J W} Y LT e DOE REVIE a - g HEGE Nl AUTHORITY - DELEBATED 8Y ERDA 8-15-77 SOREAL T G LT u!’-{.‘:.h\f\).:‘;..’., Pl KRS - OAX BATE_JUL 201979 DHOUER T, DUFF, 0og T ? AR LA .Fr'. ¥ ol B, ORNL/CO '{{"\ L \ OTEm BAK S. TETS DG P. & ;.-;zl ).‘AS-"-'-“;‘:;‘FJ“.,.:. . 3 .:!.':.-. ¥, .:3’_: ‘. '_.%:5) g ?Thfi#dpcumnt «ontdi fi?s_-*,Res-‘trmged"i@u EnetgywAc’! of-1554 in any mdm!enb 14s tmn-mmdkgufie dcg * ufluuthonze mifili‘ is prchi AIRCRAFT REACTOR TEST HAZARDS SUMMARY REPORT W. B. Cottrell W. K. Ergen A. P. Fraas F. R. McQuilkin J. L. Meem - T TR ¢ = g s T — e ' For: H. T. Dray, 'Smrvm: Laboratory. Records: Ragh - OREL - —_— OAK RlDGE NATIONAL LABORATORY OPERATED BY CARBIDE AND CARBON CHEMICALS COMPANY A DIVISION OF UNION CARBIDE AND CARBON CORPORATION - (<3 POST OFFICE BOX P OAK RIDGE. TENNESSEE A T TR B € TR T T T T P T FT T ks ool ol M g . . . i it e m LR ORNL-1835 This document consists of 156 pages. Copy 4 / of 59 copiess Contract No. W-Th05-eng-26 - Aircraft Reactor Test HAZARDS SUMMA.RY REPORT W. B. Cottrell W. K. Ergen | “A. P. Fraas Fo Ba MCQuilkin Je Lo Fleem W. H. Jordsn, Director, ANP Project S. J. Cromer, Co-Director, ANP Project R. I. Strough, Associate Director, ANP Pro,ject' A. J. Miller, Assistant Director, ANP Préject CJAN 19 1955 OAK RIDGE NATIONAL LABORATORY Operated by CARBIDE AND CARBON CHEMICALS COMPANY A Division of Union Carbide and Carbonr Corporation : - Post Office Box P Oak Ridge, Tenneéssee Series A. TTERPE YT o meRresteacermeT . o o e T ST O B TR IR T YT T TR AT T T T T AT e b Bhked adk P P R O e f i1 SN __ i | o b i oo ORNL-1835 ; i Special i 13' : ‘lk:; i b s INTERNAL DISTRIBUTION T : 1. E. S. Bettis F 2., E. P. Blizard G. E. Boyd 3. | 4, R. W. Bussard | - ; ' s 5. T. H. J. Burnett | ' 6. A. D. Callihan | T. D. W. Cardwell 8. C. E. Center ‘9. W. B. Cottrell 10. G. A. Cristy " 1l. S. J. Cromer 12. E« P. Epler 13. W. K. Ergen 4. A. P, Fraas 15. W. R. Grimes 16. H. C. Gray 17. W. H. Jordan 18. C. E. Larson 19, E. R. Mann 20. W. D. Manly 210_‘ Jo Lo b’eem _ 22. F. R. McQuilkin 23. A. J. Miller 2k, K. Z. Morgan 25. Gibson Morris 26. W. G. Piper 27. H. F. Poppendiek 28. H. W. Savage 29. A. W. Savolainen 30. E. D. Shipley 31‘339 R. I. S’brou.gh 3k, J. A. Swartout " 35, G. D. Whitman 36. A. M. Weinberg ~ 37. C. E. Winters ‘ 38-40. Laboratory Records Department 1. Laboratory Records, ORNL RC ol | o i o B i ik s o i idg EXTERNAL DISTRIBUTTON h2-59. H. M. Réth, .:Dizi'e'ctor, Research ‘ o and Medicine Division, AEC, ORO . i 5 - I o 4 4 4 4 4 i 4 [ . . ki ak old Té!.'" woche oo B ke g 0 -M&hm.mim s ik Bl oo ch e B e i Nsk b ol i g i g L S .. ket G bbbl ko ke Bee p e BERREE L sp s M 111 FOREWORD The Atomlc Energy Commission requires ‘that a Reactor Hazards Summary ‘Report be submitted and approved‘by the Advisory Committee on Reactor Safe- guards prior to the operation of a new reactor or the modification of an existing reactor in order to determine, and thus assure, the safety of the Commission's various reactor projects. In accordance with USAEC-OR-8401, Reactor Safety Determinstion, this report describes the hazards that may conceivably be associated with the Aircraft Reactor Test. All possible types of hazards are described as well ss the extent to which these hazards - have been evaluated and considered in the design and proposed operation of the reactor. , PR TRETT R ECRTIT T NI T W S o Y T caabisl L oLial Y S | '- kR . b & e Acmommmws | c The 'bulk of this report vas prepared by the au‘bhors, with the ass:Ls‘c- , __ance of the staff members of the Aircraft Nuclear Propulsion Project who are associated with the Aircraft Reactor Test. In particular, W. R. Grimes and W. D. Manly provided the portions of the report pertaining to chemical . ‘and metallurglcal problems, respectively, and E. R. Mann prepared the mate- - ,_ ~rial on reactor controls. In addition, considerable assistance has been | o solicited from seversl groups outside the project, including Robert F. Myers E‘ and R. D. Purdy of the Oak Ridge Office of the U. S. Weather Bureau and T. J , Bwne‘ct of the ORHL Hea.l'hh Physzcs Divigsion. £ R R o o i e i P wdid 5 s LA ke BddpbE B e ik . . . wonili i Bodo ke ke s bB BB 0 L BEE oo s e ——————— e —— 5. A & . e CA PBABIE OF CONTENTS INTRODUCTION AND wx L C e e THE SHIELDED REAC’EOR ASSEMBLY ,.. Reflector-Moderated Coollng System Pressure Shell . . Heat Exchanger .+ o+ ¢« « ¢ 2 « o + & Pumps . ¢« 5 ¢« « ¢ ¢ s v 6 s s o v s Shield Assembly and Testing . « « . + . . . » * * - - - * 4 & ® & ® & 3 2 B s 3 & 8 » ‘I'HETESTFACIHTYQQQQQtQ'oto s Site Building * a4 & & BT F 4 0+ & ® & 4 ® Reactor Cell # % + & & ¥ 3 3 4 = % Heat Dump System . ., « + « .+ + 4+ . Fill-and-Drain System . . . « « . . Off~Gas Disposal System . . » . . . CONTROLS AND OPERATTON - - . . - . . . fControl Phllosophy e e Scram System .+ « » ¢« v s s o« 0 s . - * & ¢ = ¢ ® & * S'tartup s T o e *7 A 4 % & & & @ ¥ 3 & % Operation Between Startup and Design Point 'De81gn~Polnt Operatlen e e e e e e s * - - . - * s s - REACTQR HAZARDS . e st 'ar‘p_e-/’v Lo, S Radlatlon Dose Levels.il;rfhféf'““?' Typlcal Qperatlona* and.Equlpment Fazlures. “Fuel. Freeze e “Sodium Fréeze i . .« 5 b owe o7 Structural Failures . . . Pump Failures . + c.v o o . . " Electrical Power PFailures . . & w d i '.a‘-_,s_,-.,'.fi' - Fuel Channel Hot Spots o e Eacessive Fuel Feed . , .. Fuel.F111-and—Dra1n SystemaFallure NaK Circuit .Heat Dump. Blower.Fallure. o Fuel, Sodlum or NeK Leak s os el ' Fuel Leak in Core . . . Fuel Leak into. the Heat Exchanger - Other Sodium or NaK Leaks . e Sodium or NaK Fires ., . . . . . 3 % 4 - & * * - * + » - » s y & % * & & . L 4 - - - <& » * T e 8 s e .. - - - 3 » » - ® & & Page R T YR TP T T T T P gy T TTETERR A T T R T . E E k i E t b i ] 1 - - - | PR | Page E ] ' - | Acc1dents Caused by the Natural Elements c s s e o .A. ... . . 40 f” q - Flood and Earthqueke « o « o o o o o o + s o s o o s o o « 40 ¢ A:;' Wlnds‘bom » - - - - - - - * .I » » .. - * - - » .. .‘ » L4 - » ll'o - .-;E 4 s , . . . . AR s T ‘ S < 't-‘k" ! Nuclear Ac01dents CauSLng Rupture of Pressure Shell « « o« » « hO ; s Fuel Prec1pltatlon in Core } ; . .f{'..} e e e e R Fuel ln MOdEI‘a'bOI' ¢« o ¢ 5 2 = » '. * & ° & + s @ .o . & » - 1"2 o ;f Penetrabillty of Reactor Cell by Pressure Shell Fragments . 43 i Sgn Acc1dents That Might. Bupture the ‘Reactor Cell ;T;“; ; . e .. b3 " E:” L"l. Hydrogen EXPlOSlon e & o e o & *» » -‘ . ® = o .. o ‘v' ‘- .r—.-."‘ * ;'o h‘ll' 7Y Demage from High EXPLOSIVES « v « « o o + o s o o o o o « o 45 ' Effectiveness of Reactor Cell in Containing Hazards . . . . L6 | Comparison of Various Reactor Assembly COntalners_.u._. . o 47 6;[ DISEERSION OF AIRBOBNE ACTIVITY . e e e .'.’. R 1 : Radlation Tolerances e s o o o .-...fl}'. « o ;u;-;%:w. e « « 50 | E Dlscharge of Activity Up the Stack c s e s s e s e e s s e Bl E Norma]. Opera‘blon « & & & ® s s s o_ . o o .— + - .o“’ e s s & 51 F‘% Operation Without Heating Stack Air . « ¢« « + ¢ s ¢ « o « - 52 Operation With No Stack Air FLOW « o « « » o = o o « « » + 58 Operation With Holdup System By-Passed « « « « « o« 2 s o & 59 ‘e - .0 - . 59 Smoke TraCklng the Stac:k Plume o o * e e . ’ - Discharge of Activity Following a Dlsaster e s e e e e 60 Minimum Heat Liberated in a Dlsaster e e e e e e e e . 6L Height Of CLOUA RISE « o o o o o « o o o o o o ¢ o o s « + 62 Fission Products in the Hobt Cloud + « o + =« « « o« o o o » - 63 Internal Exposure from the HOL CloUA « « o o o o o o « o o« Ol } External Exposure from the Hot Cloud « « « « o « « o « o » 65 b Rainout From the Hot CLOUA « + « « « o = oo o o o o o s + OF - "Exposure from & Cold CLoUG « + o o« o s o o o o s o o o o« 69 ; Beryllium HAazard .« o o o o o « o o o o o o s o s o« o o o = 69 § APPENDIXES o E A. CHARACTISTICS OF STTE = « o o o o o« o o o o o s o o o o s o oo Tl Meteorology and Climatology « « o« o « o o o o o o o o o o » o (L1 Distribution of Population « ¢ « « ¢ ¢ o o o ¢ o o o o o o o o (2 Vital Industries and Installations « « + o o o ¢ o o o o o o« T2 ' Geology and Hydrology of Site 4 e e me e s e e e e s e T8 'Seismology Of A2 « «o o o o o o s s o s s o o o o o o o o= T8 el Do B gRboede o boigs e chekeh e @ w Efnadizgmal o L T e s s Bl | e e pie B. c. D. - E. Fo -» & ..0'.9 0.’ - 0_'0 o o & o ,_',,‘_.Bas:.cFormulas ® o 6 s e e s e s 0 s s e e b o s s S UMEEBOA ¢ . e s s v et e s e e e h e e e e e e _"Calcula'hlenal P'Focedu.re o 6 o 8 * o € 8 = & e e HEAT RELEASED IN CHEMICAL REACTTONS AND RESULTING TEMPERATURES AND PRESSURES « + + « « o+ . Reaction of 70 Pounds of Sodium with Adr . Reaction of TO Pounds of Sodium with Water . . . . . Reaction of 930 Pounds of NeK with AiT + + ¢ o « « Reaction of 930 Pounds of NaK with Water . . . . « . Reaction of 1200 Pounds of Fuel with Sodium s v s e * - - e * » + - - Reaction of Sodiwm and NaK with Shield Water in Nl‘bregen Heat Absorbed in Water « « o ¢ « o = o « o o & o & Heat Absorbed in Nitrogen . « « « o ¢ ¢ o o o o o Reaction of Sodium and NaK with Shield in Air .« . . . ME?IAILURGIANDCHE&ESTRY...o............, COrros:Lon of Inconel by the Fluoride Fuel e e s s e o Mass Transfer in the Sodium-Inconel-Beryllium System . Chenmical Interaction of Fluoride Fuel and Na or NaK . Reactions From & Heat Exchanger Lesk « o o o « o o Reactions From a Core Leak « + o ¢ o o ¢ o o » o = Conclusions'..........e..,....._.... E)EEREME NUCIEAR ACCILDENT - ANAI.YTIGAL SGIMIQE c es e e GeneralEq_uat:.ons e o o o o s & 2 & & & e s 6 « o a o Veriation ef}?arameters R I EXTREME NUCLEAR ACC‘EDENT NUBERICAL semmfi aae e EFFECTS QF A NUCLEAR AcCIDE].\IT on REACT@R STRUCTURE";"; s e Stress Calculatt.ons 2 e s e o s s o e s o o 1_._',._ o o Destruc‘ba.veness of Pressure Shell Fragments . « o « Pressure Rellef Mecham.sms s s e s e ° e s e EXFOSUREHAZARDCALCULATIGNS e e '_, ; .. e :"Crlterla o o .‘. e o o o O & o6 & & o & 06 ® o6 ° 6 & 4 BV O © b & 6 & B & B . ’ & ° * e * > » . - * . . . - - ¢ > . & & » e e ® Page 102 106 119 123 126 127 131 131 131 133 13h 145 e i R R LIST OF FIGURES | | e ShaEn S A e I W W LD TS 3 ST acdai A LT SEE LTSIl e T Egaldy e ' Page Noe Title 2.1 GO-MW Reflector Moderated Reac'bor IR IR . * . 2,2 Horlzon‘bal Section T]rrrough Pump Reg;.on of Reactor . A T 2.3—-" Schematic of the Alrcraf'b Rea.ctor 'I‘est ... !". .. S T 1 % 301 o ART Faclllty - - L ] . . @ » » * * . * 70 - - » - ‘Q . - 0. * . “ * * * 16 3.2 Plan "of the ART Bulldlng VO EUTT e ar S Rime e Framoral o AR ._3-3 | Elevat:.on of the ART Bulldlng . .. .' e e e e e e ... .. .18 | '3.‘1}'5 _Bea.c'tor Assembly Cell wi’ch Water-Fllled Annulusm.: . S 20 61 He:.ght of Plume R:Lse from AR'I‘ St.ack e e e e e | A.l Map of Count:.es Surroundlng Oak R:Ldge Area. .. .. . L. 7 D.1 Reactor Power, P, Relative to the Initial Poviér, P, VerS'G.S TimeDuringExcHrSiOfi e o » . & & & = e 4 & & & & & » ‘101 D.2 Effect of Parameters on Maximumirfléac'tor Power . . . L .. . . 104 E.l Nuclea_r Excursions for Fuel Deposition in Core Increamng k_pp at Rates from 5% to 40% Ak/k per Second .« « o« o « o o o 109 E.2 Nuclear Excursions for Fuel En‘ber:.ng the Moderator : Cooling Passages and Increasing k off at the Rate : 'of60%Ak/kperSecond................... 110 E.3 Fuel Vapor Pressure as & Function of Temperature . « . « « « « 111 Wit D g o s l:l 5.1 5.2 543 6.3 6.4 605 6"_5" 6.1 6.2 Ry LIST OF TABLES Title Ajircraft Beactor'Test DeSign Pata o o ¢ o ; * B o @ Radiatidn Bose Levels ; e u'e & . °7.—, = o 8§ @ .'o ® Estzlmted Reactlv ities from Ma:jer Changes in Rea.c'ter Seurees Of Energy s 8 o n.o ¢ o b e © © » olo e ® e+ & Comparison of Hazaé.fd Dats for Several Ty:pes of : i o (.,.a"‘ Reactor containers ) ® 8 s . & o ° - * o o -®. O - & ‘,. iy o - v d - . v ,. i ‘\{ Fad {“{ ; / > > Concentration of‘ Gases Released frem Stack :Meteorologlcal Parameters for Determlning Ground o Ground Concentratmn of Gases Beleasgd from Stack (without decay cerrectmn, 6 7107 efm air). Ground Concentratlon of Gases Released from Stack (with decay correction, £-0.2 36.7* 107 cfm air) Ground Concentra‘bmn of Gases Released from Stack (w:Lthout decay correc’ca.on, 3e 105 cfm air) . . » Ground Concentration of Gases Beleased. from Stack (W:Lth no &ecay correction and no air :E‘low) Lim:Lt of VJ.Sl'blli'by of 8 Smoke lene from a 50—gph > Generator e o o o @ o & ® B o 2 o v @ o”.:o o‘a . Rural Population in the Surroundmg Counties > " .6 Total Integra'bed Internal Doses from ot Cloud | s * » W1 __”Perso el W:Lthln the AEC Restrn.cted Area . s | - | "'Metenrologlcal Pa.rameters fo:r To’cal Reactor Tragedy . ' atffllght o“o o.n-c{uig _EEJiGround Exposure Fe]_j_owing Ralnout of the Hot Cloud Total Integrated Internal Doses from Cold Cloud . . » PoPulatlon of the Surroundlng Towns '. o e s e 5 b w L sy e :.‘:::; N L ix ~ Page 3k sl 25 56 51 58 65 .67 68 ,,_,70 73 5 TN T Ty T T BT A T e T TDST TR Sl e . 4 -4 3 - L LIS'I‘ OF TABLES (Cont'd) No.'“ AL E.1 E.2 E.3 E.5 E.6 | F.2 F.3 6.1 G2 . G.3 G.b G.5 fNuclear Excur51on.CalcuLatlon for & Hypothetlcal Case - Title Vital Industrial and Defense Tnetallations in BO‘MileBadiuSOOonnoa»ooouoccv_oooo Basic Thermodynamic DALE « o « « « o o o o o o o o o o o i Fuel and Reactor Properties at 60'Mw e e e i e e e .. .1Nuc1ear Excursion Calculatlon,for a prothetlcal Cage Involving Fuel Deposition in the Core to Give an Inltlal Rate of Increase in k off of 5% per second . . ‘Nuclear Excursion Calculation for a Hypothetical Case - Involving Fuel Deposition in the Core to Give an Initial Rate of Increase in k eff Involving Fuel Deposition in the Core to Give an - Initial Rate of Increase in k ot of 20% per second . . .Nuclear Excur31on Calculation for a Hypothetlcal Case - Involving Fuel Deposition in the Core to Give an Initial Rate of Increase in k_., of 40% per second . Nuclear Excursion Calculation for a Hypothetical Case Involving Fuel Entering the Reflector Cooling Passages to Give an Initial Rate of Increase in k o, of 60%/sec ‘\ Key Dimensional Deta for the ART Pump-Heat Exchanger-'w R Pressure Shell Assenbly « ¢ o o € o o o o 6 o o o o o Dimensions of ART Detail Partis ; . . ; ; C e e e e e s Stréngth.Data for Inconel Tested in a Fluoride Mixture . FiSSion Yields 5 © | o - - ° - * s ° -0 vo. © - ¢ Q o * ,’ 0. ‘ Selected Isotope Mixture « « « « v o o e e e e e . Tnitial Dose RBLES « o o« o =« o o o o 5. 2 o o &+ o o o o o TO"S&].BOSGS.V.‘. o'oooorooo;.c.c.oo.»ooolcoo Dose from Six Selected Isotopes . . c e et e s » ... of 10% per second . « o - 'Page 76 108 » 113 114 115 116 117 120 121 153 137 . 138 140 _11+é 1hk \ TR T w: . % s S e r—— T TR AT T T e W g & g . i S /‘.y)"’ - o :r( £ A ey fiiv-fl-&"- - ATRCRAFT REACTOR TEST HAZARDS SUMMARY REPORT l, INTRODUCTION.AND SUMMARY The successful completion of g program of experiments, includlng the Aircraft Reactor Experiment (ARE),l has demonstrated the high probability of producing militarily useful aircraft nuclear power plants employing re- flector—moderated circulating-fuel reactors. Consequently, an accelerated program culminating in operation of the Aircraft Reactor Test (ART) is under way. In order to adhere to the compressed ‘schedule of the aceelerated pro- gram, it is essential that the Atomic Energy Commission approve the 7500 Area in 0Oak Ridge as the test site by February 15, 1955. This report summarizes the hazards associated with operating the contained 60-Mw reactor of the ART at the proposed Qak Ridge test site. Descriptions are gilven of the reactor, reactor cell, test site, reactor controls, end.cpérgting plan, prior to presentation of the hazards consider- ations. The hazards are classified into three major categories: 1) acci- dents with an apprec1able probability of occurring, 2) accidents causing rupture of the pressure shell, and 3) accidents causing rupture of the re- actor cell,’ ‘Category (1) accidents would involve minor difficulties in the 1ntegr1ty of the reactor system and would not result in injury to operating personnel or the surrounding population. Category (2) accidents, which are extreme nuclear excursions, might be caused by and can gause major breaks in the rédctor assembly, but due to the presence of the reactor cell which would remain 1ntact there would be no injury to operating personnel or other peoples ~ The causes of category (2) accidents are described in a general man- ~ per since it has been impossible to date to describe a specific series of events that would lead to an extreme nuclear excursions In the total facil- ity destruction; category (3), the reactor and the reactor ‘cell would be | ruptured to release the accumulated fission prodacts, but this could ‘only be accomplished by means of extremely clever distribution of large guantities - of explosives by a saboteur or by a large aerial bomb. In the analysis the _-most severe case has been presemed° namely, that cemplete volatilization of ‘all of the fuel would occur. This is highly improbable since any such acci- ;f;dent would require a large amount of heat at a high temperature, and its oc- > e»;currence_githout dispersion ‘of ‘much of the fuel in particulate formcseems to be '_@~unliké' . In addition, quantities of other materials that would also be ProPuls1on Progect, ORNL—lEBh (June 2, 1952) Program of the.Aircraft'Nuclear e T O T R o (T I T e O P~ " 277 7 et e Bt R o e Do The reactor of the ART is t0 be a 60~Mw reflector-moderated circu~ "latlng-fuel Eype whose basic design is ‘suitable for reactors to be used in aircraft.