steam pipingP 1,700 1,700 3,400
235.2 Turbine auxiliaries 250 200 450
235.3 Auxiliary cooling systems? 600 300 900
2354 Makeup and treatment 320 160 480
235.5 Condensate treatment 480 320 800
235.6 Central lubrication system 60 30 90
235.7 Reheat steam preheaters (see Table D.13) 110 25 135
Subtotal account 235 3,520 2,735 6,255
236 Turbine plant instruments and controls 330 170 500
Subtotal for account 23 31,963 7,533 39,496
Contingency: 4% materials, 8% labor 1,279 603 1,882
Spare parts 220 220
Total for account 23 33,462 8,136 41,598
24 Electric plant equipment
241 Switchgear
241.1 Generator circuity 100 30 130
241.2 Station service 1,000 100 1,100
242 Station service 450 360 810
243 Switchboards 400 70 470
244 Protective equipment 100 100 200
245 Electric structures 150 600 750
246 Wiring 2,000 2,000 4,000
Subtotal for account 24 4,200 3,260 7,460
Contingency: 5% materials, 10% labor 200 300 500
Spare parts: 0.5% 40 40
Total for account 24 4,440 3,560 8,000
25 Miscellaneous plant equipment
251 Turbine plant hoists 333 37 370
252 Air and water services 490 330 820
253 Communications 50 50 100
254 Furnishing and fixtures 350 20 370
Subtotal for account 25 1,223 437 1,660
Contingency: 5% materials, 10% labor 61 44
Spare parts: 1% 13
Total for account 25 1,297 487 1,778
26 Special materials —
264 Coolant-salt inventory” 500
265 Miscellaneous special materials 500
Subtotal for account 26 1,000
Total direct construction cost (TDC) 152,305
91 Construction equipment and services at 0.8% TDC* 1,218
921 Reactor engineering? 2,250
922 Engineering, at 5.5% TDC¥ 8,377
93 Insurance, taxes, etc., at 4.2% TDC* 6,397
94 Interest during construction, at 18.58%Y 31,687
942 Land interest during construction 420
Total indirect costs™ 50,349
Total plant capital investment 202,654

180

Table D.1 (continued)

“Estimated costs ure not for first-of-a-Kind plant but assume an established molien-salt reactor industry. Estimates are based on
January 1970 prices. Private ownership is assumed, with a prevailing interest rate of 8% and a five-year construction period.

Conlingency faclors of up o 15% have been applied. The cosl estimate follows format, account numbers, and procedures
recommended in NUS-531 (ref, 119),

®L.abor 1s for hield erection, Shop and fabrication labor are included in materials,
“For typical site at Albany, N.Y. Land cost included in indirect cost.

dAs indicated in Table D,3, basic structures include all portions of reactor and confinement buildings except dome, which is
included in account 212.3. Estimate is based on installed cost of concrete of $103/yd3 .

€Building cost based on $1.00/ft3.

fBuilding cost based on $0.65/1t°,

&Stack is 400 ft high. Based on $2000/ft.

BReactor shielding is included with structures, account 212.1.

I Average cost of Hastelloy N installed is about $14/Ib (see Table D.4).

fAverage cost of graphite is about §10/Ib (see Table D,5).

kBoiler capacity ~200,000 1b/hr.

/Based on 950 cell heaters at $200 each,

MBased on recommendations in NUS-531 (ref. 119), See text, Sect. 15.1.

"Does not include replacement reactor core.

OBased on tandem-compounded, 6-flow, 3 1-in. unit (Westinghouse price).

PBased on 900 ft of high-pressure mains at 370 Ib/ft and $0.75/1b; and on 700 £t of reheat piping at 468 Ib/ft and $0.75/Ib.
4Service water systems,

"Based on 1 X 10° Ib of coolant salt at $0,50/1b, Salt inventory is considered to be depreciating capital investment.
SFrom Fig. C-1, NUS-531, ref. 119.