§ The size and weight of +the reactor and shield will conform ‘. with aircraft requirements, and, insofar as possible in the limited time . available, the design of the 1mportant components will be based on concepts 'satlsfactory for airborne applicatlonso ' Operatlon of the ARE demonstrated that a hlgh temperature c1rculat1ng- fuel reactor could be built and operated and that materials and machlnery for “satlsfactory operatlon at elevated temperatures ‘had been developed° It show- ed that the predicted large negative temperature coefficient of reactivity ' and the resultant self-regulatory characteristics of the reactor could be V_echieved. In addition, it was found that most of the Xel3> was removed. Trom the fuel 1nto the gaiiglanket space in the pump so that the steady-state con- 'centration of the Xe was only 3% of the normal equillbrium value. The neW'princ1ple of design introduced in the ART consists of circulat-'M ing the fuel through the reactor in a single, thick, annular passage and achieving the major portlon of the moderation with a beryllium reflector which will also serve as an important portion of the shield. With the re- sultlng reduction’ 1n shlelding requirements it has been possible to design the ART reactor in such a way that the entire fuel system is contained with- in a suff1c1ently small, shielded volume to provide a low-weight shield and a useful alrcraft reactoro_ The purpose of the Aircraft Reactor Test is to 'valldate +the methods ‘of construction and the predicted operatlng character— istics of such a reflector-moderated cireulating-fuel reactoro A reactor power “of 60 Mw was selected because 1t is agproximstely the ' power that must be reached to demonstrate that the engineering problems are solved and that the operating characteristics are satisfactory for the high- er powered reactors to be used in high-altitude supersonic strategic bombers. In addition, & reactor with a power level in the 60-Mw range will provide sufficient power to fly radar picket ships, patrol bombers, and other desir- gble aircraft. For a power level above 60 Mw, the cost appears to be di- - rectly proportional to the power. Also it does appear that a thoroughly satisfactory 200-Mw reactor can be built more quickly by first building a 60-Mw reactor and then following it with a 200-Mw reactor in which the benefits of the experlence gained with the 60-Mw reactor will have been incorporated. , The design of the ART envi51ons an essentially spherical reactor 1n_’ ' which the berylllum moderator will be lumped in a central island and in an outer annulus (see Fig. 2.1, Sec. 2, this report). Two centrifugal pumps, arranged in parallel, will circulate the fuel downward between the inner beryllium island and the outer beryllium reflector and out of the bottom of the core. The fuel will then turn and flow upward through the heat ex- changer region, which is around the spherical core. From the heat exchanger region the fuel will return to the pumps and will again be discharged down- ward into the core. The reactor heat will be transferred in the heat - " B, K. P. Freas and KA. W. Savolainen, ORNL Alrcraft-Nuclear Power Plant 'Désigns, ORNL-1721, May 1954 (issued Nov. 10, 195L4). s T T T Y - ey ' , e ""aa =3 exchanger from the fuel to the secondary COblant;(Nafi) The fuel will be one of a number of fluoride salt combinations which have been shown to have acceptable physical proPertiesg 'In particular, the NaF-ZrF, -UF, fuel mixture 1s known to be satisfactory. The mixture NeF-KF-LiF-UF) has some very de- sirable properties and is unfler intensive investigation. ~ Quite a variety ef shielaing arrangements has béen censmdered for the ART. The most promising seems to be one functionally the same as that for an aircraft requiring a unit shield; namely, a shield designed to give 1 r/hr at 50 £t from_the center of the reactoro Such a shield is not far from being both the lightest and the most compaet that has been devised. It will make use of noncrltical materials that are in good supply, and it will provide use- ful performance data on the effects on the radiation dose levels of the re- lease of delayed neutrons and decay gammas in the heat exchanger, the gene- ration of secondary gammas throughout the shield; etc. While the complica- tion of detailed instrumentation within the shield does not appear warranted, it will be extremely'worthwhile +to obtain radiation dose level data at repre- sentative points around the periphery of the shield, particularly in the vicinity of the ducts and of the pump and expansion tank region° Several arrangements have been considered as means for dlsposing of the heat generated in the reactor. The most promising of these is one that re- sembles a turbojet power plant in many respects. It will employ radiators essentially similar to those suitable for turbojet operation. Conventional axial flow blowers will be used to force cooling air through the radiaters. This arrangement will be flexible and as inexpensive as any arrangement de- vised. It will give thermal capacities and fluid transit times essentially the same as those in a full-scale aircraft power plant. It will also give some very valuable experience with the operation of high-temperature liquid- to-air heat exchangers that embody features of construction and fabricating techniques suitable for alrcraft use.o In an effort to minim:ze the llkllhood of important treubles developing during the course of the test, an extensive series of cemponent development tests has been initiated. These tests have been designed to establish sound technigues for the fabricatlon ‘of pumps and heat eéxchangers and to provide detail design information on such factors as clearances, etc. The operating experience’ gained in the course of these tests should prove most helpful in ‘minimizing 0perating'trodb1es ‘with the ART and in diagnosing such troubles as may develop. These component develcpment tests include experiments with - a hot (high-temperature) critical agsembly which will consist of the pump, header tank, and core system envisioned for the full-scale reactor. The ex- pensive pressure shell and heat exchanger will not be included in this hot eritical experiment. _ ~flw¥%Experience with the ARE has indicated the advisability of building, in _1on to the crltieal assembly, a complete reactorwpump-heat exehanger- , ‘mental engineering 1aboratory. Tn a structure as eomplex ag this it is o T T T T TR T T T TR TN pr e e, e e i NG S g fl m;‘i‘ o i = felt that there will probebly be a number of mechanical problems of construc- tion. Rather than go to extreme and awkward lengths to try to correct these .'by'rew0rking the first unit, it is to be ‘built as expeditiously as possible - and operated 51mply as a high-temperature component test with no flSSlenable mater:.al° ‘The experience gained in fabrlcatlng and shakedown testing this first assembly should not only prove invaluable in construction and operation of the’ second assembly'for use with fissionable materlal, but should actually : lead to an earller operatlng date for the ART. Y & \,'-w”: “5‘1 S: As the 6551gn of the f ma. jor hazards would be much less serious than had at first been presumed. f Also, the use of circulating fuel with its high negative temperature coef- ficient gives & reactor in which a nuclear exploS1on seems almost out of the question. In order to operate the ART at Oak Ridge uhder the safest possible conditions the entlre reactor system, with the exception of the NaK-to-air radlators, will be enclosed in a cell conszsting of an inner tank within which the reactor assembly will be installed and an outer water-filled tank. The cell.vlll thus provide a water-filled annulus around the reactor assembly. ~ The very compact installation envisioned results in’'a Very low investment in sodium and NakK (about 1/20 of that required for the KAPL-SIR reactor designed - for the same power level). A relatively small amount of energy would be re- '1eased.by reactions involving the llquld metals, and therefore a corre8pond~ ingly small-dlameter cell can be used. | , ?_7, Ths cell W1th a water—fllle.“ann, us W1ll\be deqnate to absorb the 'amounts of energy that could be released in an extreme reactor catastrophe. It will be impossible for a fragment ejected from the reactor assembly by an explosion to rupture the inner tank of the cell because the pressure shell surrounding the reactor has been deliberately designed to yield at a pressure of 1000 psi and the maximum velocity of & fragment ejected at this pressure would be substantially below that required to penetrate the cell wall. ' The 60-Mw reactor test unit was 6381gned orlginally to be operated ‘at the National Reactor Testing Station (NRTS) at Arco, Idaho. It was envision- ed that the reactor could be pretested at Oak Ridge and then shipped to MRTS for the nuclear tests. waever, a survey disclosed that construction and operation at NRTS would require at least six months longer than at Oask Ridge. Delays would be occasioned by conducting a construction operation 2,000 = miles away and any small difficulty that might arise in reactor operation would be likely to introduce a major delay if that difficulty were not fore- seen and plans to cope with it made in advance. No aelay'would be ocea - sioned by construction of the small reactor assembly cell for use at the Oak Ridge site, and approval is being requested from the Atomic Energy Commission for operatlon of the Aircraft Reactor Tegt in such a cell at thg Qak Bldge. . site. . Design data for the ART ere presented in Table 1,1. o a e A -5- TABLE 1.1 ATRCRAFT REACTOR TEST DESTGN DATA Power Heat, meximum (kw) Heat flux (Btu/hr/ft Power (mex/ avg) Power density, maximum (kw/liter of core) Specific power (kw/kg of fissionable ~ material in core) : Power generated in reflector, kw Power generated in island, kw | Power generated in pressure shell, kw Power generated in lead layer, kw. Power generated in water layer, kw Materials Fuel Fuel Jacket Moderator Reflector Shield Primary coolant Reflector coolant Secondary coolant Fuel Sys@em Properties Uranium enrvichment (% I | Critical mass (kg of ‘ Total uranium inventory (k? of U255) Consumption at meximum power (g/day) "?flfDe31gn llfetime (br) :Burnup in 1000 hr at maximum pover (%) - Fuel volume in core (ft ) | 'Total fuel velume (ft ) Heutron Flux Density (avg)&' Thermal, meximum (n/cm . sec) Thermal, average (n/em”e sec) Fast, maximam (n/emS-sec) Fast, average (n/cm“-sec) . Intermediate, average (n/cm~esec) 60,000 Heat transported out by eirculating fael 2:1 1400 4500 2040 600 210 132 (b NeF-ZrF-UF) , 50-46-4 mole % or NaF-KF-LiF-UF), 11-42- ‘4h-3 mole % Inconel Berylliunm - Beryllium Lead and borated water The circulating fuel Sodium NaK 954 - 13.5 30 80 - 1000 11 2.96 5061'{' R T e A e T g W e T ; T R T T W O T T T W T TR R i ¢ b k E o e e © e b AR 3. el B AL . e iR i e L L o . o s RS S Barr Yekrs ot S S e S bRl B e o P ol S e il Fetbadial it st Sl e st S (e e f e s e b S5 L 2 R S G s . : it e e b SRl W o e .».,AW SR b Bt i e s . R R S e R S s S e et s b RS i e . R o | SRR gt S e e s L Cp b gl L 6= ~ Comtrol YT S e e B Shim control | One rod o 5% Ak/k Rate of withdrawal - 3.3 x 107* Ak/k-sec Temperature coefficient -5 5 x 10-2 (Ak/k)/oF Clrculating Fuel—Coolant Systems Fuel in Core Li Fuel Zr Fuel Maximum temperature, °F 1,600 1,600 Pemperature rise; °F 400 400 Flow velocity, ft/sec 7 7 Reynolds nnmber | ' 170,000 85,000 Fuel-to-NaK Heat Exchanger o Li Fuel Zr Fuel NaK Coolant . Maximum temperature, “F 1,600 1,600 1,500 - Temperature drop (or rise), QF koo 400 400 '“,Pressure drop, psi 35 55 - 50 Flow'rate,'fts/sep 2.7 2.7 12.6 Veloeity through the tube matrix, - Pt/sec . 8 8 36 - Beynolds number 4,600 2,300 180,000 Coolzng System for NaK%Fuel Coolant Maximum air temperature, °F T50 Ambient airflow through NaK rediators, cfm 300,000 Radiator air pressure drop, in H,0 10 Blower power required (totel for % ¥lowers), hp 600 Total radlator inlet face area, £t2 6k Cooling System for Moderator , - o Meximum temperature of sodium, ©F 1200 Sodium temperature drop in heat exchanger, OF 100 NeX temperature rise in heat exchanger, ©F ' 100 Pressure drop of sodium in heat exchanger, psil T Pressure drop of NaK in heat exchanger, psi T Flow rate of sodium through reflector, ft /sec 1.35 Flow rate of sodium through island and pressure I shell, ft7/sec 0.53 Flow velocity of sodlum through reflector and island, S ft/sec 30 Beynolds number of sodlum in reflector and island 170 OOO System Volumes and Pump ] Data Li Fuel Zr Fuel Na Coolant max Coolant Number of pumps 2 2 b ~ Pumping head, f% ' 50 50 250 280 Flow per pump, gpm . - 600 600 130 - 1300 Pump speed, rpm | 2850 2850 %300 125 Pump power per pump, hp 40 65 16 100 Ve et g m——— T t a A ik T T——n - a1 o e Core diameter (in.) - | 21 Island diameter (in.) 11 " Fuel region thickness (in.) | L Reflector thickness (in.) o : 12 ‘Shield thickness, lead; (in.) | T Shield thickmess,water (in.) - 3L -8- 2. THE 'SHIELDED REAC_TOR ASSEMBLY - g The reactor is to be of the c1rculat1ng—fluoride-fuel, reflector- o mflderated type. It will employ sodium-cooled beryllium as the reflector- moderator material and is designed to operate at 60 Mw. The reactor assem- - bly will include the pressure shell, reflector, fuel and sodium pumps, and heat exchanger assenblies. The basic design is shown. 1n‘Flg. 2.1, a vertical section through the reactor. A series of concentric shells, each of which is a surface of revolution about the vertical axis, constitute the mejor portion of the assembly. The two inner shells surround the fuel region at the center (that is, the core of the reactor) and separate it from the beryllium island and the outer beryllium reflector. The fuel circulates downward and outward to the entrance of the spherjcal-shell heat exchanger that lies between the " reflector shell and the main pressure shell. The fuel flows upward between the tubes in the heat exchanger into the two fuel pumps at the top. From the pumps, which operate in parallel, it is discharged inward to the top of the . annular passage leading back to the reactor core. The fuel pumps are sump- '~ type pumps with gas seals. A horizontal section through the pump volute -~ region is shown in Fig. 2.2.. A schematic diagram of the reactor system is - shown in Fig. 2.3. : :RefleéfierModerator Cooling System The reflector will be cooled by sodium circulated by two pumps at the top of the reactor. The sodium will flow downward through passages in the beryllium and back upward through the annular space between the beryllium and the enclosing shells. The central beryllium island will be cooled in & similar manner, except that the sodium will leave the bottom of the is- land to be returned to the top of the reactor through cooling passages in the main pressure shell. The sodium will return to the pump inlets through small torrodial sodium-to-NaK heat exchangers around the outer periphery of the pump-expansion tank region. The sodium pump and heat exchanger sub- assemblies will be positioned on either side of the fuel pump volute region. The pipe from the sodium pump discharge will make & slip fit into the re- flector sodium inlet tube. The leakage through this slip fit into the sodium return passage will simply recirculate with no penalty other than a small increase in the required pump capacity. Pressure Shell The Inconel pressure shell will constitute both the main structure of ‘the reactor and a compact container for the fuel circuit. The design has been modified somewhat from that shown in Fig. 2.1 to make the shell con- tinuous through the vicinity of the headers and thus give better contimuity of stress flow and & minimum of welding. Blisters made of l-in.-thick plate uelded to the outer surface of the shell between the NaK pipes will serve ‘both to reinforce that weakened region and to provide for sodium flow up through the shell. The imner liner assembly of the pressure shell will con- 51st of a 0. 75-in.-thick Inconel shell, a 0.125- in¢-thick hot-pressed Bhp T TR T T TR T rrrmerT e o et on sh s < . . GEERRS L L e ok w-EECREY Dwe. 22342 IMPELLER (STRESS=500 psi) UPPER PUMP DECK PUMP DECK TUBE HEADER SHEET (STRESS =1250 psi} S S e wt, B R TUBE WALL : {STRESS = 450 psi) * REFLECTOR SHELL Na-TO-NaK HEAT EXCHANGER B'® LAYER - INCONEL. JACKET Be REFLEGTOR HEADER FOR Na ,TO ’ REFLECTOR — 7] ‘Be ISLAND _ Lo MODERATOR _ OUTER CORE SHELL COOLING TUBE (STRESS =400 psi) . \ \ INNER CORE SHELL COOLING TUBE CONNECTER TUBE BUNDLE SPACERS ‘TUBE BUNDLE JACKET FOR 8'0. LAYER {STRESS = 200 psi) TUBE WALL ~ TRESS = 50 psi ) - RS b} B LAYER PRESSURE SHELL LINER Na PASSAGE PRESSURE SHELL (STRESS = 625 psi} . REFLECTOR Na o OUTLET HEADER Ny . Na 5= 1250 psi) TUBE HEADER SHEET {STRESS =600 psi) NaK OUT ¢um Fig. 2.1, 60-Mw Reflector-Moderated Reactor, SRR Y L T e T e T R Y P TR T T T Cr e T TR T wer oy 3 -2 Mu.hm i 3 .Lm".“ h.“'-m k. M’L ".. . ‘fi.m A ikmmm. ekt un-AMsh.k; e i i, i, bl o s ot : . DWG E-179868 B o . —FUEL PUMP IMPELLER INLET o ; | : FUEL PUMP VOLUTE A N / . / ] T Na PASSAGE IN ISLAND 1 f L Ne TO REFLECTOR o \ F 'l \ Na FROM REFLECTOR - < - NaK INLET Na- TO-NeK HEAT EXCHANGER NoK INLET -, ‘Na RETURN FROM ISLAND AND PRESSURE SHELL o r & ‘ y - i e T TR R T T g T TEREE T WRr T T T AR T TR ¢ " Y T WW'W" R O T T =hi- sz MODERATOR PREHEAT BURNER . 8000 cfm BLOWER ornL EDWG 3091 ' NaK-TO-AIR RADIATOR\ 6AS L L « CTO STACK: ~a— G > - , = ’ .’ DISCHARGE TO STACK & : ) S £ NeK PUMP L NaK PUMP - . | T : T - L NaK_PUMP DRAIN WATER MAIN - — _ . ; NaK FILL A _ PREHEAT DOORS PREHEAT DOORS N/ LUBRICATION- OIL DRAIN TANK Y - \ COOLER \ \f / MAIN NoK-TO-AIR MAIN NakK-TO-AIR RADIATOR — RADIATOR NaK BY-PASS FILTER / ~ ~ \ \ PREHEAT } NoK BY-PASS FILTER 1 Na -TO-NaK r 0. S/ HEAT EXCHANGER ) 5 = ] k2] 2 — - FUEL-TO-NoK j j | HEAT EXCHANGER FUEL-TO- NaK , : - /'HEAT EXCHANGER ‘ -\ GAMMA SHIELD 18 —J] m Ho0 —- —1 G - . Eo N J AIR FROM ] 5 RADIATORS B 83 o= Sa He § e B gm He é C \ g s o o — . — a g 0 I S SHELD S 4 3 | = FILL AND DRAIN ° Oy 7 4_____% =Y R TANK ~ FUEL - iz &z ~ P -1 z 52 > - 5 5 ~ s EZ g 3 > - g g N 2 = 2= N = & N\ / xg = &g \ / < / 2 & & N/ o N/ g g° \ V - = - v = AIR IN g FILL AND DRAIN 1 AR N o TANK — Na Fig. 2.3. Schematic of Aircraft Reactor Test. T TR T T T T hos s B [T S T T T e sl Seiaiid sl iitaba: L i B, A B i S S it _,. i G b bk i i ,,r\ A k\,‘} Ty '-HEat EXChanger ’ -12-_ layer, and an inner 0 062-1n.-thick Inconel can. Heat generated in the pressure shell by the absorption of gammas from the fuel will be removed by sodium flowing between the outer surface of the liner and the iunner surface of the pressure shell so that the pressure shell temperature will be held close to 1200°F. A 0.062-in.-thick gap will provide ample flow _'passage area for sodium to flow upward from the bottom of the island to the top of the pressure shell. Transfer tubes there will direct the sodium from the outer surface of the pressure shell to the sodium-to-NaK heat Jexchanger 4inlets. The hot-pressed boron carbide blocks will be diamond- shaped, with 60-deg angles at their vertexes, and they will have rabbeted edges. This design mekes it possible to cover a spherical surface with a ‘single block size and shape. To facilitate welding in the final assembly, the pressure shell will be split circumferentially at both 1°S and 35°N latitudes. By splitting the liner at 1°S5 latitude, it will be easily ac- cessible for weldlng, and the upper portion of the main shell can be lowered 1nto place for the final'weldlng operafilons. . ' The spherlcal—shell fuel»toANaK heat exchanger, which makes possible the compact layout of the reacbor-heat exchanger assembly, is based on the ‘use of tube bundles curved in such a way that the tube spacing will be uni- form, irrespective of latitude.l The individual tube bundles will terminate in headers that resemble shower heads. This arrangement will facilitate | assembly‘because 8 large number of small tube-to-header assemblies can be made leak-tight much more easily than one large unit. Furthermore, these tube bundles will give a rugged flexible construction that will resemble steel cable and will be admirably adapted to serviée in. wniflh large amounts of dlfferentlal thermal expan31on must be expected. ’ _ Pumgs _ " Two fuel pumps and two sodlum pumps are located a$ the tep of fihe reactor. These pumps are similar, but the fuel pumps have a larger flow capacity, In addition to pumplng, the fuel pump will perform several other functions. Most of the xenmon and krypton and probably some of the other fission-product poisons will be removed from the fluoride mixture by scrub- '~ bing it with helium as it is swirled and agitated in the expansion. tank. The high swirl rate in the expansion tank is also desirsble in that the centrifugal field will keep the free surface of the fuel reasonably stablé: in maneuvers or in "bumpy" flight. The expansion chamber will also serve as a mixing chamber for the addition of high-uranium-content fuel to the main fuel stream %o enrlch the mixture and to compensate for burnup. Fluid will be scooped from the vortex in the expan81on tank and directed into the centrifuge cups on the backs of the impellers. Since considerably more fuel will be scooped from the vortex than can be handled by the centrifuge, the - eXCESS'Wlll be dlrected upward 1nto the swirl pump where it w1ll be accelerated \\\\\\ l. A. P. Fraas and M. E. LaVErneg Heat Exchanger Design Charts ORNL~-1330 (Dec- 7: 1952). T 7 \ ailo it b R R il MU A, |GG L s RN G e - 213- and returned to the expansion tank. There it will help maintain a high rotational velocity. A slinger located on the pump shaft sbove the swirl pump will prevent fuel from splashing up into the annulus around the shaft below the seal. A number of other special features have been included in the pump design to adapt it to the full-scale reactor shield. The pump has been de- signed so that it can be removed or installed as a subassembly with the impeller, shaft, seal, and bearings in a single compact unit. This assembly will fit into the bore of a cylindrical casing welded to the top of the re- actor pressure shell, A 3-in. layer of uranium just above a l/2-1n. layer of’th around the lower part of ‘the impeller shaft will be at the same level as the reactor gamma shield Jjust outside the pressure shell., The space be- tween the bearings will be filled with oil to avoid a gap in the neutron shield. The pumps will be powered by d-c electrlc motors in order to provide good speed control. Shield The shield for the reactor has some characteristics that are peculiar to this particular reactor configuration. The thick reflector was selected on the basis of shielding consideration. The two major reasons for using a thick reflector are that a reflector about 12 in. thick followed by a layer of boron-bearing material will attenuate the neutron flux to the point where the secondary gamma flux can be reduced to a value about equal to that of the core ‘gamma radistion. This thickness will also reduce The neutron leakage flux from the reflector into the heat exchanger to the level of that from the delayed neutrons that will appear in the heat exchanger from the circulating fuel. An additional advantage of the thick reflector is that 99% of the energy developed in the core will appear as heat in the high-temperature zone included by the pressure shell. This means that very little of the energy produced by the reactor musfi be disposed of with a parasitic cooling system at a low temperature level. The material in the spherical-shell intermediate heat exchanger is about 70% as effective as water for the removal of fast f;neutrons so it too is of value from the shielding standpoint. The delayed “neutrons from the cmrculating fuel in the heat exchanger region might appear to pose ‘s serious ‘handicap.’ ‘However, these will have an attenuation length ' ‘much shorter than the corresponding attenuation length for radiation from thexeore.,'"hps, from the outer surface of the shield, the intermediate heat ’chaager will appear ‘as a much less 1ntense source of neutrons than the 'more’ deeply buried reactor core. The fission-product decay gammas from the heat exchanger will be of about the same 1mportance as secondary gammas from the beryllium and the pressure shell. - e hot reactor pressure shell l s> P o from the gamma Shleld, which 'is a layer of lead about 7 iun. thick. The lead, in turn, is surrounded by & 3l-in.~thick region of borated water. The slightly pressurized water shield is to be contained in shaped rubber bags similar to - fitted aircraft fuel tanks. Cooling of the lead shield will be effected by T T 7 3 e a i i 3 e 3 k- B f: Al . i S8 ] 2 e o -1k circulation of water through coils embedded, in the outermost'portion of the lead and through auxiliary coolers. The borated water shield will be - cooled by thermal convection of the atmosphere in the reactor assembly cell. ' Asseflbly and Testlng As each component of the reactor 1s constructed 1t wmll be cleaned ‘and leak tested in the Y-12 area. The whole assembly is designed so that leak testing can be carrled ‘out as the components are added, piece by piece. Mass spectrographic techniques will be used for leak testlng. The com- pleted assembly of circulating-fuel and moderator-cooling systems can be leak tested before it is moved to the site. All ‘thermocouples and other 1nstrumentation will be checked as the assembly proceeds.-* T Assembly of the radlators, blowers, NaK pumps, £111 and draln tanks,' and other auxiliary equipment will proceed concurrently with the assembly of ‘the reactor$at Bulldlng 7503. Cleaning and leak testing procedures will be the same as those for the reactor. All instrumentation in the building will ' bé checked out, as far as possible, prior to installation of the reactor package._ The reactor Shleld assembly (w1thout wamer in the Shleld)'Wlll ‘be moved as a package from the Y-12 area to the 7503 building. After all connections have been made, flnal leak testlng will be carried out. The NaK system'wzll be Filled with NaK and, ‘with the beat ‘barrier doers closed, the NakK pumps will be started. A heat input of approximately 300 kw will be attained by circulation of the NaK, and this energy in the NeaK will be used to preheat the reactor. The presence of leaks, if any, from the NakK system to the fuel system will be determined at this time by a flame photo- meter. In addition to the leak check, the circulation of NaK in the system will permit the checkout of all instrumentation and the determination of the system characteristics. After cireulation for several hours, the NaK will be sampled and analyzed for oxygen content. If the oxygen content is within the capacity of the purification system, the NaK will be left in the system for the remainder of the test; otherwise it will have to be replaced. If the NaK is to be replaced it will be dumped hot 1n order to carry the ox1de W1th it. ' The sodlum system for cooling the berylllum Wlll be filled next, and finally, the fuel system will be filled with a barren fluoride mixture. Circulation of the fluids in these circults will permit final cleaning of the systems, operational checks of instrumentation, and a determination of the system operating characteristics. As with the NaK, samples will be taken and analyzed for purity. The barren carrier will then be drained from the fuel system and replaced with a fluoride mixture containing 80% of the requlred‘uranlum (as determlnea from the crltlcal experlment) i “‘m.'-"A - ) P . . e e B B T = TR R P T TEroI i TR T T TR Y T e B e et bl Crk ailbie Sl i . 3. THE TEST FACILITY Site The AB.E facility, ORNL Building 7503, ‘a8 noted "'in the ARE Hazard.s Sum- - mary Report,™ 1s located at a site 0.75 mile southeast of the center of the present ORNL area and about 0.24 mile northeest of the Hemogeneous Reactor = Test (HRT) facility. This ARE location is near the center of Melton Valley . which is approximately 4 miles long and 0.5 mile wide. With the exception of the nearby HRT, this valley is unoccupied. Between the ARE site {elevation | 840 £t) and Bethel Valley, which contains ORNL (elevation 820 £i), is Haw ; Ridge which averages 980 £t in elevation. Within & radius of 1.9 miles, all 1 the land is owned by the AEC and is already a se@umtyucontrelled area. Within a radius of 2. 3 ‘miles 'bhere 18 approximately C.3 square miles of farm land . ‘that 1s not AEC owned or comtrolled. Additional information on the smounding _'_'area. and the natural characteristice of the site:is. presented in Appendix A. Bu1lding . To modify the"ARE Bua.lding to”"accomodate the ART, it is pla.nned that an "addition to the south énd will be constructed to effect a 6h-t extension of the present 105-ft long ‘building. The ART shielded reactor assem’oly and. its container will be imstalled within this addition to the ARE Buildlng. Such an arrangement will permit the use of ARE services and facilities that exist in o ‘this installation wh:a.ch has now gerved the purpose for which it was erected. """ For example, {tems such a@ the control voom, offices, change rooms, toilets, storage area, water supply, power supply,, portwns of experimental test pite; access roads, security fencing, and security lighting are available for in- corporation in ART planso Fig. 3.1 shows the preliminary degign of the facil- ity in perspec tlve. TR T T T AR | | The pla.n a.nd ‘eleva'bmn wings of the ART f’aeillty are ‘shawn in F':z.gs. ok SR 3 .2 and 3.3. The floor level of the addition will be at the ARE basement o - floor grade (ground level at this end of ‘the building), and the cell for housing the reactor assembly “will be supk in the floor up to 3 £t below the E L bolting-flange 1eVel. The resctor cell will be loeated in approximately the .é center of the he-f't wide by 6h-Pt long higha‘bay extension and. directly in L - line with the ARE experlmental bay. The reactor assembly will be positione& - so that the top of the shield will be at the building floor elevation. i cverhead crane facility w:.ll “be revised from 10-to-20~ton capaclty to per- S : mit use of the experimental pits for installation of auxiliary equipment i i 1. %, H, Buck and W. B. Cettrellg 'ARE Hazards Summary‘Rep@rtp aRmL-lhe7, 1952}« o R REE b T e o o o0 o o o | g o I O ircraft Reactor Test Facility. Fig. 3.4. A e Y e Sy o e UNCLASSIFIED kS ORNL-LR-DWG 4471 A 4—] 1 i o : . _ . . 1 - 106 ft 11in. - 64 ft O in. ADDITION L [ ] [ ] ] [ fi = r T T . — v : - R T LT IINE T | : LOGKER ROOM e ” INSTRUMENT COUNTING OFFIGE CONFERENCE > S ROOM ROOM , ROOM A : S 2w ' — g ze INSTRUMENT ey A . g B SHOP = | / o 4 2 Eo ooqu& CORRIDOR ‘ 5 § 3 = 1% Yyin, . z: | g et ezl Vet = = K . O w ] Pk . GENERAL CONTROL ROOM { w ‘> SHOP : ¥ . 7 Q = / | 2 <« / + [ & & 7 } L-EXISTING 12-X 12-ft 5-x 8-ft CONTROL ROLL-UP DOOR, 3- X e ' T T T V1 [T T T ° TRENGH 7-ft SWING DOOR AND 0t GOLUMN TO BE RE- MOVED. f L < = TI LL de b ol ”‘E”—._l' |_- ot e ap ot il e o %EUZJ o o o s i 11 ¢ i i Hr éjg <~ NEW 26-f-HIGH x ‘ B X3 30- ft-WIDE ROLL- . Bad UP DOOR. s I [11) p i ] i 1|t 1 . ] . e o ap o o e a0 4 . & — HATCH : g i i I PART FIRST FLOOR PLAN EXISTING BUILDING T MONORAI LEGEND ABO POURED CONCRETE . - 20 4 8 12 16 20 SOLID CONCRETE BLOCKS e — et e 77/} CORED CONCRETE BLOCKS - FEET PART BASEMENT PLAN ; ? 7 ® i Fig- 3.2. Plan of the ART Building. R O T T T R W g e T e e ey - v e v w K;’F‘ rrm rm e T =1tin. 42 f+-01n. 39 #+-Oin. ' ' ' , EXISTING B I ) - N 106 t-11 " CORRUGATED ASBESTOS SIDING 4211-10in. 10ft-Sin. 26ft-7in. 27t-0in, 3‘f’r—0in. 145+-7%In. {7 f+-0in. e \.3; 12Ft-0in. sicoi 7, NOTE: FOR PLAN OF ART BUILDING SEE Fig. 3.2 . _al_ SECTION A-A I—-‘IO + MONORAIL, INSULA REMOVABLE SCREENING — 1) 10f+ METAL DUCT ¥ RADIATORS BLOWER LEGEND POURED CONCRETE —~ — SOLID CONCRETE BLOCKS ; { i b | PN / | | L CORED CONCRETE BLOCKS ' SECTION B-B 10 5 0 5 40 5 FEET Fig. 3.3. Elevation of ART Building. UILDING 7503 UNCLASSIFIED ORNL~-LR-DWG 4472 ADDITION ©64ft-0in, METAL. DUCT i t | t | |l | l HE 1 \ 7/ | ~ 1]t 11 (] e e —————'———-FLJ_!.j Fhdg |- L C_dJd WEST ELEVATION —{ [=4ft-4in. 7] 8t-Oin. e 74 . FIRST FLOOR ——y, o] BASEMENT SECTION C-C O 4 8 P e FEET Y ¢ " v T Ty P P T BT e ST T I T TR s e ool R R i e AR X e e & e ke i e e SRR kb e «10~ afid.possibly for underwater reactor disassembly work after resctor oper- ation. Also, the truck door in the north wall of the ARE Building will be enlarged to provide & large entry door to the ART arsa. Field maintenance and 1aboratory facilitles will Dbe 1nstalled in the area east of the new bay and south of the low bay of the ARE. This area and the old expérimental bay will be partitlioned from the new bay with gbout & 16-in.-thick solid, concrete-block shield wall. This wall will not be erected until after placement of the upper sections of the reactor assembly container. The only other major modification to the ARE facility to accommodate the ART will be that of modifying and equipping one of the ARE experimental pits for underwater disassembly work on the reactor after operation. Reactor Cell . The cell designed for housing the reactor assembly is shown in Fig. 3.4. As may be seen, the cell is to consist of an inner and an outer tank. The heat dump equipment will be located outside the cell, but nearby. The space between the two tanks will be of the order of 18 in. and will be filled with water. The inper tank will be sealed so that it can contain the reactor in an inert atmosphere of nitrogen.at atmospherie pressure, but it will be built to withstand presgures of 100 psi, The outer tank will be merelyaa water contalner. ‘ . The inner tank will bu approx1mately 2# ft in dlameter with a stralght section about 11 £t long and a hémispherical bottom and top. The outer ' tank, which is to be cylindrical, will be approximately 27 £t in diameter and about 47.5 ft high. When the reactor is to be operated at high power, - the space between the tanks and above the inner tank will be filled with water so that in the event of an accident so severe as to cause a meltdown - of the reactor the heat given off by the decay gamma act1v1ty will be car- ried off by the water. Since the heat transfer rate-to water under boiling - - conditions would- be exceedlngly high (of the order of 320,000 Btu/hre ££2) -+ and since the thermal conductivity of the fluoride fuel is relatively low, . the water-szde temperature ‘of the inner tank will not exceed the water tem- . perature by more than 4OCF. The water capacity of the space between the - tanks together with the water in the reservoir above the inner tank (ap- fiiq{proximately 10 £t deep) will be of the order of 1,000,000 gal. ' Boiling of . .. the water in the annulus and above the inner tank will sufflce to carry S off all the heat generated by the “fission products after any acc1dent with— ,Hfjout any additional water being supplled to the tank° \/ the space inside the tank will be shielded fairly well so that it will be possible for a man to enter the inner tank through a manhole for inspection or repair work, even if the reactor has been run at moderately high power. L _ ”‘will be above floor grade. This portion of t ertank, ag well as the top hemisphere of the inner tank, will not be - attached until completion of the reactor installation and preliminary shake- ~ down testlng.' Since the shielding at the reactor will be quite effective, T T RN W I, . b o, Sisaiis il e ... ad e o s il o S e G5 . . S ORNL-LR-DWG 3568 © a0 ey LN L7 2q y ~ Fig. 3.4. Reactor Assembly Cell with Water-Filled Annulus. —20 -~ ERERTER - . N T o b LR G G . ..2]_- The unshielded reaetor assembly will weigh approximately 10,000 Ib, the lead gamma shield, approximately 30,000 1b; and the water in the shield, ap- proximately 34,000 1b. The first two of these items can be handled conven- iently with a 20-ton crane, while the borated water will be pumped in after the rdbber tanks have been Lnstalled for the water shleld. - The reactor assembly, with lts aireraftutype shleld will be mounted in the inner tank on vertical columms with the reactor off-center from the vessel axis and about 6 ft above an open-grated floor. This pos1tioning . will provmde the space needed for movement of the portable fluoride fuel and sodium moderator coolant containers to their operating stations under the re- actor. The off-center location will also serve to minimize the length of the NaK plplnga ‘ The NaK and off-gas plping connected £0 the reactor will pass through a thlmble-type passage or bulkhead in the double-walled cell. The openings will be covered with stiff plates which will be welded to the tank walls. The piping will be anchored to the inner plate and connected with a bellows- type seal to the outer one. The volume within all bulkheads will then be malntalned at a pressure above that of the inner tank to prevent out- leakage from the inner tanko A doubly-sealed Junction panel for controls, 1nstrumentation, and aux11iary services will be installed through the tank below the building floor grade as a part of another bulkhead to pass wires, pipes, tubes, etec., reqguired for the circuits and systems. The various thermocouples, power wiring, ete., .will be installed on the reactor assembly in the shop and fit- ted with disconnect plugs so that they can be plugged into the panel in a short period of time after the reactor assembly has been lowered into position in the test facility. This will minimizé the amount of assembly'work re- quired“in'the'field.” o Two more bulkheads, in the form of manholes, will be installed in the ‘ upper portlon of the contalner.” Ore manhole will be about 3 ft by 5 £t and located just above the flange on “the inner tank to allow passage through _ ;:both container walls and thns provide an entrance to the inner tank for use = ,‘after placement ‘of the’ upper “sections of the containero_ The second opening will be a manhole gbout 5 ft.in diameter in the hemispherical top of the - inner container to prov1de overhead crane service after placement of . the top. Sufficient catwalks, ladders, and hoistlng equipment will ba installed withln the inner tank to provide easy ‘access for servicing all equipment. The control bulkhead in the cell will be located so that the assoclated »' control junction panel and the ‘control tunnel will extend to the auxiliary equipment pit (formerly the ARE storage plt) The pit and basement equip- - ment will inelude such items as the lubricating oil pumps and coolers, borated shield water make-up and fill tank with a transfer pump, vacéuum pump, relays, switch gear, and emergency power supply. The reactor off-gas flow, diluted with helium, will be piped through the NaK.plping bulkhead to the disposal facility outside the building. - TR ey EmpTr WY TTTTWT T woom T T I A T S . M E R T " Heat Dump System -20. The ha31c requirement of the ART heat dump system is to prov1de heat dump capacity equivalent to 60 Mw of heat with a mean temperature level of 1300°F in the NeK system. It has also seemed desirable that the ART heat dump system should simulate the turbojet engines of the full-scale aircraft in a number of important respects, such as thermal inertla, NakK holdup, and f&bricational methods. Slnce heat dumps are required for use in heat exchanger test rlgs, and since work of this character ‘has already progressed to the point where the cheapest ‘most’ compact, "and most convenient heat dump is. currently proving to be a round-tube and plate-fin radiator core, 1t is believed that this basic type of heat transfer surface should prove both’ sufficiently relisble and sufficiently'well-tested to serve for the ART. The round-tube and plate- fin radietor planned for the ART makes use of type 310 stainless steel clad " copper fins spaced 15 per inch and mounted on 3/16-in.-0D tubes placed on 2/%-in. square centers. Individual radiator cores will have an inlet face. 2 ft squareal Mueh information has been dbtalned in tests of 51mllar unlts. Sl - HaK w1ll be clrculated through five separate systems.' Four will con- stitute the main heat dump system, while the fifth will be the moderator heat dump system. In the main heat dump system a group of four fill-and- drain tanks will be used, but these will not require the remotely operable - couplings desired for the sodium or fuel systems. The NaK will be forced | into the main cooling circult by pressurizing the tanks. The 2k tube hundles‘ of the ‘fuel-to-NaK heat exchanger will be manifolded in four groups of six each. The NaK will flow from these tube bundles out to the radiators which will be arranged in Pour vertical banks with four radiator cores in each bank. The NaK will flow upward through the radiator bank to the pumps . A small by-f pass flow through the expansion tank will allow it to serve as a cold trap. A filter to remove oxides will be placed in the return line from the tank. - The moderator heat dump system will be essentially similar except that its capacity will be about one-~quarter that of one of the fow circults of the main heat dump system. NeK will be circulated to the Na-to-NeK heat ex- changer in the top of the reactor where the NaK will pick up from the sodium the heat generated in the island and reflector. The NaK will pass. to & small NaK-to-air radiator where it will be cooled and returned to the pump suction. An expansion tank and by-pass filter will be ‘included, as in the main NaK system. It is planned to have only drain and filter by-pass throttle valves iri the NakX systems, since the NaK w1ll be drained if any ‘: f' repalrs are requlred° As shown iu Flg. 3.2, ‘the NaK-tb-air radistors will be mounted in an air duct close to the reactor cell° This duct traverses the southwest cor- ner of the bulldlng addition. The radmators Wlll be located at floor grade‘””' 5V, 5. Farer 5L L., Freliuiuery Design end Performence of Sodiun- " to-Alr Radiators, ORNL-1509 (Aug. 20, 1953). o R, SN B o P Ry, s b R : be supplied by a 2- ft-dla blower 0perat1ng at about 3&00 rpm° _23- over the NaK pipe-llne pit° Four axial-flow blowers will force 300 000 cfm | of air through the radiators and out through a 10-ft-dia discharge stack T8 ft high. Since the axial-flow blowers will stall and surge if throttled, control can be best accomplished through by-passing a portion of the air '”around the radlatorso' This arrangement required only constant speed a-c motors and simple duct work with a controlleble louvre for by-passing. The ~ heat dump rate will be modulated by varying the number of blowers in oper- ation. A set of counterweighted self-opeming louvre vanes in the inlet or dlscharge ‘duet from esch blower Will prevent backflow through the blowers not in operatlono Thus each blower will be driven with an a-c¢ motor inde- pendently of the others, and the heat dump capacity can be increased dn in- erements of 25% from zero to full load. An additional % £t by 8 £t set of controllable louvres will be mounted in such & way as to bleed air from the plenum chamber between the blowers and the radiators to get vernier control of the heat load. ! Heat barrlers mounted on either side of the radiators will be requlred to mlnlmlze heat losges during the warmup operations. Warmup will be accom- pllshed by-energlzlng ‘the ‘pumps and driving them at part or full speed. ' Since approximately 400 hp must be put into the pumps in the NaK circuits, this power will appear as heat in the fluld pumped as a result of flubd frictional losses. A mechsnical power input of 400 hp to the NeK pumps will produce a heat input in the NeK system of approximately 300 kw. This should be enough to heat the system quite satisfactorily with the radiator cores ‘blanketed to ‘prevent excessive heat losses. Relatively simple sheet stain- - less steel ‘doors filled with 0.5 in. of thermal insulation when closed over both faces of & radiator (6h ft2 inlet-face area) filled with llOOOF NeX will give a heat 1oss of 30 kw. Heat load control for the low-power range ‘presents some problems. The plenum chamber préssure with the 4 £t by 8 ft by-pass louvres wide open and. one blower on will be about 0.5 in. HyO.