"From Fig. C-2, NUS-531, ref. 119.

WBased on five years construction time-at 8% interest compounded annually and typical cash flow curve shown in Fig. C4,
NUS-531, ref. 119.

VBased on seven years ownership at 8% interest compounded annually,
Windirect costs amount to about 33% of TDC cost.

Table D.2. Estimated fuel-cycle costs for the MSBR power station

A. Estimated cost of equilibrium inventory in primary circulation system

Total weight of fuel salt in system: 1720 ft° x 208 = 357,760 1b
Total moles of fuel salt: 357,760/64 = ~5590 Ib-moles

Total
TLiF: 5590 X 0.72 X 26 X $15/1b% $ 1,570,000
BeF;: 5590 X 0.16 X 47 X $7.50/Ib 315,000
ThF4: 5590 X 0.12 X 308 X $6.50/1b 1,343,000
S0 1223 kg at $13/g 15,900,000
233p, 7 kg at $13/g 94,000
2387 112 kg at $11.20/g 1,252,000

$20,474,000
181

Table D.2 (continued)

B. Estimated cost of salt inventory in chemical processing plant
Total weight of barren salt in chemical plant: 480 f’ x 207 = 99,360 1b

Total moles of barren salt; 99,360/63.2 = 1572 Ib-moles

TLiF; 1572 % 0.72 X 26 X $15/1b%
BeF,: 1572 X 0.1G X 47 X $7.50/1b
ThF4: 1572 X 0.12 X 308 X $6.50/1b
233y;. 63 kg at $13/g

333p,; 103 kg at $13/g

C., Makeup salt cost (per year, based on 15 cal-year cycle)

LiF: (1,564,000 + 441,420)/15
BeF,: (315,276 + 88,658)/15
ThF4: (1,342,942 + 377,657)/15

D. Chemical processing plant equipment costs?

Direct construction cost equipment and field labor (cell construction cost is included
in account 21, structures) (allow.)
Indirect costs at 35%°¢

Total
E. Operating cost (per year)
Payroll and overhead directly associated with chemical processing system. Say,
F. Estimated fuel-cycle cost (mills/kWhz)
Fixed charges on salt inventory at 13.2%
Makeup salt

Fixed charges on processing equipment at 13,7%%
Process plant operating costs

Production credit, based on 3.2%/year yield
Total estimated fuel-cycle cost

$§ 441,000
89,000
378,000
815,000
1,336,000

§ 3,059,000

§ 134,000
27,000
115,000

§ 276,000

$10,000,000

3,500,000
$13,500,000

$700,000

@Based on "Li at $55/Ib, or $120/kg.
bBased on 233U at $13/gand 2*°U at $11.20/e.
©Based on 2¥?U and ?*3Pa at $13/e.

dThe estimated cost of the MSBR fuel processing equipment is not precise at this time. Figure 15.1 shows the

effect of the fuel processing equipment cost on the fuel-cycle and total power production costs.

Indirect cost of 35% is approximately the same as for the main plant. See accounts 91 through 94, Table D.1.
IFixed charges to be applied 1o the capital cost of the fuel-salt inventory over the 30-year life of the plant
cannot be precisely estimated because of the changing fuel-pricing and tax structures, and because of the
uncertainties in the handling and cleanup costs involved in recovering the fuel salt for reuse at the end of the plant
life. The fixed charges would probably [all between the 13.7% used for depreciating equipment (see Table D.14)
and the 12.8% used for nondepreciating items, as recommended by NUS-531 (ref. 119). An average value of 13.2%

has therefore been used.
fFixed charges on depreciating equipment are explained in Table D.14.
182