- This will give a heat dump capa- city of 3 Mw if all the radigtor he barriers are opened. Lower heet - “loads can be obtained by varying the number of heat barriers opened. Oper- - ating the ‘heat barrier doors against a pressure of O. 5 in. HQO (2 6 1b/ft2) .« should not be difflCUltor | \._,1,. .'f*éjTfiE heat appearing in the" moderator w1ll ‘be’ about 3 5% of the reactor -‘“pofiér dutput. "The ‘moderator codling circuit will also remove heat from the ore g@glls ‘and the’ pressure shell so that the total amount of heat to be ‘Mf?om the moderator coollng circuit will be about 6% of the redctor tput. This must be ‘removéd at & méan NaK circuilt temperature of about ~-~1050°F. A radiator having a 242 inlet-face area and the same proportions as those used for the main heat dumpe will be employed. This radiator will 5 e "3.1&.»3: ST A :,%,_;rif;{;i‘;"} :;,‘%‘,. ,-&fiwwm AR e L S i To permlt fairly easy fuel 1oading and removal, an effort is being made | to develop a good, reliable, relatively simple fill-and-drain system incor- - porating & remotely operable coup;l_ing° Such a pilece of equipment will permit T T T T Y T T QR ey ey T T ST e prp : - LT on i ChThac om0 e S n Lol g o . p o : g e el e Ll u 5 A & h e kel i Y- = il oMk uo oMl i oo e Ol removal of the fuel from the reactor and provmde considerable flexibility - in the conduct of operationsq It is believed that by removing the fuel the radiation level can be cut for maintenance operations. Easy installa— _tion or removal of the fuel will also fac1litate peprocessing the fuel or " modifying the fuel composition. Should a reliable remotely operable coup- " ling not be developed in time, welded attachments will be made between the ~ various flll tanks and the fluid systems, | ‘._w‘ E \\‘_‘_;_ S f- m; e o : For handling the heavy, shielded fluoride and sodium contalners in 1de¢ the pressure vessel, a track will be installed on the floor and 1n51de the wall. Wheels will be mounted on the tank dolly, both on the bottom and on. one ‘end, so that the assembly can be lowered by the overhead crane to the floor track with the end wheels on the dolly rolling against the vertical track. Once on the floor track, each dolly will be moved to its operating station under the reactor° Fach track pair in this area will be mounted on ‘a 1lift for raising ‘the tank connection nozzle to the contaet positlon w1thin - _the reactor shield. o et Other requirements of the fill-and-drain system include provision for _ accurate measurement of the quantity of fluld in the drain tank at all times during either the filling or the draining operation. This is particularly important in connection with reactor fuel systems because it is important that the'exact amount of fuel in the reactor be known at all times. . The shielding required for the fuel tank will be 10 in. of 1ead to re- duce the dose to 1 r/hr at 5 £t from the tank one week after full-power operation. The resulting shield weight for a T7- ft3-capacity tank will be about 15 tons. The lead shield required for the 1- £t0 sodium drain tank will be 5 in. thick, and it will weigh approximately 2 tons. Fill and drain systems for fuel or for sodium from the moderator circuit will include provision for both preheat and the removal of decay gamma heat. These functions will be carried out by diverting NaK from the radiator cir- cuits and directing it through a jacket surrounding the drain pipe and through coils in the drain tank. i Off-Gas DlSpOSal System The design of the off-gas system was based on the pessimistic assump- | tion that all the fission products will be given up to the off-gas system as they are formed and will be swept out with 1000 liters of helium per day. They will be passéd through a long charcoal-filled pipe designed so ‘that no more than 0.COLl curie/sec of radioactivity will go up the stack. About # Mw of the 60 Mw will appear as fission-product decay energy. Since the design was based on all of the fission products being released to the off-gas system, a decay of t™V*< was used in calculating the heat released. ~ The gases will be removed from the fuel in the expansion tank at the top of the reactor and vented through a 1/4% in. Inconel line 20 ft long to a 2-in. steel pipe 1050 £t long. The first 50 £t of this pipe will be ”open, but the balance w1ll be filled with activated charcoal. The gas T Y T T T TR T S o o T (-"!_"r‘-r % - -25- LT A R AT a0 DA o ML .+ R R AR i kS will exhaust From the charcoai bed to the stack where it Will mix with the 670,000 cfm of hot air from the radiators. On this basis, the holdup time and heat generation for each component of the off-gas system will be as follows: @ L o Component Volume Holdup Time ‘Heat Generation Expan31on tank (gas volume) ‘“:f{' X lO5 | 6007 o 13000 1/4 in. line 20 £t long 200 17 10 2 ino line 50 ft long 3L x'th 3 x 107 - 3p5 o in. 1ine 1000 £t Qomg T T oot (assuming 1/2 volume 5 | L is charcoal) 3.4 x 10 3 x 10 665 heat dissipation. Additional shielding will be uSed'as needed. If the gases are held up for one or twgadays, calculatbions sggw that Kr ™~ presents the greatest hazard, since K§8 and its daughter Rb““ give out about 2. 4 Mev per dlsintegratlon of XKr°®. The activity in curies/sec is 6 x 107(w) X 3 x 10 (flssions/sec‘w) x O, 01} (atoms/fission) x e~ At 3 7 x 10T (dis/sec~curie) - 1.9x 100 ye Rt ',1.‘-For a decay constant of 6. 9 x 1077 . dis/se%~atom and after a holdup time © of 48 hr, there W1ll remain less than 10 curies/sec° - charcoal bed, £ is the volume flow rate of the off-gas, and K is a con~ stant with'a ‘magnitude of about 500 for the type of charcoal to be used. : or a holdup time of 48 hr and a flow rate of 1000 liters/day, it is found " that only L liters of charcoal w1ll be needed, whereas 600 llters will be avallable. - 5; The formula, which was obtained Trom M. T. Robinson of the ORNL Solid State Division, is based on experimental evidence. The entire 1050 ft of 2-in. pipe will be in & trench under & ft of water for [ — RETTIT OMITRTRE e e ot i e e Rk R ackbdah PR O _26_ Since about 0.5 Mw of heat will be given up by the fission products on the charcoal, the limiting factor will be the heat transfer. Upon entering the charcoal, the gas will be absorbed very rapidly, and the first few feet of charcoal will soon rise in temperature to around LOOCF. The temperature will start to decrease in less than 10 ft, and, -by the time the gas has passed through about 100 ft of charcoal, the temperature will be down to near the | ambient temperature of the surrounding water. A.by-pass line ‘around the'Offégas system i& to be provided for ‘uge in case there is a leak along the pipe in the trench. In such an event, the reactor will be shut down and the gas will be vented into the reactor cell. After a two—day holdup, & vent line will be opened directly to the stack. This auxiliary vent line can also be uwsed if fission-product gases leak from - any of the reactor components into the reactor cell. Monitrons will be pro- vided at suitable locations in all gas lines. In the event that the off-gas system is to be operated at a time when no power is being abstracted from the - reactor, the air from the blowers will be &ucted around the radiators to avoid difficulties which would otherwise follow from cooling of the NaK._ ———— | AR i ML -27- ok, ccamom AI\ID OPERATION Control éfiilésophy The early ORNL effort to develop the circulating-fuel type of aircraft reactor wvas motivated in part by a desireble control feature of such reactors. Thls feature is the inherent stability at design point of the over-all power plant that results from the negative fuel temperature coef- flcéent\of reactivityo In a power plant with this characteristic the nuclear power source is a slave to the turboaet load with a mlnimum of exter» nal control dev1ces.vllv oo This predlcted master-elave relationshlp hetween the lead and pewer source was verlfled by the ARE. Controlwise the power plant consists of the nnclear source, ‘the heat dump (in the case of ART), and the coupling between source and sink (the NaK 01rcu1t) Control at design point can be effected to some extent by nuclear means at the reactor, by changing the coupling (i.e., changing the NaK flow), or by chenglng the load (i.e., the heat dump from the NaK radiators) _ For the ART at design point the regulatlng rod will be used mainly for adjusting the reactor mean fuel temperature. In particular, an upper tem- perature limit will cause the regulatlng rod to insert until the fuel out- let température does not exveed 1600 ¥. This 1imit will override any normal demand for rod withdrawal. Furthermore & low NakK outlet temperature from the heat dump ‘radiators will sutomatically decrease the heat load to keep the lowest NaK temperature of the system at no less than 1050°F. This lower temperature limit will overrlde all ether demands for power» _ Critlcal experlments will be performed W1th the system 1setbermal at . 1200°F. Thls temperature ‘was chosen to incredse the life expectancy of the f_j;beryllium moderdtor. The méan fuel tenperature at design point will be . 1%000F, thé moderator being held at 1200CF. Since the fuel temperature | coefficient is 5 x 10 = per °F, 1.0% Ak/k will be required to raise the ‘mean fuel temperature this 200°F» ‘ ’?fThe total worth of the regulating red over 1ts stroke w1ll be ebeut “;A,Q_S% Afi/k. In -addition to its use in changing the mean fuel temperature, '+ .this amount of rod will supplement the fuel addition by solid pills (dis- sedfat the end of thls Lhapter) 1n eompensating for burnnp and f13310n— o W i - peration in the deszgn‘901nt‘range (from 20 to 120% of desxgn-"' P r, which is the useful range for an aircraft power plant), manual change in load demand or in qperatlng temperature will be restricted by the meximum rate at which the load can be changed or by the maxinum rate at which the regulating rod can be withdrawn, respectively. In the design point power range the maximum rate of withdrawal of the regulating rod will be obtained by adding Ak/k at the rate of 3.33 x 10~* per second. TErTETY BT T v e o p—— T ¢ R T - T e e T T W R T R 2 N T T 4 - grrrr o TR 7 o L e sl il vion, . -28- e This will raise the fuel outlet temperature at the rate of 12°F per second until the meximum fuel outlet temperature of 1600°F is reached, at which - point the temperature limit will hold. In this range a permissible load ‘change rate of one-half design point power In 1 min is comparable to the requirements of engine performance for a nuclear-powered aircraft. Load ~ changes are effected by manual demand for changlng the air flow over the . NaK rad:.ators o Control of the ABT is classzfled in three dlfferent categorles of operation: namely, (1) startup, (2) operation between startup and design point, and (3) operation in the design point range. For the second and third of these categories the nature of the reactor and power plant is so different from that of conventional high flux reactors that control mnst be based on inherent characteristics of the resctor to a large extent rather than on conventional reactor conmtrol art. There is no conventional art for these categories with high flux reactors. Control at startup utilizes, in . principle, old reactor control art with short-period "scrams! that are con- ventiénal in princlple. Exyerimentation will take place primarily in the startup and design-point regions. In the intermediate region between these - two, little testing will take place. Consequently, operational procedure will be followed to take the reactor from the low-level adeguately control- led region to the high levél region in one simple manner. This procedure will be assured by permissive instrument t interlocks that are described in a following sectlon of thls chapter. Agvf»"* Fissionflchambers and compensated fon chambers Will,be located beneath%if°¢" the reactor shell between the fill and dump tanks and the reactor. The ' reglon ‘around the pipes between these tanks and the reactor will be filled | with moderator material, either Be or BeQO, through which cylindrical holes for théseé chambers will run radially out from the centerline of the systen. From féur to six such holes will be available. Chanber Sensit1v1t1es will be adequate for the entlre range of nuclear operatlono - The fuel expan51on chamber is a key item in prov1ding safety for the | ART. The fuel temperature coefficient of reactivity provides stability for the system by the expansion of fuel from the eritical reglon._‘Adeqnate expansion volume will be available atpall times, Scram System A conventional scram system achleved‘by droyping Poison rods into the o critical lattlce will not be used with the ART. The reasons for elimina— ‘ ting this feature are “the follow1ng ' 1. In the de31gn-point range described, analysis shows that llmlting the rate of rod withdrawal and the rate of load increase will limit the - period of the reactor when it is operating normally. A limited rate of ‘rod w1thdrawal and a limited rate of load increase near the design point in the ARE gave s minimum perlod of about 10. sec.. The same technlque will be used on fihe ARTo ") L T P A T e e —eee- b A G . o -29- Short periods, of the order of 1 sec, can occur in the design—point range only in the event of structural failure. The total Ak/k required in " the rod to override an increase resulting from such failure cannot be ob- tained from.one rod nor could such a rod, were it available, be inserted fast enough to prevent s serious accident. The temperature coefficient will react so rapidly “that it will limit the signal which would normally actuate a scram, ‘except for an extremely high rate of increase in reactivity. It hag not been 90531ble to devise a control system that would react rapidly enough in such cases.to prevent the accident. Therefore, in the deS1gn— point reglon the conventional scram would be of little merlt. -8 For the'lnitlal 1oading and crltlcal experlments a scram system will. be used, “but it Wlll not {avolve dropping the one control rod. The method described below was’ ‘proposed because the single-rod system lacks the safety feature of a plurality of rods, as ordinarily found in conventional reactors. The actuating signal will be a short period, & high flux, a menual scram, or any of & number of failures in the system, and the signal will be supplled - through an auction circult in the conventional manner. Essentlally, the safety of the system lies in the procedure of adding _fuel in a subcritical external loading tank and forcing it against gravity into ths fuel systemo This will be done by pressurizing the loading tank with hellum.through a valve which will fail closed. Two parallel helium ‘ outlet 11nes frem the loadlng tank to the off-gas system will fail open. All 1oad1ng tank and the reactor locked open, and the signal from the auction circult will actuate the solenoid in the helium-pressurizing system in the menner described’ ‘above. Actually; two parallel dump lines, one to the £ill- and—draln tank and one to the emergency dump tank, will contain valves which will be ‘actuated smmultaneously on the dump signal from the auction amplifier . clrcult. Thls system ‘has the merit that if too rapid addition of fuel to the system’ causes & short period, reversal of the operation will reverse the perlod° Control of the hellum,pressurizing system will limit the rate at which the fuel © i8"added to the’ system. The helium system can be designed so that it will fail , except for the case of a plurality of simultaneous failures comparable '”f3“1n‘probab111ty to the failure of a plurality of magnetic clutches all of which slmnltanecugly'fa}l to. open in the conventional rod-dropping reactor scram. Startup - A rather close estimate of the crltieal concentratlon shoald be avall- ablé from the hot critical experiment so that 80% of the uranium will be in the fuel at the time of the ART startup. ' The fimnal 20% of the required uranium will be added in steps. After each uranium addition the fuel will be forced frem the £ill tank up into the reactor by'means of helium pres- sure. The scram czrcumt will be available, ag described previously. The rod will be inserted for each step, s glven smount of uranium will be added, the rod will be slowly withdrawn, and a count will be taken on the fission chambers to determlne the suberitical multiplication as a function of uranium concentration° ‘A polonium-beryllium source of approximately 15 curies strength . will be 1nstalled 1n tha Pentral 1sland of the reactor to provide neutrons for startup. T T T MO - Ty £ ot w e TO TR ST T ool o kA A T " ki S i ek b S il AR | S RSt e (e i S S e i S T Ll i R T b b S e e Lo i - . P o - TR T s s s s sk e b bR KR Skt it e e R LS L b b e bl e b Lk SRR st s e b LR L B i G Bnt s e el e | _crit:;.cal‘ The_ whole system is to be isothermal at 1200°F determ:.nation of the ‘temperature coefficient, ete. ) will be manual with the maximnm rod speed providing a rate of change in Ak/k of 3.33 x 10~% per _‘bhe leading tank as described above, ‘Operatmn Between St&rtup and Design Point After going critical, the reactor will be leveled out menually at about " lowed to go on a period. This will be a check on the effects of flow rate . withdrawn; and the reactor will be allowed to go on a period until a level - - ingerted to drive the reactor suberitical. This procedure will be repeated ~ to give 2 or 3 calibration points on the regulating rod. Since a similar - red will have been carefully calibrated in the hot cr:l.t-lcal experiment, onl:y' _ ' a few rough check p01nts will be necessary. ' ~ menually with the rod. At this power level, the shielding and off-gas The pump speed will then be increased so that the fuel flow rate will be - the temperature coefficient provides stability while the reactor gets its ' - ~ power demand from the lecad. Accordingly, a single operation procedure for S every’ ‘operation in this range will be followed, The load will be interlocked T ge that permission to start adding the load will come only when a compensated _30- =, " R T ke Before beginnmg the eritical experiment the speed of the pumps will be set so tha.'b the frlow through the core will be about 50 gpm. Thus over one-half the delayed neuwtrons will be available for control Whlle go:mg COntrol for zero "‘power oPerat:Lon&: rod ca 1bration, fuel enrichment, second. Overrlding the manual rod withdrawal will be a 5-sec period rod reverse end a l-sec period fuel dump by rellevmg ’che hellum pressure in TR T g ” v L - - A hold:.ng servo system will be used at ZeT0 power for exper:.ments re- quiring constant neutron flux. Operation with the servo system will be essentially the same as that for the ARE. Limits will be maintained on the rod speed, and overra,des will be maintalned on per:.od and tempera‘bure. o SRR b . Since most of the nuclear data will have been obtained from the hot eritical exper:.ment, the low power operation will be held to & minimum, 10 to 100 watts. The pumps will be stopped, and the reactor will be al- | on the reactivity contribution of the delayed neutron fraction. When the power level has reached about 1 kw, the pumps will be started and the rod will be inserted to drive the reactor suberitical. Sufficient uranium will be added to give sbout 0.5% excess reactivity. The pumps will again be stopped and the reactor will be brought to about 10 watts; the rod will be of 1 kw is reached. The pumps will again be started, and the rod will be The power 1evel will then be elevated %o about 16 W and leveled out o o systems will be checked out thoroughly without great hazard to persomnel. increased f‘rozn 50 gpm to the design flow rate of 1200 gpm. This will cause a decrease in reactivity of the order to O. 2% Ak, and the rod will be with- dram accordingly. The reactor will then be ready to del:wer power, ' T T The negat:_ve fuel ‘bempera'bure coefficient of the ART mkes manual con- | trol mandatory in taking the reactor from zero power to some power at which - ! 3 T R T sl sl jon chamber current reaches some prescribed value. This value will be determ- ined in ‘bhe manner deserlbed, below the first time the reactor- is taken to power. b R S R, LS + . - e \ ' Lo , . o s r"}vq‘: ,: .::.g E -is : With all loop flow rates at design point and the reactor at &bout 10 kw and isothermal at 1200°F, the regulating rod will be withdrawn until the reactor is on a positlve period. This period will gradually increase Cuntil it becomes infinite and finally negative because of the temperature - coefficient.’ MEanwhlle, bothithe log N and micromicroammeter readings will go through a maximum. This meximum log N reading will provide the signal +to permit opening of the heat barrier doors to the NaK radiators. Natural convection from the radiators with these doors open will be about 300 kw. 'Accordlngly if these doors are opened when the log N reading reaches this value, the température coefficient will always suffice to provide regula- tion and stability, provided the rate of load demand above this mlnlmum is restrlcted tq the value 01ted previously¢' : If after shutdown the flux exceeds the log N reading requlred to open the heat barriers; opening of the barriers will not "shock" the system even though the reactor may'be suberitical at the time the doprs are opened. If on the other hand ‘the flux is too low to permit opening of the barriers, there 1s only one precedure for getting perm1651on. - De51gn-P01nt Qperation Wlth the reactor at about 500 kw (estimated from the power extracted by opening the heat barrier dcors) the blowers will be started and heat will be extracted from the NaK, which, in turn, will extract heat from the fuel. The reactor will be leveled out at 3, 15, 30, and 60 Mw. A heat balance will be obtained at each level of extracted power vs nuclear power. The Qperation of all components will be observed at each power level. . Care will be taken not to exceed the maximum temperature of 1600°F .or fall ‘below the minimum temperature of 1150°F. : The reactor will then be operated for 1000 hr at 60 Mw. Xenon will be removed contlnuously from the fuel by hellum injected into the pump chamber and escaping in the swirl chamber. The rate of re- ..moval by this means can be determined only by operating the power plant. However, experience with the ARE has indiestéd that less then 1% Ak/k of regulating rod will be needed to cope with the xenon that is not re- m eq. Should the purging be much less than is anticipated, the xenon €0 w&iand would under some circumstences shut the reactor down. The low- 'mperature ‘1imit on the NaK radiator outlet temperature will automatically remove the load to effect this shutdowna_ In case this happens the fuel " w111 bevdumped umtil the Xenon decaysa uel e Ptp compe-s te'for burnup Wlll be accompllshed by dding fuel inlthe form of high-UP37-content "pills" ef solid fluoride 'i*fuel, These will be introduced into the reactor fuel circuit through an 5fifentry provided in the fuel expansion tank located on the north head of ' ‘the reactor. The pill-addition meghanism will be carefully designed and tested to make it jam-proof and incapsble of ejecting all its pills in - one spurt. It will permit the 1ntroduction of only: @ne plll at a tinme to the reactor system. T T R T T T T RS T R T szl dkalidd A N - TR g r e Gt Goe. i e -32- The total burnup is equivalent to about 2.5% Ak/kq The capac:;.ty of the pill machine will be such as to hold no more pills than that amount ‘equivalent to 2.5% Ak/k. Accordingly, the rod will always be capable of ‘overriding any fuel addition. Furthermore, the rate at which successive ' pills can be added will be much less than that which can be effectively can- celled by movement of the control rod. Compensation for burnup and fission- -~ product poisoning can be accomplished by both conmtrol rod'W1thdrawal for ‘fine control and'by fuel enrlchment for ce&rse control. A 1 L The’best'ch01ce for a plll material 18 tie com@ound EaQUFs. “This wvas used as the enriched fuel component for the ARE test. Its melting point is apprOX1mate1y'1160QF and its solid density at 1000°F is about %.9 g/cm3. ‘ It is composed of approximately 60% U233, 114 Na, and 29% F (by weight). Several pill dispenser (and container) designs have been prepared that are based on the use of pills 1/2 in. in diameter by 1/% in. thick. Pills of this size have a volume of 0.80 cm3, they weigh 4.0 g, and they contain about 2.% g of U235, The rate of pill addltien required to maintain & constant reactor fuel inventory will thus be 0.5 pill per Mwd of operation, For operation at & power level of 60 Mw, this will require the addition of 30 pills per day, or & total of about 1200 pills for the 1000 hr of full- power operation. 1 Bok & bt R SRR, - 'f“'radlators. o An attempt has been made to envision as many hazards as possible that might occur during the course of the operation of the Aircraft Reactor Test. Included in this chapter, therefore, is a discussion of . the normal radi- ation hazards, the hazards resulting from operational or equipment failures, and fluid leaks, as well as the nuclear and chemical hazards peculiar to the cycle. The dispersion of airborme activilty, either from the ‘off-gas system or following a hypothetical accident in which all the fuel is volatillzed, is described in the’ follewzng chapter, "Dispersmon of Alrborne Activ1ty. The radloactivlty of the ART will be 1nherently conflned‘by the nature of the design and materials in such a manner that the uncontrolled disper- sion of the activity outside the resctor cell will be virtually impossible. Consequently, the hazard from most failures will be negligible, since the o only action required will be dumping of the fuel; the activity would not ‘even be released to the cell. Furthermore it is shown that wh;le a hypo- thetical nuclear accident could rupture the reactor pressure shell, the reac- 'tor cell Wmuld remain 1ntact and the accldent would he safely contalned. Some con31derat10n'has been glven to cases in whlch the reactor cell, as well as the pressure shell, would be ruptured, and the resulting subse- quent d;sper51on of activity has been examined in detail. It is believed that such an accident could occur only as the result of aerial bomblng or sabotage. : _ Radlatlon Dose Levels The radiation dose levels to be expected at representative stations ab the facility have been estimated for a variety of conditions and have been tabulated in Table 5.1. 'The shielding assumed for these egtimates was the following: (1) the primary airchaft-type reactor shield designed to give - 1 rem/hr at 50 £t at full power, (2) the reactor cell steel and water walls, =1 (3) 16 in. of concrete block stacked around the reactor cell to a height of 10 ft, (4) 16. in of concrete block stacked between fhe reactor room and ‘the maintenance shop and between the reactor room and the former ARE main " test bay, (5) concrete block stacked around and on top of the air duct for ‘the NaK~to-air radistors in such & way that the equivalent of 12 in. of oncrete will be imposed along any radlal line extending outward from the ’room and around the air ‘duct and radiators are ‘intended %o provxde “‘shieldlng in case a fuel leak developed e;ther into the reactor cell or - 1nto the NaK systems, TR e e m— T W TR R R R T g = pe T o o ‘o ahis | mblcsae Rk oo - oo . . & i . iy e N i i et vk i e M iR w e i e o B e e e i e an et et RS ' EORMAL OPERATION DOSE IEVEL DOSE LEVEL (rem/hr) WITH DOSE LEVEL {(rem/hbr) WITH {rem/hx) 1% OF FUEL IN NeK IN RADIATOR ALL FUEL IN BOTTOM OF CELL 15 min - 10 days 15 min days 15 min 10 days Full After After Full & After - After After After Location Power Shutdown Shutdown Power Shutdown Shutdown Shutdown Shutdown Reactor shield : - 5 5 3 4 surface 100 4 1 % x 10 10 2 x 10" 107 - o,x 103 ) Ooutside reac- -2 1 " | ' | | gfi tor cell 10 h x 10 - 10 1500 500 120 60 15 - [ . Outside reac- - ) % i - | - Hor room 10 4 x 10 10 3 ' 1 0.25 0.5 0.08 control room 1077 L x10°7 107¢ 0.6 0.2 0,05 0.0k 10,01 Road 5 x 10 2 x 10 5 x 10 0.6 0.2 0.05 Oll - 0.02 e e B B T R A s L SR RN m55- e As may‘be seen from Tab Te 5 l, ‘the readtor wall be adequately shielded so that the control room operators will receive much less than 1 ren/hr even with 1% of the fuel in the radiator and much less thap 0.1 rem/hr even if the pressure shell is ruptured. The dose rabes would be consider- ably higher, however, if it were postulsted that the reactor cell was also ruptured, in which case the activity would no longer be confined. This ex- treme situation is considered in the following chapter, "Dispersion of Aifiborne Act1v1ty.? - : nal and.Eq, Eggnt‘Failures There are any nufiber of operatzonal or quxgment fallures that can be eHV1310ned in & system as complex as the Aircraft Reactor Test. In this section are listed those failures which would have the greatest effect on the operation and which therefore seem to offer the greatest hazards. For each failure some probable ceuses are given, as well as the resulk, and the action reqnlred in order to minimize the hagzard is stated. In all cases, ~as will be shown, the faillure would be inconvenient, but no serious danger would ensue, since the most drastic action required would be dumping of the fuel (and/or sodium) into the dump tanks. Therefore it is also apparent that the operabllity'of the dump sgstem mnstrbe assured, - Fuel Freecze, ExceSfilve coollng of ths §r1mary'HaK clrcuit could ‘canse fuel to freeze in the heat exchanger and stop the flow of the fuel. The heat exdhanger wauld not be seriously damaged, but some cracks might form in tube wallso_ ‘Because of fuel flow stoppage and consequent lack of cool- ing, the temperature of the fuél in the fuel ecircuit would rise 13°F/sec (boil in 2 min) as a result of fission-fragment decay heat, which will be 3800 Btu/sec (6% of power) immediately upon cessabion of cooling in the circuit. In ‘the reactor structure, the cooling available from conduction after dumplng of the fuel would net be adequate 50 keep pump blades, wells, and other points where fuel might be trapped from being raised to fuel vaporization temperature. Excessive cooling in the primery NaK circuit “would, therefore, require that the fael be dflmpe69 T L g, o EER erator éoalzng clrcult. The temperature of ‘the beryllium.moderator W'°F/sec at fullfipower operation anda‘lflF as a result of It wpuld.be : excegsive heating, and‘pressure surgefi’would'be The possible causes of structural failures. If a pressure ¢éised & 0.020-in. expansion of the ouber core shell, there would_ be T tivity change of +0.002." If the outer shell were to collapse under - vEnexeéssive exterpal pressure.load, there would be a large redctivity de~ ierease, and the' possikility of leakage of sodium into the fuel circuit. The results of deformetion would be similar, but the effects would be of a lower magnitude. A failure of the pressure shell would release fission products; hot, highly radiocactive fuel and abiendant decay heat; and NaK. .PE k L - g e, N ey - TR E R T R -36- : A fallure in the fuel-to-NaK heat exchanger weuld cause & NaK or a fuel 1T leak. Deformatlon 1n the fuel_to~NaK heat exchanger would possibly result i _ in sllght changes in pressure drops and heat transfer characteristics. It - would prdbably be necessary to dump the fuel if any of these postulatea - events ‘occurred, with the poss1ble exceptlon of slight deformatlon in the fuel-toéNaK heat exchanger. . Pfim Faalures. Loss of power to pump drmves, as Well as selzlng of shaftsz bearings, or impellers, could cause pump stoppage. If one of the two fuel pumps stopped pumping, the fuel flow patiern would be altered and roughly one-half the fuel-to-NaX heat exchanger would be starved. 'There would be a consequent reduction in power output. If both the fuel pumps failed, fuel flow would stop and the fuel temperature would rise. 15°F/sec é : because of fission fragment decay heat (see item on "Fuel Freeze" above). 1 "~ If only one fuel pump failed the fuel would be dumped or the reactor would 1 ~ be operated at reduced power; if both pumps failed it would be necessary to - : dnmp the fuel. , , *, bt e s e pog i Iffone sodlum pump falled at full power operatlon, the temperature of the sodium would rise 0.25°F/sec to accommodate the increased heat load on - the opersble pump, and it would be necessary to dnmp the fuel or to oper- ~ ate the reactor at reduced power. If both sodium pumps failed, the tempera- tures of the beryllium in the moderator and the sodium would rlse Q. 5°F/sec, and 1t would'be necessary to dump the fuel. : The fallure of one Naxlpump in the prlmary heat exchange 01rcu1t would reduce NaK flow and conseguently reduce the reactoxr power output. As with other pump failures, it would be necessary to dump the fuel or Lo operate the reactor at reduced power. If all the NaK pumps in the primary heat exchange system falled, the fuel temperature would rise l3°F/sec (see item on "Fuel Freeze" above), and the fuel would be dumped immediately. Faile ure of the NaX pump in the moderator ‘cooling circuit would ‘cause tempera- tures of the sodium and the beryllium to rise 0.5°F/seec. As in the case of the sodium pump failures, it would be necessary to dump .the fuel, . . If the pumps for prov1d1ng coollng 011 to the pumps were to f&ll, there would be a slow increase in temperature of the oil coolant, the. fuel punp shaft, the bearings, and the gas-seal mechanism, .Failures of this type would be taken care of by switching to the auxlllary pump and repalr- ing the pump that failed.. and all .ingtrumentation and essential equipment would be transferred to it. Therefere, there would be no immediaté hazard following such a failure. - The emergency power system will be adequate to operate at least one fuel pump, one sodium pump, one NaK pump, one blower, and all the necessary in- struments. This equipment will be sufficlent to prevent excessive tempera- ture rises from the- fuel afterheat. All possible measures will immediately be taken to restore the normal power supply as rapidly as possible. How= 7 ever, if the failure lasts an extended period of time, 1t may be necessary to anmp the fuel. e DR Sl ot SSsas i mase tt st i s s st BESRG, R R R S s R U i s e SRS S e s S e e L T - Electrleal Power Fallure. Afi*émérgefiéyfiaser*sfigpiy'fiiiilfiefiésaiiéfiie o i i Rt 2 B bl A, LA o A — e 2 dbide ke e .f:-;:""uFuel Sodium or NaK Ieaks ..37- Fuel Channel Hot Spots. Flow separation in the core or fallure of the core-shell coolant system could cause hot spols in the fuel channel. In this event there would be the possibility of fuel boiling in the core and causing irregularities in power or increased corrosion. The powver level would be reduced until the fuel boiling ceased, or, if necessary, the fuel wou.ld. be é.um;ped. Exaessive ‘Puel Feed. A failure in the ‘enrichment system might result in the addition of excess fuel. In this event the reactor would heat to a new and higher equilibrium temperature. An excess of 0.6 1b of URD) intro- duced instantaneously would make the reactor prompt critical. A Ak./k of 0.002 would oceur and result in an immediate fuel temperature rise of LO°F, The reactor would qulc}fl.y level out at the new temperature., If the equi- librium temperature were excess:.ve » the fuel would. be dumped. Fuel Flllwa.nd-])raln ..:;,rs’cem Failure. The fuel fill-and-drain system might fall because of jammed valves or a coolant system failure. If such a failure occurred before the fuel was enriched, there would be no hazard. The system would be repaired, if possible, or the nonradicactive fuel would be dra:.ned on the floor of the reactor cell. In the event of an emergency drain of radiocactive fuel coincident with a failure of the fill-and=drain system that prevents drainage, the reactor will fail at the weakest point and release hot fuel, fission products, sodium, and NaK in the reactor cell. " If the drain system functioned satisfactorily but the dump tank cooling sys- tem failed after the fuel was drained, the tank would fail at its weakest point and release large quantities of hot fuel and fission products to the reactor cell. The reactor cell is designed to contain the hot, radicactive fuel y as described in a follemng section. If it were desired to drain the rad:.oactz_ve fuel under normal operating conditions and drainage was pre- vented, it would be necessary to cool the radiocactive fuel with the normal heat removal system until the decay heat had dropped sui‘flc:.ently to permit shutdown of the NaK system. | NaK C:chult Heat Dg_g_gg Blower Failure. I:E' the blowers falled, heat loss : .:E‘rolm the radiators would be by nabtural convection only and would be about © .3 Mw. Fissioning would ‘stop in the reactor, but the fuel temperature would ‘rise lO"F/ sec because of decay hea.t. In this event, the :t‘uel would be , ., dumpedoq ‘Binee ea.ks' betwee :"the variousfi flu:.d. systems can g:Lve rise to some - o:f' the most serious accidents that can be postulated for this reactor sys- tem, it is important to examine the conditions which could cause a leak, the ‘reactions which would subseq_uen‘hly becur as a consequence of the leak, and the ultimate hazard. If the many welds are good, as will be determined by radiographic techniques, as well as by prellmnary test:.ng, any leaks that might ocelr would probably result either from corrosion or a fatigue crack. Corrosion is far more probable and is of particular concern, largely because of the uncertainties associated therewith. S EmTwr o e . e g e Y o- ey . TUT Fremmmeer 100 TR e T il *mflmw rE bk i T T W T PRy R A T W e e e e o T TR T g g v BT T e o | ek ¢ e G SR e 4 e ik ik i o LBl ke oh Bl BEE i ~38- The corrosion process is discussed in‘Apyendixrc.' From the test data included there, it is estimated that the corrosion penetration in the hot zone of the ART would be of the order of 15 to 18 mils if no reductious, in comparison with present experlence, can be effected in the corrosion rate. The wall thicknesses to which the fuel is to be exposed will be 125 mils in the core and 25 mils in the heat exchanger.