Table D.3. Summary of special materials in reactor building

Stainless Carbon steel” Concrete Insulation
steel (Ib) (tons) (ya®) )
Confinement building &
Nome (v48)” 18,276
Reactor cell 48,992 955 3,184 11,268
Waste storage cell 153 c
Floors 25,183
Control rod storage 8,238 72 ¢ 1,935
Spent core storage cell 12,567 130 ¢ 3,188
Replacement core cell 12,574 22 250
Spent heat exchanger cell 9435 100 1,929 2,445
Chemical processing cell 57,445 i
Freeze-valve cell 5,028 55 45 1,351
Off-gas cell 28,722 c
Hot cells 12,724 617
Auxlliary equipment cells 333
Drain tank cell 12,571 130 250 2,250
Miscellaneous concrete 32
Reactor building

Floors 4,340
Exterior walls 5,698
Interior walls 4,933
Steam cells 94,423 956 c 23,328
Coolant-salt drain cell 94,423 e

Total 397,142 2,573 65,070 45,765

@Carbon steel for containment and shielding only. Does not include reinforcing or structural steel.
BIncluded in account 212.3, Containment Structures.
“Concrete included elsewhere in Table D.3.
183

Table D.4, Estimated cost of Hastelloy N in reactor vessel?

Weight Cost per pound Shop labor Materials and labor

(Ib) (dollars) (10® dollars) (10 dollars)
Removable upper head assembly
Cylinder extension (18 ft OD, 13 ft high, 2 in. thick) 68,130 10 341 681
Flange (20.66 ft OD, 18 ft ID, 6 in. thick) 22,480 15 225 337
Head? (18 ft diam, 3 in. thick) 40,800 15 401 612
Control rod pipe (18 in. diam, 20 ft high, 0.56 in. thick) 2,420 25 48 60
Miscellaneous internals (allow.) 1,000 25 20 25
Reactor vessel, permanently installed
Upper flange, as above 22,480 15 225 337
Cylinder extension, as above 68,130 10 341 681
Portion of top head (22.53 ft OD, 18 ft ID, 2 in. thick) 15,410 15 150 231
Head skirt (22.4 ft av diam, 6 in. high) 3,260 10 16 33
Vessel cylinder (22.4 ft av diam, 13 ft high, 2 in. thick) 84,920 10 425 849
Bottom head? (22.53 ft diam, 3 in. thick) 63,920 15 639 959
Bottom well (3 ft diam, 4 ft high, 1 in, thick) 1,750 15 18 26
Bottom ring (1 ft 8 in., 3 in,, 17 ft 6 in, ID) 14,003 25 280 350
Top ring (1 ft 9in., 3 in., 17 ft 8 in. 1D) 14,861 25 297 372
Reflector retainer rings (2 in., 4% in., 21 ft diam) 10,208 25 204 255
Bottom ring (3 in., 6 in,, 16 ft 2 in, ID 3,627 25 73 91
Nozzles, etc. (allow.) 5,000 38 165 190
Miscellaneous internals (allow.) 2,000 25 40 50
Replaceable core assembly
Internal head? (18 ft diam, 3 in. thick) 40,800 15 401 612
Bottom ring® (92 in.2, 16.3 ft diam av) 18,200 25 364 455
Miscellaneous internals (allow.) 1,000 25 20 25
Alternate removable upper head assembly (see above) 134,830 10-25 1035 1715
8946
Transportation to site 200
Total (does not include field labor) 9146

aEstimated weights based on Hastelloy N density of 557 1b/ft?, and on Figs. 3.1 and 3.2.

b2 1+e
PInside surface area of ellipsoidal head = ma’ +-"27 In

=0.9D?,
l1—e

where
a=D/2,
b= 3a/11 (for MSBR), and

e =/ (a® - b*)a=0.962.