- In the core the wall thickness is believed to be ample; in the heat exchanger the metal would sustain a steep temperature gradient because of cooling by the NekK s0 thet the penetrations given above may not @pply. However, should com- ponent tests fail to justify this assumption; heavier tubing will be used. If a leak, elther from corrosion or fatlgue, did occur, various chemi- cal reactions could occur between the fuel and the Na or NeX, dependlng upon the fuel composition, the size of the leak, and whether the leak was into or out of the fuel system. The consequences of the chemical reactions from each of 16 distinct leak conditions are also discussed in Appendix C. It is apparent that the resulting hazards are dependent upon the assumptions ‘made regardlng the leak size, extent of completion of the reaction, and the pre01p1tat10n of inmsoluble partlcles. The most severe case 1mag1nable is discussed in a follcw1ng sectlon ‘entitled "Accidents Causing Rupture''of the Pressure Shell." The probable consequences are, however, much less severe and are discussed below. The action taken in each case would be to dump the fuel and the subseqpent hazards would be small Fuel Leak in cOre, Slnce the fuel pressure is malntalned below that of the sodlum, a core leak would most probably result in the addition of sodium to the fuel. The sodium would dilute the fuel mixture and, in the reaction of sodium with the fuel, UF5 would be produced. The reactlon could continue ubtil metallic uranium was produced, which, in turn, would be deposited in the hotter part of the system (i.e., between the core and the heat exchanger). This situation would be handled by dumping the fuel--no hazard'would ensue. . On the other hand, if the fuel were to leak into the sodium in the moderator region, the result would be more serious.because such a leak would certainly allow excess uranium to be present in the core region. In this event, the fuel would be dumped as soon as poss1ble. Fuel Leak Into the Heat Exchanger. A leak in the ‘heat exchanger wculd '--éither admit fuel into the NaK system or vice versa. A fuel lesk into the NaK system would increase the activity outside the primsry shield, while a NaX leak into the fuel system might result in excess uranium in the core. ‘Either of these situations would be undesirable, but it is now felt that a fuel leak into the NaK system would be the more hazardous; therefore -the pressure of the Na2K in the hest exchanger will be maintained higher than that of the fuel in the fuel system. Accordlngly, a leak in the heat exdhanger would.probably result in NaK enterlng the fuel. As with sodium, UFz would be formed and the fuel mix- ture would be diluted. Eventually urahium would be formed and would be de- posited in the hotter section of the system. - This situation would e handled by dumping the fuel and no hazard would result. T T T v e T T A it okeii Py -59.. If the fuel pressure were to be grester t.han that of the NeK and a leak occurred, some of the fission products would move outside the reactor cell, The radiation ¢ doses which Would. be experienced at various locations in the facili‘l}y with ‘as much as 1% of the fuel in the NaK system are given gbove in Table 5.1. The dose rate in the control room 15 min after the ~shutdown is _shown to be only 0 2 rem/hr. The fuel and then the NaK systems | ether Sod.mm or ‘IiiaK Leaks, j In addrl::l.on +0 leaks involvmg the fuel s the NaK systems might leak externally, and a leak might develop in the Na- to-NeX heat exchanger. A leak of sodium into the NeK would increase the gamma activity of the Nak, wha.le a NaK leak into the sodium system would decrease reactivity in the core, since potassium is more poisonous te the reactor than sodium is. Helther of the above situnations presents a serious hazard. !_a would require enly "l:.ha:b the systems :Ln ques‘b:.on be d.ra:x.ned.. In “the event of a external? Hak eak ‘co air, “the resulting fire would release some act:.v:.ty (55 cur:.es, mex:r.mum) he system would be dumped and ‘the :f.'ire would, be extingulshed. - , Sod.lum or KEK Fires. Wa:hh regard to fire as a bhazard, 'the bullding (Flgs. 3.l and 3.2) carries a Uniform Bullding Code fire rating of 2 hr. Inflammeble materials are not used in any appreciable guantity in the con- struction of the building or the reactor. The reactor, as well as the associated plumbn.ng , pumps, and heat transfer equipment, has been examined rather closely from this standpoint.because of the high (up 1o 1600°F) operatz.ng tempera‘bures involved.. Howeveér, except for the use of Na and NakK as coolapts in the system, éven “the high temperatures (:.n the a’bsence of combustible ma'ber:.al) prec.en'b no hazard The possn.bll:.tles of a sodlum or a N4K leak have been preva_ously dis=- cussed, However, during operation, the reactor cell will be filled with belium {or nitrogen) in which sodium and NeK are not flammable. In fact, - experiments have shown that even the potentia.]ly dangerous reactions of NaK ... apd water are greatly reduced in the absence of oxygen. A Nek leak external to the _gell would, howeéver, résult in a fire, and therefore the Alkali _Metals Area Sa:f‘ety C%uidel will be employed. Materials which will safely L gxtmgu:.sh “a"NaK ‘or “Na fire are graphite powder and Ansul~Met-eL~X (sodium . chloride cga’ced to prevent the absorption of moisture).2 Adequate Quantl- "t:.es of he'se materials m.ll he kept at convem.ent locat:.ons. Since such conventlonal: extmgulshers as water 3 002 » and sand should ot be ‘applied to a NaK fire, the conventional sprinkler system has ‘been o fomfbted from the design of the building; hewever C, & :f:‘:.re hyd.rant on a 6-113. main is ;prov:.ded 20 f’c :E'rczm ‘I:.he bun.ld.:x.ng. e . '2. A trade compound deve.Loped by Ansul Chemical CGmpa.ny 9 Marmette ’ W’.LS. T P. ToED, e e Area STty G{ime ,""3-811 (Aug. W T TR MRS R ot e oo . ?‘ F £ - i ki 'and material of comstruction in the building, there is little probability of any damage occurring from‘natural elements whlch'could creete a‘hazard. serious hazard to the ART. As a consequence of the particuler topography s S i o 'I-i. e bR ok i S hdil i ke 'Further, the data on the frequency and severity of earthquakes in the Oak 'mal working stress of 20,000 psi in the steel structure. A review of the " winds of this magnitude will even be approached at the sheltered site of the J“Huclear Accmdents Cau81ng Rupture of the Pressure Shell cen51dered the worst concelvable nuclear accident. Therefore the estlmated ~ %this or any other Ak.could.be introduced into the reactor. | '33;; Bm L. Myers and J. Ze Hblland, A.Meteorologlcal Survey of the - “40- Acc1dents Causeé.by the Natural Elements As a conseqnence ‘of the partlcular ‘Ioeation of the ART and the type Flocds and Earthquakes. Nelthcr £100ds not earthquakes present g selected for the site, a flood, and therefore flood damage, is impossible. Ridge area show that the probability of earthquake damage is extremely small (see section on seismology of area in Appendix.A, "Characterlstlcs of Site") | | * Windstorm. With regard t6 windstorms, the building is designed to the Uniform Building Code crlteria. These criteria provide for design against ° wind loads of 20 1b/ft“ (about 100 mph) Without exceeding the allowsble nor- meteorologlcal data for thls ‘area” shows that it is highly improbable that ART. o It is always instructive to con51der the consequences of what mlght be reacfi1V1t1es from various changes in the reactor are given in Table 5.2. From this table it may be seen that a Ak/k of 0.09 is the highest that may be expected. It is then necessary to consider the maximum rate at which TABEE 5 2. ESTIMATED BEACTIVITIES FROM MAJOR CHANGES IN BEACTGR Tota.l va.lue of control rod T, 665 Ai{/k _Control rod.mctlon | - 1 ¥ _ 0 00035 QSk/k)/sec ”JTcmperature coefficient of | . 5 reactivity : -5.5 x 10 L&k/k)/° Removal of sodium from passages | * through reflector | .00151Ak/k Removal of sodium from passages - through island 40,0005 Ak/k Fuel replecing sodium in core shell | an@ reflector cooling passages 0.09 Ak/k Gak Ridge Ares, GRO-99 (NOV, 1953). o \ _é o beiliboch i K G & Sl G L ik G S s e e Lo LA sk, I S TR AT 41 ‘"It has been possible to conceive of two extreme situations which appear t0 establish an upper limit on the rate at which the reactivity could increase. They are as follows: (1) fuel abruptly begins to pre- cipitate out in the core and the fuel stream enters the core at the nor- - mal rate but no uranium leaves in the exit stream, and (2) fuel abruptly enters the moderator cooang passages ‘replacing the sod;umdn A third . situation~~one in whlch the ‘beryllium reflector melts and mixes unlformly with the fuel--was examined, but it developed that this accident would not give nearly as large a rate of increase in keff ‘because the high heat capacity of the Peflector would keep the berylllum from melting rapidly. The rate of temperature rise in the portion adjacént to ‘the fuel region would be only 20°F/sec at 60 Mw. The two most severe accidents are dis- cussed below. In addition, the vulnerability of the redctor cell to pene- tration by pressure shell fragments, if the pressure shell were to rupture as a consequence of other ac01dents, is examzned. The 1nvest1gat10n of these two extreme nuclear accldents may be sum- marlzed as follcws' ~f;i l,_ The 1nherent stablllty of c1rculailng~fluar1de-fuel reactors derived from their high negative temperature coefficients make the pro- posed 60-Mw reactor self-regulatlng even for extremely rapld changes in eff' S 2; Even 1f the pressure shell were to be ruptured as’ & result of an extremely high rate of increase of Kerp [h()% (Ak/k)/sec:l , the pressure shell fragments ejected would not plerce the l/2~inw wall of the reactor cell. , - 3. The radlatlon hazard from all the fuel in the bottom of the , reactor cell would'be even less than for the case with 1% of the fuel in the NeK system, i.e., only 0.0k rem/br in the control room 15 min after the accident (see Table 5. l) i FiR sam Fuel Pre01pxtation in the Core.? In fihe event oan sodlum; r‘NaK leak f' into the fuel system, uranium might be formed.wh_ch could dEPOSlt out in’ the core (Appendlx c). It has therefore been assumed for con51derat10n of ' ‘this bazard that (1) all the uranium fluoride in the fuel is rediuced to ~ free uranium, (2) this uranium.prec1p1tates out in the core as fast as it *" 'is formed so that no uranium leaves with the exit Tuel stream, and (3) the fuel flow (and hence uranlum flcw) 1nto “the core is maantalned at the maxlef' ums ratas., fffiél clrcult._ -the rate of the pressure increase, The fluid pressure ‘would rise until ° the pressure shell ruptured. Thermal expan51on of the fuel would continue until the reactivity was reduced 40 less than unity and the power dropped back to a low level. Boiling of the fuel might or might not take place. T T T e e is case have , nceived The power and.hence the tempera-”_ b e Tu \ Thas pressure rise would be pr@pagated with the'veloclty'of”sound T.ea,y 8t a hlgh rate compared w1th | F T e T rrewer PR R B - R SR P e Ca e Bl BRELL i e g If boiling does take place, the heat of v&porlzation of the ZfiFh is. 40 kecal/mole. This accldent ‘has been considered analytically in Appendix D and by a nfimsrical a@proadh in Appendix E. The results of the btwo methods are B reasenably consistent and are summarized in Fig. E.l of Aypendix E. For this type of accident the pressure shell is ruptured only if the reactivity ~is increased at a rate higher than that corresponding to half the rate-at which fuel s weuld.be pumped into the core, If the pressure shell should rupture, ’ - the reactor fuel, and sodlum'would.be spewed into the shield where they would ‘mix and’ react with éach other and with:‘the -shield water..:The resulting chemical resctions are discussed in Appendix B. As may be seen in this Appendlx all the chemlcal reactions may be contained, with the possible ex- ceptlon of the case discussed in a following section in which there is an ~alr atmosphere in the reactor cell and a hydrogen explosion becomes a possibility. Fuel in Moderator. if”tfiewédaififimfiiegéfirgéié'nét mainfiéifiEd higher ~ then kbat of the.fuel in the core, fuel would énter the moderator region in the event ef a core shell leak. It has been assumed that as a conse~ quence of unforseeable events an abrupt rupture of the reflector shell close to the core inlet would begin to discharge fuel into all the cooling pess-- ages through the beryllium at a rate equivalent to that given by the sodium velocity. While the cooling passages would probably plug close to their - inlets it is interesting to construct the probable course of the accident if it is assumed that mno plugging occurs, The initial fuel velocity through the passages would be about the same as that for the sodium, i.e., 30 fps. Because off the higher density of the fuel, its velocity would fall off as it penetrated the reflector. The worst case would be that in which the fuel entered all the coolant passages simultaneously, thus giving the maximum ~ rate of increase in reactivity. The initial fuel veloclity through the re- flector would give a transit time of about 1/7 sec, or about 1/7 of the corresponding value for the core. The volume of these passages is about 0.16 ft5 as compared with 3 £t in the core. The increase in Kepp if fuel filled the passages in the reflector and island has been computed to be 0.09.- The average power density in the fuel in the reflector was calculated to be six ‘times that in the core. This would make the average rate of tem- perature rise in the reflector fuel about equal to that for the fuel in the core, However, the rate of temperatmre rise at the nose of the fuel columns in regions of high importance would be at least twice the average so that: ~ the rate of temperature rise there should.be at least twmce the awerage value 1n.the core. , Calculatlons were made for this case, except that the raie of 1ncrease in keopr was taken to be twice as great to give a truly extreme case. In this instance the power was found to rise to about 6000 Mw in about 80 msec, at which,301nt the fuel in the beryllium would begin to boil. The pressure in the reflector cooling passages would rise sbruptly to about 500 psi, the erl.would.be expelled from them in ebout 10 msec, and, at that point, the core fuel would have reached a temperature of about 300°F above nermal the . Ireactar Wbuld then be on a 20~msec negative perlod. 3 3 3 B e s R LG . The tabulated ealculatlons fcr this accldent are glven in Appendix E, ' | and the results are shown in Flg. E 2‘ Again, the resulting chemical reac=- to rupture the reactor cell unless, possibly, oxygen was present in the cell. Penatrablllty of‘Beaetor"Cell by Pressure Shell Fragments. The ‘pressure 'shéil will be constructed to burst open at a pressure ‘of 1000 psi. The velocity, V, of fluid escaping thrcugh a crack if rupture should occur at thls pressure can be computed as follows: V'= NeeE , where g is the acceleratlon and H is the height of a fuel column giving 1000-psi pressure. With a fluld density of 200 Ib/ft53 th;s helght would'be 1000 (1b/m.2) x 1hk (11_102/ft2) __;" 720 £t , 200 (1b/£42) ‘ and 215 fps . ll v \Te x 32.2 (ft/secg) x 720 (:f‘t) ThlS veloc:ty of 215 fps is certaanl& a reasoaable one and should not glve ' any particular trouble., It is also 1m@ortant because it represents the maximum possible velocity of a fragment that mlght be broken out of the pres— sure shell in the event of a hydrostatic rupture. This velocity has ‘to be compared with the velocity required to penetrate the l/2-in.‘wall of the en- closure. As discussed in Appendix F, the penetrating power of projectiles varies with their shape, hardness, strength, mass, and velocity, with a 6-in. casteiron sphere giving a rough standard for this case. As shown in Appendix F, the velocity for penetration of l/2-1n. steel plate by such a sphere is roughly 380 fps. Since the pressure shell is designed to rupture at the bottom, and since the shielded drain tanks for the fuel and sodium will be carried on a substantial floor 1mmed1ately under the reactor; the tank bottom is well protected from any pressure shell fragments, ‘particu= S larly 51nce ‘only small fragments could have a velocity as high as 215 fps. . Thus it is clear that, even in the most pessimistic case, penetration of © the cell wall 1s out of the question; even if no account is taken of the energy loss of ‘the fragment durlng 1ts travel through the lead and water. It was demonstratedwln the precedlng ‘section that the reactor cell will ,t be. ruptured &s a consequence of even the worst conceivdble nuclear acci- dents. If it is postulated that one ‘such nuclear accident mey coincide with - ‘the production of ell the heat that could possibly be released from the " chemical combinations of various materials in the Tédctor and in an oxygen "atmesphere, ‘enough additional energy-would‘be available to give still higher cell pressures and possibly rupture the reactor cell As pointed out in the previous section, since the chemical reacttons are more serious with an oxy- :gen aimcsphere, the cell atmosphere will be maintained greater than 99% T T T T TR T ottt o T T —l{-"l' - | nitrogen. The oxygen concentratlon will be monitored, and the instrumen-~ tation will be interlocked (durlng power operation) to dump the fuel when- ever the oxygen concentration is greater than l% Tt then appears that the only means by which both the pressure shell and the. reactor cell could be ‘ruptured simultaneously would be as a result of bomb damage. The possi- bilitles of such a bombing, either by sabotage or aserial bombing, are re- mote. The principle hazard associated with the simultaneous rupture of the reactor pressure shell and cell would be from the dispersion of the activity. This is dlscussed in the following chapter on “Dlsper31on of Airborne Act1V1ty, ‘ Hydrogen EXploélon. In considering the various chemical reactions (see "Appendlx B), one way that the reactor cell might conceivably be ruptured would be by a hydrogen explosion following the reaction of the sodium or NaK , w1th water in an air atmosphere. As mentioned above, nitrogen is to be main~- talned'w1thin the reéactor cell once any apprecisble radiocactivity has been generated in the fuel. Oxygen absorbers will be exposed to the atmosphere within the cell and an oxygen monitor will be interlocked to open both the _ fuel and the NaK drain valves if the oxygen concentration goes above 1%. The energy produced from the various reactlons, including those reqnlr-‘ ing an oxygen atmosphere, have been calculated in Appendix B and are summar- ized in Table 5.3. The most serious event would be that in which the shield water would combine in stoichiometric proportions with all the sodium and NeK in the system, and the resulting bydrogen would burn in the presence of the available oxygen in an air atmosphere in the reactor cell. If no heat were lost to the cell walls, the pressure in the cell would reach 181 psia.. 0bv1ously, this is an overestimate because comsiderable heat would be re- moved by the surrounding water durlng the course of the reactlon. ' DABIE 5.5. SOURCES S oF ENERGY fiéat from reaction of 1000 1b of Na”'¥ " - 6'3 . and NaK'W1th'water . - - 2.07 x 107 Btu Eeat from reactlon of hyflrogen with | ' ~ available oxygen in air-filled | 6 - ‘reactor cell | - 1.56 x 10° Bhu Heat from reaction of 1000 1b of Na S and NaK with air - 2.90 X 10~ Btu | Heat from reaction of 1200 1b of 6 - zirconium~base fuel with sodium | 'O 98 x 10" Btu _Ehéflfrém exiféme nuclea} aCcident'; o o. 3 X 106 Btu jfif Flssionsproduct decay heat emitted . . during first 2 hr after shutdown | : (assumlng no flSSlon-product removal) - 8 x 10" Btu o W‘W N e G e B g T s o b 4 v R W G e . S Gl e e Al e i B ook Rk L_seems exceedlngly 1mprdbdble. “H5- If all of the hydrogen were to mix with the oxygen without igniting, detonation could take place and give a shock wave that would increase the stresses in the‘walls of the reactor cell, oOf the 181 p81a only part, or 66 psia, arises from the hyflrogennoxygen reaction. If it is assumed thai the pressure increase assoclated with datonatlon is twice the normal pres- sure rise, or 132 psia, and that the pressure from all reactions prier to the explosion is not relieved by heat removal by the walls of the cell, then a peak pressure of o7 psia would result. This value is about double that for the case in whlch a nltrogen atmosphere is malntained.withln the tank. ‘ In all prdbabillty the hyflrogen Would‘barn as 1t was formed from the»»,, NaKéwater reaction. The simultaneous occurrence of several unlikely events would ‘be necessary for an explosion; namely, the oxygen absorbing and moni- torlng system'would fail, an alr lesk inteo the cell would occur, a leak between the sodium or NeK and the shielding water would occur, and the hydrogen would not be ignibed as it was formed but would react only after the major part of the NeK-water reaction had gone to completion. Even if all of these things dld happen, it seems unlikely that the resulting shock wave could rupture the tank wall, particularly in view of the inertia of the steel wall and the water surrounding 1t. , £Mmage From High E391051ves. BlGW1ng np ef the ABT with explos1ve charges set from within by sapoteurs would be feasible, as for any instal- lation. However, the effectiveness of such action would depend primarily on the proximity of the charge with respect to the most critical components, i.e., the reactor pressure shell and the fuel drain system. Accessibility to these components and therefore vulnerability will determine the effec- tiveness of this type of sabotage. The double-walled reactor cell serves as a barrier against entry and would be formidable as protectzon against external explosions. However, the successful placement of an explosive charge within the contalner durlng servicing operations could be effective. In that event, the reactions of the reactor fluids, including the shield fwa$er, would take pTace as deserlbed in the preceding sectzon. 