€An irregular shape; see Fig. 3.2
184

Table D.5. Estimated cost of graphite for MSBR

Pound Cost
St (thousands of dollars)
(L T v+ d
TYCENLS UL Blapiine
Zone I, 13% salt
Octagon (14.33 ft across flats, 13 ft high) 221,400
Zone 1, 37% salt
Axial (9 in. thick, top and bottom octagon) 18,496
Upper end elements (3 in. thick, top octagon) 735
Radial (16.83 ft OD, 14.5 ft high) 55,055
Salt inlet, upper part? 880
Radial vessel coolant plenumb 3,360
Radial reflector, 1.2% salt (22,16 It OD, 17.16 ftID, 14,5 ft high) 254,395
Axiul reflector, bottom,? 3% salt (20.2 1 effective diameter) 54,816
Axinl reflector, top, 3% salt (sume as above) 54,816
Outlet pnssnge" 5,400
Summary of graphite weights and costs
Extruded elements and shapes (at $11/1b)
Zone I and zone II axial 240,631
Zone I1 radial 55,055
295,686 3252
Reflector pieces (at $9/1b)
Radial 254,395
Axial, top 54,816
Axial, bottom 54,816
Qutlet passage 5,400
Coolant plenum 3,360
Salt inlet 880
373,667 3363
Alternate head assembly (at $9/1b)
Axial reflector, top 54,816
Outlet passage 5,400
60,216 542
Total graphite, including alternate head assembly 729,569 7157
Replaceable core assembly
Zone | 240,631 "
Zone 1) radiat | 2t $11/10 55,055 e
Axial reflector, bottom 54,816
Salt inlet, upper part } at $9/1b 880 S
351,382 3753

@Weights based on graphite density of 115 Ib/ft>.
PErom estimates by H. L, Watts.

“Based on volume of spheroid: ¥ =4/3 na®b, where (for MSBR) 4 = 10.1 ft and b = 2.3 ft. Thus ¥ = «D?/52.7 (for one head),
185

Table D.6. Estimated cost of primary heat exchangers®

Description for each of four units
Material, Hastelloy N

5543 tubes, 0.375 in. OD, 0.035 in, wall thickness, 22.07 f{
long (each unit)

Total surface, 12,011 ft? x 4= 48,044 ft?

Sheil ID, 66.2 in.

See Fig, 3.33
Weights of Hastelloy N?
Tubes (70,800 Ib at $30/1b) $2,124,000
Cylinders (192,400 Ib at $10/1b) 1,924,000
Heads (7,800 Ib at $15/1b) 117,000
Tube sheets, rings, etc, (149,100 1b at §20/1b) 2,982,000
$7.147,000
Installation labor 200,000
$7,347,000

Time did not permit revising the above cost estimate to agree
with the latest primary heat exchanger data, as listed in Table
3.15.

bWeights are for total of four units.

Table D.7. Estimated cost of steam generators
Total of 16 units

Total surface: 56,432 ft?
Material: Hastelloy N
Tubes: 380 tubes. 0,50 in, OD, 0.077 in. wall thickness, 70.9 ft
av length (each unit)
Total weight? = 170,609 1b at $20/1b $3,400,000
Shells: 18 in, ID, 0.375 in. wall thickness, 71 ft av length
Total weight = 95,122 1b at $10/Ib 950,000
Spherical heads: 28 in. OD X 4 in. wall thickness, total 32
Total weight = 74,661 1b at $15/1b 1,120,000
Nozzles, baffles:  Weight = 16,000 Ib (say) at $20/1b 320,000
Installation 480,000
$6,270,000
@Total weights are for 16 units.
Table D.8. Estimated cost of steam reheaters
Total of eight units
Material: Hastelloy N
Total surface: 2253 X 8 =18,024 ft?
Tubes: 392 tubes. 0.75 in, OD, 0.035 in, wall thickness,
29.27 ft long (each unit)
Total weight? = 27,911 Ib at $30/1b § 837,330
Sheli: 21 in. ID, 0.5 in, wall thickness, 30 ft long
Total weight = 31,352 1b at $10/1b 310,352
Tube sheet: 21 in, diam, 4 in. thick
Total weight = 7145 1b at $§10/1b 71450
Heads:; 10.5 in. radius, 0.75 in, thick (assumed hemispherical)
Total weight = 3215 b at $20/1b 64,000
Baffles: 21 in. diam, ¥, in. thick, 70% cut; total 36 per unit
Total weight = 8440 1b at $§10/Ib 84,400
Nozzles, etc., say 4000 1b at $25/1b 100,000
Installation labor 200,000
Total $1,668,000