1ves Were detonated gust “%e released ‘and a most serious smtmatlon would.result. Both might also be ruptured by aerial bombing. Either of these cases . The qgestions of the strategic 1mportance of &wfi hickly settled areas must be baken imto account. A.bombing attack would most certainly be expected under wartime conditions, and ap= o proprlate measures ceuld.be taken at that tzme sh@uld the eventuallty occur. e :s E_ i v e . Sl R . e i s BASRELE o Zo Vi it e REMRD bl G s e s e e ST G e % “ e e e i L g dhdie £ s o o R LR o ik o Rk R s o i i § RGN R, . b v b adi ST RGNS ) e e et s o S LG s - -46- P ETARS e : Effec'blveness of the Reactor Cell in Conta:.n:.ng Hazards Eo allowance for “the temperature ‘rise of the atmosPhere within ‘the : double-walled rea.ctor cell associated with the relea.se of fissien-preduct afterheat has been made. Calculations weré mde %o include this factor as “well, and it was found that the’ ‘temperstures and préssures would fall off slowly after the accident. Thus these values represent meximums; any . severe accident would almost certeinly involve marked quenching by the : 'shiel_d wa.ter y probably to the extent that little pressnre and ’cemperature rise 1n ‘bhe tank would be exper:.enced. | L B One of the main reasons for surrounding ‘I:.he inner ta.n.k with a th:x.ck L layer of water was to provide a simple, positive cooling system. The worst . heat load 1ikely to be thrown on this cooling system would be that resulting ~ from discharge of Puel into the bottom of the tank after a long period at ‘high power. The lower portion of the tank will be designed so that the fuel will be spread out in a layer 2 in. at the thickest point. For this thick- ness of fuel and a 1/2-in.-thick steel plate tank bottom, 1t can be shown that- | Pcwer density in salt frem f::.ss:.on-product act.wity o o 1 min after 1000 hr at éo:Mw = .33 x 300 1o-w/cm Hea't release rate from 2-in.-'bh:1ck layer | w7000 w/f'b . £l EE : - o 160 000 Bte/hr-ft _At in steel pla.te | C woo%r Ae in vater ?1ln S wF A’c in fuel la.yer (essumlng no convectlon) ltOOoF Aree. of layer agalnst shell o | 36 242 Amonnt of water required to absorb 8 x 106 Btu (heat from fission products in first 2 hr) | I 1ts temperature rise is 130°F 62,000 1b (1000 £t3) If 1ts temperature r:.se is 130°F S 3 - and it vaporizes 8,000 1b (32 £t°) 'A;mount gf vater required to absorb & total of . 20 x 10° Btu (heat from fission products in : next 22 hr) | | e its temperature rise is 130°%° 15h 000 1 (2500 f‘t3)__ If' 1‘ts temperature r:.se is 130°F 3 and it vaporizes 20,000 l'b (32 £t ) TR o premising test :E'a.czls.ty. =47~ | Amount of water in 27-ft-dia outer tank 1,000,000 gal (14,000 £t) Amount ef'water in 27-'ft-d:fa gisk 1 £t thick = 4,200 gal (570 £49) Actually; any'acci&ent severe enough to dump all the fuel charge into the bottom of the tank would prabably also rupture the shield so that the shield water would float on the surface of the fluoride. It would boil vio- lently, and water vapor would rise, condense on the tank walls, and draln back to the bottom of the tank. : The double-walled cell was devised primarily in an effort to give a thoroughly reliable means for absorbing the heat evolved in any accident, no matter how severe. No pumps or other motor-driven equipment would be required and, even if the electrical power supply were to remain inoperable for days after the accident, there would be enough heat capacity in the water so that little, if any, of 1t would vaporize. The high surface heat transfer coefficient associated w:th boiling of the water should ensure good eoellng of the cell walls. Comparison of Various Reactor Assembly'Containers Because of the 1mportance of contalning the products of & remctor accident, several different potentlal reactor containers were examined. A comparison of key data for the most promising types of container is given in Teble 5.4, and the worst set of conditions applicable to each case is pre- gumed. As mey be seen, the reactor assembly cell propoged in this report compares favorably with the hemispherical and ellipsoidal buildings. The double-walled cell also appears superior in that it would be less subject to sabotage. Even if both the inner and the outer tanks were ruptured by sabotage and the reactor melted down, the residue would tend fo sink to the bottom of the tank pit where it would be flooded by the water that had filled the region at the top and between the tanks. This water would serve both to | absorb the heat of any reaction and as & shield to reduce the radiation 1evel» at the top of the pit. After careful review of these and ‘a host of lesser considerations, the 2h-ft-d1a double-walled cell was chosen as the most e o, TRTPr TR T TR T BT ST CNEE K e e S TR gy TR g e T T ey e T AT T o o R s R o 3 o . - o S e b roaaemite, || e . el el s i ki Ly TABLE 5.k, COMPARISON OF HAZARD DATA FOR SEVERAL TYPES OF REACTOR CONTAINERS . ol -ft-Dia 200-ft-Dia 1l5-ft-Dia - - Double-Walled Ellipsoidal Hemispherical y: Cell with L1-I% Buildi Buildin Straight Section uLLaing & Heat released, Btu 2.4 x 109% 107 107 L | 6 Container volume, :fi't5 12,230 l.2x l06 0.k x 10 Container surface area, :E"c.2 - 2,640 h.3 x .'.l.Olp 2.1 x -ZLOlL Peak gas température, °F 2,792 130 220 Peak gas pressure, psig 18 1.k 3.9 L ' Required shell thickness, in. | 1.0 0.15 0.09 @ (for allowable stress = 18,000 psi) p _ Weight of steel in inner shell, tons 36 L3l%x 40 * The fission product afterheat was not included in calculating the peak gas btemperature and pressure for the dou‘ble-walled cell because adequate cooling had been provided, Includes steel framing. o s s e e e s e e ey T B RN R ET—————"— 33% - - e ” - =ho- oo . 6, DISPERSION OF ATRBORNE ACTIVITY Meteorological data have been used to calculate the possible radiation hazards to the Laboratory and civilian population as a result of both nore mal and accidental release of radicactive materials from the ART. The nor=- mal method of discharging activity will consist of directing the off-gases through a chercoelnfllled pipe which will remove most of the activity and effect a more than two-day holdup before ejecting the residual gases up the stack. The stack in carrying the large volumes of heatéd air produced by the process systems will give very large plume rises in winds under 10 mph. The resultant ground exposures will be always and everywhere substantially below tolerance. In fact, even when there is no process air flow up the stack, the resulbing exposures will be below tolerance.. The process alr'will;' however, effect a reduction in the resulting ground concentration of around 10, which will be deszrable in some emergency 51tuatlons. ( L In the calculatlons for an dccidental release, a value of about 6 X 108 | | cal ves used. as the -minimum ameunt of heat which could cause all the fission 3 activity to be given off in a gaseous cloud. The total activity present was i then assumed to be that given by Millfs formula,2 even though ARE operation : indicated that some of the act1v1ty'was continuously removed and therefore not available 0 the disaster. This activity will be safely dispersed from a hot daybtime cloud, but it will exceed tolerance at night by 2 meximum fac- tor of 5 aor 17, depending upon which tolerance value is used. As would be expected, the doses from the cold cloud or from rainout of either cloud ! would be well above tolerdnce. It is worthy of note +that the assumptions made for the calculations which follow, ‘both for the case of the discharge of activity up the stack and for the. daspersmon of activity from a disaster, have been comservatively A_ 'be.ken _atsevery step. The result:.ng safety factors combine to effect indi- - "Q_ geeonebly be expected to occur., Since 1t wonld become w tedlously'repetlfilous if this censervatlsm were to be noted for every situ- .~ ation ig the follewmng seetloe, the more. pertinent factors are llsted end e “of 50 ilters per mlnute s used, ‘even ‘though this is the breathlng rate for an excited man and such a rate cannot naintained by an individusl over a prolonged perlod. ‘The krypton toler= “ance of 6.3 x 1077 curie/m> for continuous exposure used in the calculations glves only 300 mrem over the duration of the test (1000 hr). Ko decay was taken in the krypton act1v1ty after 1t left the holdup system. o : T. Most'ef T SerTon e e Tthen by R. F. Mgefé_afid'n.‘h, Pardy of the | U. S. WEather Bureau, Oak Rldge, Tennessee. 2. M, M, Mllls, A.Stu@gfief Reactor Hazards RAA-SR~31, p.'na (Bec. 7, 19&9) T e VT O il OO i B ;T QORI e . i e e my———— T TR e LTI MR T T il e A ol - Holdup System. In the charcoal-filled pipe there will be over 100 times the minimmm amount of charcoal required to effect a two-day holdup of krypton. In the calculatioms, no credit is taken for residence time in ‘bhe reactor system or travel time in the off-ga.s system. Meteorologlcal Parameters. A sta'bility :f:'a.ctor (m.gh‘btime) of 0.4 was - used . in the disaster calculation rather than the more likely value of 0.35. - The hot cloud and the stack plume rises were limited in some instances, ' - even though greater heights would have effected greater dispersioms. Although the nighttime stable conditions will not last much longer than 16 hr, the -unrealistic assumption of continuous nighttime conditions was not used for continuous exposures or for disaster-cloud travel times of greater than about 16 hr. ~ Low average wind velocity values were used for the hot cloud, even thongh equally justifiable higher values would have effected grea.ter » dispersions. Also, for each calculation that gives the maximum dose at any location, it is assumed that the wind always blows in that direction with a constant optimnm va.lue . ' FlSSlOB Products in Ho’c C'loud. fhe amount of heat in the hot clowd is 8 1ower limit for the amount of heat required to veporize all the fuel be- -cause the heat of vapor:n.zatlon of the first component which comes off (as the fuel is heated) is used for determining the heat reguired to vaporize all the fuel. For determining the actin'by in the cloud, an upper limit 'is used, since all the fuel would mever be in the right place at the right time to be vaporized. Also, operatlon of 'bhe ARE 1nd1cated that some ac'blv:Lty may be removed continuously. Radiation Tolera.nces | The ma.ximwn total dose which the civilian populat:.on should be permit- ted to receive in any accident is 25 rem. The maximmum permissible exposure to contaminated a'bmosphere which will give a dose of 25 rem by inhaletion is therefore of considerasble interest. A value for the maximum permissible exposure of 10 curleasec/m has been given by Marley3 for total fission products. However, Marley used a breathing rate of the order of 6 liters/min that is considerebly lower than that for the average excited man, which is around 30 ll'bers/mln In applying Marley's tolerance to a given condition the total radiation is assumed to decay accordn.ng to t"o'2 as the radio- active cloud moves out from the source. | S On the other hand, T. J. Burmett of the ORNL Health Physics Division has calculated,™ on the basis of a selected group of 30 long-lived f'lssion products, that the maximum permissible exposure to these isotopes is 1. hh curie- sec/m3 af'ter 39 d.ays5 of reactor operation. The 30 isotopes 3. We Go Ma.rley, Eealth P‘bysics Cons:.dera‘tlons in a Beactor Accu.dent, o - R/SAF/WK/3 (no date). k. Appendix G, "Exposure Hazard Ca.‘Lculations, '_' this report. 5. | Th:.rty-nlne days is used here (rather 'bhan 1+l, which would 'be a closer - . approximation to the antlcn.pated. 1000 hr of operatlon time) because %% seme’ calculations were already avellable with this time and the dif- ference is sma.]_'!. e B B b el g o e o AR R e 'forKr i the stack m'th 37x 100 cfm of amblent ‘air amd st;.ll no‘b exceed a0 3 m’bernal | -51_ selected defme the llm'blng tolera:afle for all fn.ssmn products . After 1000 hr of eperatlon, these 30 1sotopes represent about 876 of the total activity present at 1 sec, if £=0.2 i5 assumed for all the activity. Since these are leng-lived isotopes, no decay correction is applied. (It is further shown in Appendix G that this group of 30 isctopes may be reduced to a group of six isotopes which combribute over 957( of the dose to the bone.) . - As was noted previously the off-gas system is designed to routinely dischar Sge less than 0.001 curie/sec of Kr88 to the stack. The tolerance -RbS8 has been calculated® to be 6.3 x 10-8 curies/m3 for contin- uous exposure. This calculation permitted s dose rate of O 3 mrem/hr for the 10“ hr of ccmtemplate& regctor opera‘blono Discharge of,Ac-:t;ivi“ty up the Stack | The ART stack is 78 ft high and 10 ft in diameter. Its dimensions were largely defined by the air flow from the process systems since it was desired to employ this air in order to effect greater dlspers:.on of the off-ga.s.' The design capac.i'by of the air blowers is 3 % 102 cfm of air at ambient temperasture. This volume of air when heated in the hot-air radi- ators to T50°F expands to 6.7 x 105 cfm, and ‘therefore, during power opera- 'bmn, thls 1atter qtrby of air is dlscharged up the stack. The d.ispersa.on of ac.,t:.w.ty from the stack has been considered for three conditions of air flow: (1) 6.7 x 107 cfm of T50°F air, (2) 3 x 107 cfm of embient air, and (3) no air flow. With the two-day holdup provided in the off-gas system before the off gases reach the stacks, the off gases may be discharged without anywhere exceeding the k58 tolera:nce of 6 3 x 10~-8 cnrle/m for continuous exposure for the first two condrbmns. "For no’ st.ack air flow, the Kr88 tolerance will be exceeded by a factor of k. Further- more, if it becomes necessary following a ‘contained disaster to dispose of large amounts of activity, as much as 8 curies/sec. could be discharged up ‘Normal Operation. During normal operation the A:chraft Reactor'fest o ‘will produce sbout 6.7 x 102 cfm of air at T50°F from the air radiators. ;é:lr w1ll:“b<=,~ egec‘bed up the stack snd hence will aid in the dlspersal | o gaseous fission prod'acts which Will be given off by “the bhot fuel. o ’;The work of Damdson'T (which has been compared with actual smoke observa- tionsd from a large TVA steam plant stack for verlflca‘bion) has been used = | T '_’_\W. F, Dav:.d.son, "The D:Lspersmn Spreadmg of Gases and Busts :Erom - Chimeys, " Ind. Hyg. Foundation Amer.,, Trans. BuJ.l Ho. 13 (19&9) 8. F. W. Thomas, TVA, Wilson Dam, Ala., Plume Ihservatlons, Watts Bar Steam Plant (1952), meubllshed manuscript. T TN TR T IR T eI -52- in estlmatlng the plume rise expected from a lO-ft-dla stack, 8 ft hlgh. Figure 6.1 shows estimated plume rises versus wind speed for both - 6. T x 109 cfm alr_at 750°F and 3 X 105 cfm of air at amblent temperature. The calculated plume rise values were used, together with ‘sppropriate ) " meteorological data Gerived from the dbservatlons reported by Myers and Holland,9 to .compute the maximum ground concentrations and the distance from the stack of the maximum ground concentration under a wide range of wind speeds. The full plume rise was used in the unstable or daytime case, but the plume rlse was limited to about 600 meters at night, which corre- | ' sponds. to the upper limit of ‘observed rises from,the large TVA steam plant stacks durlng ‘stable condltlons. The calculated rises below the 600-meter ‘level were used for the nighttime rise. This treatment of the stable case minimizes the safety factor of the light-wind stable case which might be the most doubtful. The important dispersion conditions which are considered are those'whlch oceur with winds under 10 mps and are representative of 99. % of the hours of'w1nd dbserved at the site. The parameters used for these calculatlons are given in Teble 6.1. decay’ correction given here is that determined by the decay rate for equi- _ llbrlum reactor flSSlOD prodncts, £=0. As prev1ously noted (sectlon on “Radlation Tolerances") the concentra- tion of Kr®® for continuous exposure is 6.3 x 10~ curle/m « Furthermore, for a reactor operating at 60 Mw the equilibrium dlscharge rate of Kr after a two-day holdup in the off-gas system is 0.0009 curie/sec (section on "Ooff-Gas System"). The maximum concentration (with no decay correction) oceurs during the day at 0.37 mile from the source when the wind speed is ‘10 meters per second (1 mps = 2.24 miles per hour). This concentration (6.08 x 10~9 curie/md per 0.00L curie/sec emitted) is a factor of 10 below the toleranee for continuous exposure. Furthermore, it should be noted that the w1nd.w1ll not blow in the same direction and with the same high (10 mps) veloc1ty for 1000 continuous hours of daytlme conditions, and, in additionm, . the actual off-gas holdup system will have over 100 times the amount of char- coal required to effect the two-day holdup which was used in the above cal- culations. QgeratiOfifii%houfijfieetigg'Stack Air. Except when the reactor is oper- = ating at full power the air flow up the stack will be less than 6.7 x 107 cfm and the air temperature will be less than T50°F. For the limiting case with no power bheing removed from the system, there will be 3 x 107 cfm of ambient temperature air flow up the stack. A by-pass air duct will be provided around the radiators so that the air may be sent up the stack without cool- ing the reactor. The lower temperature and smaller air flow will produce +the lower stack rises which are given by the lower curve in Fig. 6.1. These plume rlses, together Wlth the meteorclogical deta derived from OR0-99, were i st 9.'