T otal weight is for eight units.
186

Table D.9. Estimated cost of fuel-salt drain tank

Description: 13 £t 9 in. OD, 21 ft 9 in. high; see Fig. 6.2 and Table 6.1
Material: Hastelloy N
Heads: Total of four. 13 ft 9 in, diam, 1'4 in. thick

Aspect ratio, 7,

2 Cmemes Ve =V Fl L
Aren of head 0,.9D% (cae footnote b, Tubla D 4} onc head hias & S-{t-diam hole ai cenier hine

Weight, 4 X 557 (21.7 ft? — 2.45 ft*) = 42,800 Ib at $15/1b 642,000 $ 642,000
Cylinders: Totalof two, 13 ft 9 in, diam, 16.5 ft high, 1 in. thick = 66,200 Ib at $10/lb 662,000
U-tubes: Total of 1500. % in. OD, 0.042 in. wall thickness, average length, 17.5 ft

Weight = 18,950 1b at $30/1b 567,000

Headers: Total of 40, 3-in. pipe inside 6-in. pipe, about 6 ft long
Weight = 7260 Ib at §15/1b 108,900
Nozzles, baffles, ete, (allow,) 2000 1b at $25/1b 50,000

Heat-removal system cost allowances?

Salt-to-steam exchanger $ 200,000
Steam-to-air exchanger for 40 MW(t) 200,000
Piping, etc. 250,000
$2,679,000
Installation labor 300,000
Total $2,980,000

@Capacity = 18 MW(1).

Table D,10, Estimated cost of fuel-salt storage tank

Description: Tank is essentially same as fuel-salt drain
tank, except that cooling tubes in tank are
salt-to-steam transfer as in MSRE and no
intermediate heat exchanger is required

Material: Stainless steel (p = 495 lb/fta)
Tubing: 16,800 1b at $5/Ib = § 84,000
Cylinders: 58,600 1b at $3/1b = 175,800
Heads 38,000 Ib at §5/1b =
Headers 6500 1b at $5/1b = } 232,500
Nozzles, etc. 2000 1b at §5/1b =
Installation labor 150,000
Heat-removal system cost allowance
Steam-to-air exchanger 100,000
Piping 50,000
$642,300
Installation labor 70,000
Total $712,300

Table D.11. Estimated cost of coolant-salt storage tanks

Total of four units, 12 ft diam, 20 ft high, useful storage capacity 2100 ft> each
Material; stainless steel (p =495 1b/ fta)
Cylinders: 12 ft diam, 20 ft high, 1 in. thick
Total weight = 124,460 Ib at $3/1b $375,000
Heads: 12 ft diam, 1.25 in. thick
A =1.0D? (assumed aspect ratio for heads)

Total weight = 71,304 1b at §5/Ib 350,000
Nozzles, etc., say 8000 ib at $5/1b = 40,000
Installation labor 35,000

Total $800,000

Table D.12. Estimated cost of maintenance equipment for

the MSBR

Costs in thousands of dollars

Polar crane
Cask
Hoists
Transition piece
Maintenance containment cover
Maintenance closure
Disassembly and storage cell equipment
Maintenance shields
Long-handled tools
In-cell supports and mechanisms
Transfer cask for miscellaneous components
Maintenance control room equipment
TV viewing equipment
Decontamination equipment
Remote welding equipment and controls
Hot cell equipment
Miscellaneous

Total

600
125
150
25
120
75
500
250
400
250
50
150
150
100
1000
50
50
4500

187

Table D,14, Fixed charge rate (percent per annum) used for
investor-owned MSBR power station?

Return on money invested? g
Thirty-year depreciation® 1.02
Interim replacements? 0.35
Federal income tax® 2.04
Other taxes 2.84
Insurance other than liability® 0.25
13.7

9This table is for depreciating equipment. For non-
depreciating items, such as land, a fixed charge rate of 12.8%
was assumed, as recommended in NUS-531 (ref. 119), See Table
D.2 for fixed charge rate on fuel salt.

bReturn based on 52% in bonds at 4.6% return, 48% in
equality capital at 10%.

“The sinking-fund method was used in determining the
depreciation allowance for the 3(-year period.

dln accordance with FPC practice, a 0.35% allowance was
made for replacement of equipment having an anticipated life
shorter than 30 years. (Reactor core graphite is included in a
special replacement cost account — see Table D.15.)

®Federal income tax was based on the ‘‘sum-of-the-year
digits”” method of computing tax deferrals, The sinking-fund
method was used to normalize this to a constant return per
year.

IThe recommended value of 2.84% was used for other taxes.

£A conventional allowance of (.25% was made for property
damage insurance, Third-party liability insurance is listed as an
operating cost.

Table D.13. Estimated cost of reheat steam preheaters

Total of eight units

Material: Croloy

Total surface: 781 X 8 = 6248 ft?

Tubes: 603 tubes. 0.375 in. OD, 0.065 in. wall thickness, 13.2 ft long (each unit)

Total weight? = 15,591 Ib at $2/1b

Shells: 20.25 in. ID, 7, ¢ in. wall thickness, 13,6 ft long

Total weight = 11,880 b at $§1.50/1b
Spherical heads: About 31 in. OD X 2.5 in. thick

Total weight = 30,000 1b at $2/1b

Installation labor

$ 32,000
18,000

60,000
25,000

$135,000

@7 otal weights are for eight units.
Table D.15. Cost of replacing reactor core assemblies

in the MSBR

In thousands of dollars

Cost of assembly

Hastelloy N — see Table D.4 1,092
Graphite — see Table D.S 3,753
4.845

Chargeable power reyenue loss during core
assembly replacement?
Special labor cost per replacement? 500

Total cost per replacement 5,345
Effect on power production cost, mills/kWhr¢  0.17

A1t is assumed that the MSBR core assembly can be
replaced during the plant downtimes for inspection and
repair of other equipment, such as the turbine-generator,
which are accommodated by the 80% plant factor, and
no additional plant outage is chargeable against core
replacement.

bThe labor force for making core replacements is
assumed to be in addition to the normal plant operating
and maintenance crew.

€While various methods could be used to estimate the
cost of future core replacements, a sufficiently represent-
ative and straightforward method is to assume an extra
amount charged per kilowatt-hour, which is set aside, at
8% interest compounded annually, so that at the end of
four years the total cost of a replacement will have been
accumulated.

Repl. t
K b P $5,345,000 x 10°

10° % 365 X 24 X 0.80 (1.08% + 1.08% + 1.08 + 1.00)

= 0.17 mill/kWhr

For simplification, this method ignores the small effects
due to no accumulated funds needed the last two years
of plant operation and the fact that it is unlikely that the
plant would be shut down exactly after 30 years of
operation with 2 years of useful life remaining in the
reactor core,

188

Table D.16. Estimated annual costs for plant operation
and maintenance?

Staff payroll? $ 800,000
Fringe benefits? 80,000
Subtotal — plant staffine RRO NOD
Consumable supplies and equipment 400,000
Outside support services 140,000
Miscellancous 80,000
Subtotal 1,500,000
General and administrative 225,000
Coolant-salt makeup® 9,000
Nuclear liability insurance
Commercial coverage (net) 240,000
Federal Government coverage 67,500
Total direct annual cost 2,061,500
Fixed charges on operation and maintenance 38,800
working capital
Total annual cost $2,080,300
Contribution to power costd 0.30 mill/kWhr

@Based on cost breakdown and computation prescribed in
NUS-531 (ref. 119), The values agree reasonably well with those
reported by Susskind and Raseman (ref. 121), Costs do not
include chemical processing, which is included in the fuel-cycle
cost, nor special costs associated with periodic replacement of
the core graphite.

bBased on NUS-531 (ref. 119) recommended values for July
1968 escalated 8%.

“Makeup cost assumed to be 2% of inventory,

dBased on 80% plant factor,
189

Table D.17. Cost penalties for use of wet natural-draft cooling tower instead of fresh once-through
condensing water supply in 1000-MW(e) MSBR station as compared with penalties in light-water
nuclear power stations?

MSBR Light-water reactor?
Increased capital cost of plant due to towers, $4.,000,000¢ $6,000,000
pumps, ete.
Estimated loss in generating capacity due to 13,000 kw? 20,000 kW
heat rate increase from 7690 to 7800 Btu/kWhr
Estimated capital cost of increasing thermal $1,000,000¢ $1,500,000
capacity of plant to give 1000 MW(e) net output
Annual operating cost for towers $15 0.000f $150,000
Annual additional fuel cost due to higher heat rate $71,0008 $165,000
Increases in power production costs,” mills/kWhr
Capital cost of towers, etc. 0.08 0.12
Capital cost of additional capacity needed 0.02 0.02
Operating costs of tower 0.02 0.03
Increased fuel cost due to higher heat rate 0.01 0.02
Total increase 0.13 0.20

@Use of wet natural-draft cooling tower will increase the turbine back pressure from 1'6 to 2112 in. Hg
abs, Performance and costs of MSBR with cooling tower are taken as proportional to the effects of adding
a tower on light-water reactor performance, as estimated by Hauser.!??

DEstimated by Hauser (ref. 129).

CCapital costs of towers, pumps, etc., taken as proportion of the tower costs for light-water
reactors’ 2° on basis of amount of heat rejected to the condensing water,

dEstimated loss in capacity (and increase in heat rate) based on ratios of enthalpy drops in steam
turbine to 2‘/2 in. Hg abs vs l‘/; in. Hg abs, and equivalent effect on light-water reactor cycle.! 29

€Capital cost of increasing reactor plant capacity, flow rates, etc., to achieve 1000 MW(e) net plant
output estimated at $75/kW, as was assumed in ref, 129,

frower operating costs assumed to be the same as those for light-water reactors.

£Based on same 10 cents/MBtu chargeable to fuel-cycle cost as in MSBR reference design,

hBased on ~14% fixed charges and 0.8 power factor,

129
e

. Adams

Affe]

!'fl;nF’!-m_Zj-niu:-n
o
2
3
=
o
-
o<
(4]
-

- E.

o

. S. Billington

. E. Blanco

. F. Blankenship
. 0. Blomeke

B

. G.

> A

i o

. E.

. Bra

. A. Bredig

. B. Briggs

. R. Bronstein
. D. Brunton
. W.

, Can

. W Cardwell
A e

B

Al 7
) ;3

B

arlsmith
Carswell

I. Chang
H Chapman
J. Claffey

. F. H. Clark
. Nancy Cole
. C. W. Collins

191

INTERNAL DISTRIBUTION

48.
49.
50.
51.
52.
53.
54.
55.
56.
37
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
193
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.

Compere
. Cook
Cooke
Corbin
Crowley
Culler

Cuneo

Dale
DeBakker
DeVan
Distefano
Ditto
Doss
Dworkin
Eatherly
Engel
Federer
Ferguson
Ferris
Fraas
Franzreb
Frye
Fuller
Furlong
Gabbard
Gallaher
Gehlbach
Gi
G

M

.
.
.
.
.
.
.
.
.
.
.
-
.
»
.
.

bbons
ilpatrick
ldberg
Grimes
Grindell
Guymon
Hamner
Handley
H

L
W
T
L
L
. M.
b 4
JH,
. R.
J
A
S
R
|
E
P
LK
H
C
. H.
B.
. E.
8 ;
0.
. Go
s
. H.
iy A
:H
. A. annaf ord
LTHL
8.
.N.
. K.
. W.
. G.

H
R.
P.
.K.
. R.
v,
- E.

.M.
. F.

HOWASTROSEVEHAIA P SO =RPROE

ORNL-4541
UC-80 — Reactor Technology

94.

9s.

96.

97,

98.

99.
100.
101.
102.
103.
104.
105,
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139,

Long
Lundi
Lyon
Macklin
M

. MacPherson
a cPherson

wmwwmeHOHfi?W;PP¢F;HO$SOPN§;§WPF¢FZH

.
. A
- .
. .
.

-

pu;x:OOp;:_iOp.‘-
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.

. McNeese
. McWherter
. Malinauskas

ueller

Perry

Plper

Ragan

HH@w;O;e>wwmrfimH::38c0?wfi?:>~v

421.
422.
423.
424.
425.
426.
427.
428.
429,
430.
431.
432.
433.
434.
435.
436438.
439440.
441.
442.
443-666.

192

170. D. M. Richardson 209. R. E. Thoma
171. M. Richardson 210. L. M. Toth
172. G. D. Robbins 211. D. B. Trauger
173-182. R. C. Robertson 212, Chia-Pao Tung
183. R.G. Ross 213, W.C. T. Stoddart
184 T Roth 214, W_E, Unger
185. J. P. Sanders 215. G. M. Watson
186. H. C. Savage 216. J. S. Watson
187. W. F. Schaffer 217. H L. Watts
188. Dunlap Scott 218. C. F. Weaver
189. J. L. Scott 219. A. M. Weinberg
190. H. E Seagren 220. J. R. Weir
191. C. E. Sessions 221. H. L. Whaley
192. J. H. Schaffer 222. M. E. Whatley
193. W. H. Sides 223. J. C. White
194, M. Siman-Tov 224. R. P. Wichner
195. M. J. Skinner 225. L. V. Wilson
196. G. M. Slaughter 226. Gale Young
197. A. N. Smith 227. H.C. Young
198. F. J. Smith 228. 1.P. Young
199. G.P. Smith 229. E. L. Youngblood
200. O. L. Smith 230. F. C. Zapp
201. I. Spiewak 231-233, Central Research Library
202. H. H. Stone 234. ORNL — Y-12 Technical Library
203. R. A. Strehlow Document Reference Section
204. R. D. Stulting 235-384. MSRP Director’s Office (Y-12)
205. D. A. Sundberg 385419. Laboratory Records Department
206. J. R. Tallackson 420. Laboratory Records, ORNL R.C.
207. E. H. Taylor
208. W. Terry

EXTERNAL DISTRIBUTION

Kermit Laughon, RDT Site Office, Oak Ridge National Laboratory
C. L. Matthews, RDT Site Office, Oak Ridge National Laboratory
D. F. Cope, RDT Site Office, Oak Ridge National Laboratory

T. W. McIntosh, Atomic Energy Commission, Washington

F. N. Peebles, University of Tennessee, Knoxville

H. M. Roth, Atomic Energy Commission, Oak Ridge

M. Shaw, Atomic Energy Commission, Washington

J. A. Swartout, Union Carbide Corporation, New York

W. L. Smalley, Atomic Energy Commission, Oak Ridge

David Elias, Atomic Energy Commission, Washington

W. T. Hannum, Atomic Energy Commission, Washington

D. R. Riley, Atomic Energy Commission, Washington

J. J. Schreiber, Atomic Energy Commission, Washington

F. N. Watson, Atomic Energy Commission, Washington

M. J. Whitman, Atomic Energy Commission, Washington
Director, Division of Reactor Licensing

Director, Division of Reactor Standards

Laboratory and University Division, AEC, ORO

Patent Oftice, AEC, ORO

Given distribution as shown in TID-4500 under Reactor Technology
category (25 copies — NTIS)