-3. L Myers and J. Z. Hblland, A,Meteorologlcal Survey of the Oak R;dge o Area, ORO- 99 (Nov. 1955) T T Hake il . PR & .=‘ < : . B Lo - SR L RIETEE S et - ¢ - S A A i po R T - ults of'these calculatlons for steady eniission of 0 OOl curle/secf are glven in Pables 6.2 and 6.3, with and without the decay correction. The 3 ¥ - T m'» T —m—-‘ P B T I R e M T PR, T T sy g v e TABLE E-2. NUCLEAR EXCURSION CALCULATION FOR A HYPOTHETICAL CASE INYOLYING FUEL DEPOSITION IN THE CORE TO GIVE AN INITIAL RATE OF INCREASE iN k¢ OF 5%/sec Conditions: Accident relieved by fuel e.xpulsion from core forced by thermal expansion Fuel: NqF-ZrF4-U.F4 Time Interval: 0.020 sec ‘Mean Neutron Lifetime: 4 x 10~% sec (@) ) () (d) (e) o (e () () ) (k) (D (m} (n) Time, t k Excess, k Period, T Power, P Ne:rr:o;:::r:;:::rval Temperature Rise in Change in Fuel Total Change in Change in Volume Change in Yolume Fuel Exit Velocity, Velocity Pressure in (mse::): 0.00005 [(a) '— (th/2)} (msec;: (Mws: 07 (Btu): ’ Timé interval, AT’ (°F): FuelATemp:ra-fu;e Rise, Meon I:uel Temperature, Yelume in :fimg Fuel Volume, Av (H3): Absorfaed.in Pressure3 Absorbed gof Blowout Disk, (0 u (fps): ] 2.]-in.y.diq 1et, AP Ar =20 ~G),_,/15 0.4/() (&) _, PO [y v @), (e)/180 T CFF 20 TpCF): )+ 1400 aterval, AVZ(frh: =) 7k 1074 Shell D"_‘?'{”' Avp ) AV (54) e (e 00216 (1) " _ 120l A2 () 5.7% 10 7.6 x 1075 ()~ (m) _,] e x 10%/(At x 0.025) 0 0 o 60 0 0 0 1400 0 0 0 0 20 0.0005 800 61.5 15 0.08 0.08 1400 0.000046 0.000046 0 40 0.0015 267 T 66.4 79 0.44 0.52 1401 0.00025 0.00030 0 0.5 60 ©0.0025 160 75.4 218 1.20 1.72 1402 0.00069 0.00098 0 1 80 0.0034 18 89,0 444 2.50 4.22 1404 0.00142 0.00240 0 3 100 0.0043 93 110.0 790 4.40 - 8.62 1408 0.00255 0.00492 0 5 120 0.0051 78 142 1,320 7.30 15.9 1416 0.00410 0.00910 0.0001 8 1.38 140 0.0058 69 190 2,120 10.8 26.7 1427 0.00616 0.0153 0.0001 12 3.11 160 0.0064 62.5 262 3,320 18.5 45.2 1445 0.0106 0.0260 0.0005 21 9.52 180 0.0067 59.7 366 5,080 28.2 73.4 1473 0.0161 0,0420 0.0010 32 22,0 200 0.0066 60.5 500 7,460 41.4 114.8 1515 0.0236 0.0652 0.0019 47 47.5 220 0.0060 66.7 675 10,550 58.6 173.4 1513 0.0335 0.0990 0.0024 | 60 78 240 0.0048 84 856 14,110 78.5 251.9 1652 0.0447 0.144 0.0046 : 80 140 260 0.0028 143 980 17,160 95.0 346.9 1747 0.0540 0.198 0.0058 100 216 280 0.0002 2000 990 18,500 103.0 449.9 1850 0.0590 0.260 0.0036 110 263 300 -0.0027 -148 860 17,300 96.0 545.9 1946 © 0.0548 0.312 0.0000 110 261 320 ~0.00535 ~75 658 13,980 88.0 633.9 2034 0.0501 0.361 ~0.0022 105 235 340 ~0.00769 -52 448 9,860 55,0 688.9 2089 0.0314 0.392 -0.0078 79 133 360 -0.0088 ~45.5 299 6,270 35.0 723.9 2124 0.0200 0.411 ~0.0056 _ 50 56 380 ~0.0091 ~44.0 189 3,680 20.5 744.4 2144 0.0117 0.425 ~0.0039 | 29 19 400 ~0.0089 -44.8 121 1,900 10.5 754.9 2155 0.00600 12 N 420 ~0.0083 -48 80 810 5.0 760 2160 T AW e s v ————— T ™ TR T > e v 113 i Al s a ek Loom o AR = el T TABLE E-3. NUCLEAR EXCURSION CALCULATION FOR A HYPOTHETICAL CASE INVOLVING FUEL DEPOSITION IN THE CORE TO GIVE AN INITIAL RATE OF INCREASE IN keff OF 10%/sec Conditions: Accident relieved by fuel expulsion from core forced by thermal expansion Fuel: N¢;|F-Zr|"'4--UI='4 Time Interval: 0.020 sec Mean Neutron Lifetime: 4 X ]0"'4 sec (k) (n}) (a) (b} () () (e} " (&) (h) (1) ) () (m) Net Hoat per Interval Change in Fuel Change in Volume _ Fuel Exit Velocity, , _ Time, ¢ k Excess, &k, : Period, T Power, P from ?xcursion, Temperature Ris’eoin Fuel Temperature Rise, Mean Fuel Temperature, Volume in Time Total Change in 3 Absorbed in Pressure Change in Volume. « (Fps): Ve]oc.:ny If’ressure in {msec): 0.0001 ‘[(a)—- (At/2)] (msec): (MW)A:t/(C) Q" {Btu): Time Interval, AT (°F): ATn CF): 3 () TF CF): (g) + 1400 \nterval, AV’(ha): Fuel Volume, AVn_ift ) Shell Dilation, AVD (h3): Absorbed at Blogout Disk, () — (&) — )] 2.1--|n'.~d|a Jet, A}; Ar=20 ~{j),_1/15 0.4/(6) (d), _qe @), + @),y (€)/180 (/) x 5.7 x 10~4 (g) x 5.7x 10 7.6 % 10=5 [(n). — (n) Av, () » 103/(At x 0.025) (psi): 0.0216 (k) ~ 120} At/2 n G 0.060 o0 60 0 0 0 1400 0 0 0 0 0 0 20 0.001 400 63 30 0.17 0.7 1400 0.000095 0.000093 0 0 0 40 0.00299 134 73.1 161 0.89 1.06 1401 0.00051 0.0006 0 0 1 0 60 0.00496 80.6 93.7 468 2.60 3.66 1404 0.00148 0.00209 o 0 3 0 8;.1 0.00686 58.3 132 1,057 5.87 9.53 1410 0.00335 0.00543 0 0 7 1 100 0.00864 46.3 203 2,150 11.9 2.4 1421 0.00678 0.0122 0 0 14 4 120 0.01019 39.3. 338 4,210 23.4 44.8 1445 0.0133 0.0255 0.0012 0 25 13 140 0.01130 354 595 8,130 45.2 90.0 1490 0.0258 0.0513 0.0022 0 45 44 160 0.01158 34.5 1062 15,370 85.4 175 1575 0.0487 0.0998 0.0066 0 77 130 180 0.01033 38.7 1781 27,230 151 326 1726 0.0861 0.186 0.025 0 124 330 200 0.00659 60.7 2476 41,370 230 556 1956 0.13] 0.317 0.030 0 196 650 220 -0.00014 -2857 2459 48,150 268 824 2224 0.153 0.470 0.011 0.033 215 1000 240 -~0.00834 -48 1621 39,600 220 1044 2444 0.125 0.595 0.000 0.019 215 1000 260 —0.01467 -27.3 779 22,800 127 117 2571 0.0724 0.667 =0.021 0 185 740 280 ~0.01250 —22.9 325 9,840 54,7 1226 2626 0.0312 7 0.699 -0.032 o 127 325 300 -0.01758 -22.8 135 3,400 18.9 1245 2645 0.0108 0.710 -0.017 o 58 75 320 —-0.01630 —24.5 59.7 747 4.15 1249 2649 114 TABLE E-4. NUCLEAR EXCURSION CALCULATION FOR A HYPOTHETICAL CASE INYOLYING FUEL DEPOSITION IN THE CORE TO GIVE AN INITIAL RATE OF iNECREA.SE IN & g OF 20%/sec ' Conditions: Accident relieved by fuel expulsion from core forced by thermal expansion ' Fuel: Nc:F-Zer-UF4 Time Interval: 0,010 sec Mean Neutron Lifetime: 4 X 104 sec (a) (&) (© () (e) f) () (b) ) ) k! o m) (n) | Net Heat per Interval ' | Change in Volume . . . Time, ¢ k Excess, &k : Period, T Power, P from Excursion, Temperature Rise in LT Ri " Foel T . \Cl:h;:nge in :f"el Total Change in Absorbed in Pressure Chc::e ': \‘;o::nme Fuel Ex;: V;lomty, Velocity Pressure in {msec): 0.0002 .[(a) - (At/2)] (msec): (MWK:/(c) Q’ (Btu): Time Interval, AT’ (°F): ' ve AT"’“(‘;‘:‘;:”'E" m‘s"' ;‘"‘(o;‘; (;'“:'2’:0‘;’*' lm;‘;’;“’ AI; '(;’;‘;‘): Fuel Volume, AV _ (5+3): Shell Dilczl;ion, Blo‘:::msk' o 'i (:)“' ol 2.1-.in.-dic| Jet, Ag Ar=10 ~(),1/15 0.4/(b) (d),_y ¢ (), + @, _, (e)/180 n F . 57 104 (g) % 5.7 % 10=4 iy z};/D (F): AV (@) < 103/(As x 0.025) (psi): 0.0216 (k) - 120] At/2 631072 {m) ~ (,,_y] ¢ 0 0 o 60 0 0 _ 0 1400 0 0 0 0 0 0 10 0.001 400 61.5 7.5 0,041 0.041 1400 0.00002 0.00002 0 0 0.1 0 20 0.003 133 65:2 33.5 0.185 0.226 1400 © 0.00011 0.00013 0 0 0.4 0 30 ‘0,005 80 73.8 95.0 0.53 0.756 1401 0.0003 0.0004 0 0 1.2 0.03 40 0.007 57 87.7 208.0 1.16 1.92 1402 0.00066 0.0011 0 0 2,64 0.15 50 0.008¢9 45 109.5 386 2.15 4,07 © 1404 0.00123 0.0023 0 0 4.9 0.52 60 0.0108 37 143 665 3.70 7.77 1408 0.0021 0.0044 0 0 8.4 1.52 70 0.0127 31.5 197 1,100 6.1 13.9 1414 . 0.0035 0.0079 0 0 14.0 4,20 80 - 0.0144 27.8 283 1,800 10.0 23.9 1424 0.0057 0.0137 0.0007' 0 30 | 9 9 0.0160 25.5 420 2,901 16.1 40.0 1440 0.0092 0.0230 0.0010! 0 32 22 100 0.0173 23.1 648 4,740 26.4 66.4 1466 © 0.0150 0.0380 0.0025 0 50 55 10 0.0182 22,0 1020 7,740 ' 43.0 109.4 1509 0.0245 . 0.0623 0.0057 0 75 130 120 0.0183 21.8 1610 12,650 70.0 180 1580 ' 0.0400 0.1025 0.0114 0 114 280 130 - 0.0173 23.2 2480 19,850 110.0 290 1690 0.0630 0.1650 0.0222 0 163 573 140 0.0146 27.4 3570 29,700 165.0 455 ' 1855 0.0940 0.2600 0,033 0.017 215 1000 150 0.0096 21.6 4530 39,900 222.0 ' 677 2077 0,127 0.386 0.000 0.073 215 1000 160 0.0022 182 4780 45,950 256.0 933 2333 0.146 0.532 0.000 0.092 215 1000 170 -0,0066 ~60.5 4070 43,650 242 1175 2575 0.137 0,670 0.000 0.083 215 1000 180 -0,0148 -27.1 2880 34,150 190 1365 2765 0,108 0.778 0.000 0.054 215 1000 190 -0,0208 -19.3 1720 22,400 124 1489 . 2889 0,071 0.849 0.000 0.017 215 1000 200 ~0.0240 ~16.7 941 12,700 71 1560 , 2960 ' 0.0405 0,890 -0.010 0 ' 190 200 210 -0,0250 ~16.0 502 6,600 37 1597 2997 0.0211 0.910 ~0.025 0 140 | 550 220 -0.0246 -16.3 272 3,270 18 1615 3015 0.0103 0.920 -0.020 0 100 © 300 230 ~0,0233 -17.2 152 1,520 8 1623 3023 0.0046 0.926 ~0.015 0 80 200 240 —0.0217 ~18.5 89 | 655 4 1627 3027 0.0023 0.927 ~0.005 | 0 50 100 250 -0.0198 -20.2 54 116 0.6 1628 3028 0.0003 0.927 0 : 0 0 0 260 ~0.0178 ~22.5 35 -155 -0.9 1627 3027 - ~0.0005 0.926 ‘ 0 0 *Pjece shears out of pressure shell through heat exchanger outlet bell. 115 e o e o g = gy e g g T YR - W ™ a1 gy o vl TR TR ———— ey T T w3 = it~ g T o emn WO vy T TABLE E.5. NUCLEAR EXCURSION CALCULATION FOR A HYPOTHETICAL CASE INVOLVING FUEL DEPOSITION IN THE CORE TO GIVE AN INITIAL RATE OF INCREASE IN &_ . OF 40%/sec | Conditions: Accident relieved by fuel expulsion from core forced by thermal expansion Fuel: NaF-ZrF4-UF4 Time Interval: 0.010 sec Mean Neutron Lifetime: 4 x 1074 sec (a) (8) () (d) () (f) (g) (5) (i) (M | (®) (1) (m) (n) (o) Net Heat per Interval Temperature Fuel Change in Fuel Change in Yolume Pressure to Time,t k Excess, k_: Period, T Power, P from Excursion, Rise in Temperature ~Mean Fuel Volume in Total Change in Apcoihed in Pressure Change in Volume Fuel Exit Velocity yoiociry Pressure in Accelerate (msec): 00004 [(a) ~(Aw/2)] (msecl: (Mw); Q"(Br): Time Interval, Rise, 1°TPSITVC Time Interval, Fuel Volgne, Shel Dilation, Absorbed at u (fps): 2.1-in-dia Jet, AP 30-in.-dia Inconel Ac=10 () /15 04/(8) (), LD (D, + (@), AT F): AT (°F): TpCFE ave e AV R Avp, (#): Blowout Disk, [G) =B =] (i) 0.0216 (2 and Lead Disks ~120] At/2 (e)/180 2 () (6 +1400 (1) 575107 @X5TX107 7451078y ~(m 1 AV () x 10%/(Az x 0.025) (psi) o 0 o 60 0 0 0 1400 0 0 0 0 0 0 0.002 *200 63.3 16.5 0.09 0,09 1400 0.000057 0.000057 0 0.24 0 20 0.006 66.7 73.5 84.0 0.47 0.56 1401 0.00027 0.00032 0 1.08 0 30 0.00978 40.1 94,3 239 1.33 1.89 1402 0.00076 0.00108 0 3.04 0 40 0.01393 26.8 136.0 551.5 3.06 4.9 1405 0.00174 0.0028 0 7.05 1 50 0.01781 22.4 211 1,135 6.31 11.3 1411 0.00360 0.00644 .0 14.40 4 60 0.02157 185 362 2,265 12.6 23.9 1424 0.00718 0.0136 0.00100 25 13 70 0.02509 15.9 679 4,605 25.6 49.5 1450 0.0146 0.0282 0.0029 49 51 80 0.02812 14.2 1,373 9,660 53.7 103 1503 0.0306 0.0587 0.0087 90 165 90 0.03008 13.3 2,912 20,825 116 216 1619 0.0661 0.125 0.028 0 160 520 100 0.02967 13.5 6,108 44,500 247 466 1866 0.141 0.266 0.036 0.049 215 1000 10 . 0.02427 165 11,198 85,930 477 943 2343 0.272 0.538 0.000 0.218 120 300 120 0.01014 394 14,433 127,555 709 1652 3052 0.404 0.940 ~0.068 0.448 120 300 300 Fuel in core boils 121 ~0.0130 ~30.8 10,400 12,500 70 1722 3122 0.051 0.991 ~0.008 0.011 230 122 —0.0160 ~250 10,000 10,000 55 1775 3175 0.062 1.053 0 0.011 230 122 ~0.0200 ~200 9,500 9,500 53 1828 3228 0.073 1.126 0 0.014 300 124 —0,0245 ~16.3 8,900 9,100 50 . 1878 3278 0.087 1.213 0 0.014 300 125 ~0.030 ~13.3 8,300 8,500 47 1925 3325 0.101 1.314 0 0.014 | 300 126 ~0.036 “1n 7,600 7,900 44 1969 3369 0.115 1.429 0 0.015 320 127 0,043 -9.3 6,800 7,100 39 2008 3408 0.130 ©1.559 0 0.016 330 128 ~0.051 ~7.8 6,000 6,300 35 2043 3443 0.146 1.705 0 0.016 330 129 —0.060 -6.7 5,200 5,500 30 2073 3473 0.162 1.87 0 0.016 330 130 ~0.,070 ~5.7 4,400 4,700 26 2099 3499 0.178 2.05 0 0.016 330 131 ~0.081 ~4.9 3,600 3,900 22 2121 3521 0.194 2.24 116 TABLE E-6. NUCLEAR EXCURSION CALCULATION FOR A HYPOTHETICAL CASE INVOLVING FUEL ENTERING MODERATOR COOLING PASSAGES TO GIVE AN INITIAL RATE OF 60% Ak/k PER SECOND Conditions: Accident relieved by boiling of fuel in moderator passages and fuel expulsion from core forced by thermal expansion” Fuel: NaF-ZrF,-UF, ' Time Interval: 0.010 sec a4 Mean Neutron Lifetime: 4 X 107~ sec (a) (5) (c) (d) (e) (f) (g) (b) () () (k) (0 (m) (n) Net Heat per Interval Temperature Fuel Mean Euel Change in Fuel Total Ch . Change in Volume Ch in Vol F uel Exit Velocity P Time, ¢ k Excess, kex: Period, T Power, P from Excursion, Rise in Temperature an Fue Volume in ota ange 1 Absorbed in Pressure a::e ': do vme Velocity, .e;cnt-y r.e ssure (msec): 0.0006 [(a) - (Av2)] (msec): {(Mw): Q ”(Btu): Time Interval, Rise, _ Te;rpe(r:;ure, Time Interval, er‘i V(C;I;Te’ Shell Dilation ' Bl sor teD‘qL u (fps): n 'A];?'-d_')a Jet, ~120] Ar/2 ()/180 2 & () x 5.7 x 10=4 187X 7.6 x107° ()~ () _,] e x 10%/(At x 0.025) ' 0 0 L 60 0 0 0 _ 1400 0 0 0 0 0 _ 0 10 0.003 133 64.7 23.4 0.13 0.13 1400 0.000074 0.000074 0 0 0.3 0 20 0.009 44.4 81.0 128 0.71 0.84 1401 0.00040 0.00048 0 0 1.6 0 30 0.01497 26.7 118 395 2.9 3.03 . 1403 0.00125 0.00173 0 0 5.0 0.540 40 0.02089 19.1 199 985 5.47 8.05 1408 0.00312 0.00484 0 ; 0 12,5 3.40 50 0.02668 15.0 388 2,335 13.0 21,5 1422 0.00741 0.0123 0.0011 0 25 ' 14 60 0.03219 124 869 5,685 31.6 53.1 1452 0.0180 0.0303 0.0041 0 55 69 70 0.03699 10.8 2193 14,710 81.7 135 1535 '0.0466 0.0770 . 0.018 0 » 116 ' 300 80 0.03988 10.0 5961 40,170 223.0 358 1758 0.127 0.204 0.053 0.020 215 - 1000 Boiling Expels Fuel from Moderator 90 ~0.013 —30.8 4320 50,900 283 640 2041 0.161 0.366 0.000 j 0.107 215 1000 100 ~-0.0238 - =16.8 2380 32,900 183 - 824 2241 0.105 0.471 0.000 0.048 215 100 110 -0.0308 -13.0 905 15,800 88 912 2312 0.050 0.520 —0.004 0 215 | 1000 120 —0.0340 -11.8 388 5,880 33 945 2345 - 0,019 0.539 ~0.026 0 180 600 130 ~0.0353 -11.3 159 2,140 12 957 2357 0.007 0.546 —0.030 0 100 200 140 -0.0353 -11.2 65 520 3 90 2360 0.000 0.546 ~0.016 | 0 54 60 150 ~0.0358 -11.2 . 27 ~140 -1 959 2351 0.000 . 0.540 0 0 0 0 117 J P R — [ v o om M e ———— o mam e e et i B E e - L -119- f Aggendix F | EFFECTS OF A NUCLEAR ACCIDENT ON REACTOR STRUCTURE The detailed design of the reactor and related equlpment has been predicated upon the use of a stress level approximately one-fifth of the stress for rupture in 1000 hr. Thus the system is one in which no burst-type of rupture will be likely to occur. The type of failure to - be expected would be the result of either a fatigue crack or a leak caused by corrosion. In either case, the failure would develop slowly s0 that there would be ample indication of the character of the trouble before anything serious.developed. It should Dbe noted, however, that while this philosophy has been applied to the types of accident and haz~ ards to be expected, no advantage of this design basis has been taken ‘in, consideration of extreme accidents. Such accidents could take place only if a burst-type of rupture occurred, and such ruptures have been pre- sumed, even though no reasonable mechanism for causing them has been en- | v151oned. In addltlon, twn major tenets of the design phllosophy have been that the pressures throughout the systems should be kept low, particularly in the hot zones, and that all structure should be cooled to a temperature approximately equal to or below that of the secondary coolant leaving the heat exchanger. Great care was exercised in establishing the proportions of the designs presented in Tables F.l and F.2 to satisfy these conditions. The stress values calculated for the various stations in a typical design are indicated in Fig. 2.l. The stresses in the structural parts have been kept to a minimum and the ability of the structure to withstand these stresses has been made as great as practiceble. Thermal stresses are not indicated since they will be indeterminate, and it is felt that they will, to a large degree, anneal out at operafilng temperatures. of a nuclear acczdent, it is possible £0 envision fairly well the sequence - of events tha$ would lead to a failure in the reactor structure.. The Pirst 3;1 cons1derat1on in any such analysis is the -strength of the material of the _structure. Table F.3 presents strepngth data for Inconel, the structural “'material presently being considered. Note that the yield point and ulti- ~ mate tensile strength are much higher than the lOOO-hr stress rupture Jimit. Note also that the percent elongation. is substantlal and thus much plastic ” ':distortion wnuld take place before rupture would occur,k Since the struc- ture 1ncorporates substantial stress concentrations, the lccal yleldlng may be substantial, but the volumetric change in the pressure shell will be 7 'smfll. T e i b E | Core diameter (in.) | | 21 G B s o S S b i i e o e B =120~ TABLE F.1l. KEY DIMENSIONAL DATA FOR THE ART PUMP-HEAT EXCHANGER- : PRESSURE SHELL ASSEMBLY Core Island outer dismeter (in.) - 10.75 Core inlet outer diameter (in.) . Core inlet immer diameter (in.) Core inlet area (1n°2) Fuel System Fuel volume in core (ft3) (30 in. length)3 Fuel volume in inlet and outlet ducts (ft°) Fuel volume in heat exchanger (££3) Fuel volume in pump volutes (££3) Total fuel volume in main circuit (ft3) Fuel expansion tenk volume (ft3) (8%0) Expansion tank diameter (in.) Expansion tank ‘height (1n ) Fuel Pumps . . Centerline to centerline spacing ' 2 V'Volute chanber width (in.) 12 Volute chamber length (in.) 32 Volute chanber height (in.) _ 1 I e - Tmpeller speed (rpm) - 2850 ‘Estimated impeller weight (Ib) 11 Critical speed (rpm) | 6000 Sodium s ' S ' ' Speed %rpm) | 1300 " Impeller diameter (in.) Impeller inlet diameter (in.) Impeller discharge height (1n ) Sodlum.System ' “Expension tank volume (ft°) (1076) Sodium in beryllium passages (££3) Sodium in pressure shell (ft3) 3 Sodium in pump and heat exchanger (ft~) ‘Sodium in return from.moderator (££2) - Total sodium volume (ft ) Maln Heat Exchanger Volume (ft3) Nunber of tube bundles Number of tubes per bundle (11 x 12) Total number of tubes ‘Latitudé of header centerline (deg) Sodlumrto—NaK Heat Exchangers - “ Number of tube bundles , 2 Number of tubes per bundle N ' ' . 300 Moderator Regions : Volume of beryllium plus fuel (ft ) 27.2 Volume of beryllium only (ft3) 2h.0 Cooling passage diameter (in.) 0.187 Nunber of passages in island | 46 Number of passages in reflector - - 232 vor Bk FerPRER O o L™ ANAFONO N O TR T TR TR * * »B8ounb BJIT o N B ON R OOOC OO0 O O o e » W B £ O\ Voo W e ey vmwwm"-@‘ T T P m P —————r L T T L T P o I T Ty T T P Wy » & e o L L L e Lk e Elo o ENEERNGE L i Bk e e U G MG . R e e = .’5lffBot Fuel Pump Assembly'Dimenblans (1n ) -121- . MABIEF.2. nnmszbns OF ART DETATL PARTS gguatorial Diameters (in.) Island ; | Control Rod Thimble, ID = = = = = = = = = = = =« = = = = Control Rod Thimble, OD - - = « = = = = = = = = - - = « Be Islanfiv OD = = = = = = o = = - = .- = .-m - - | Tnner Core Shell, ID = = = = = = = = = = = = = = = = = Tnner Core Shell, OD = = = = = = = = « = ¢ = = = = = = Inner Core Shell Thickness = = = = = = = = = = = = « =« Reflector | o Outer Core Shell, ID = = = = = = = =~ — e - m - - - - - Outer Core Shell, OD =~ - - - = = = = - = - - = - - Outer Core ‘Shell Thickness = = = « = = @« = =« = = = = = Be Reflector, ID = = = = = ~ = - - - - - - - - - - - Be Reflector, OD = = = = = = = = = = = = = @ = = = = = ‘Reflector Immer Boron Jacket, ID = = - = = = = = = = = Reflector Inner Boron Jacket, OD = - = =~ = = ~ = = = ~ Boron Layer, ID = = = = = = = = = -~ « = - .- - - - Boron Layer, OD = = = = = = o 0 = = o = = = o5« - - Reflector Outer Boron Jacket, ID = = = = = = = - - e - Reflector Outer Boron Jacket, OD = - = = = = = = = - - ‘Reflector Shell, ID = = = = = = = = = = = = = = = - - Reflector Shell, OD - - - Pressure Shell Boron Jacket, ID - - - = - - e e m - - N ~ Boron Jacket, OD - = = = = = = = = = « = = - - - - ~ Boron Layer, ID - = = = = = « = = = = - = - ———- - ~ Boron layer; OD « = = = = = = = = = = = = = - - - - Liner, ID =« = = = = = = = = n = = = v - - .- .- = Liner, OD - = = =« = = = = = = e e e e .- e - Pressure Shell, ID = = = = = = = = = = = = = = = = = = - Presswre Shell, 0D - - == - - --------s o © Pressure Shell Thickmess - - - == - == - - - o= -1 Vértlcalwnlstance Above Fquator (in.) "*?3 Flopr of Fuel Pump TIniet Passage "W£m %T“’__ CeTete. 278, m of Lower Deckim = =~ = = = = = = = = = =2 -2 - - - 19 . Top Of LoWer Deck - - = = m = = 22 L aaEa 4ot {,Bottom.of Upper Deck = = = = = = = = = = - - “Top of Fuel Pump Mownting Flang . Top of Na Pump Mounting Flange - - - - -,—r;_-q-7-‘5m-'-'f” ’ Sbaft | ' - S Fe ‘Thrust Bearing Jburnal, OD = = = = = = = = ‘= = = = = = OD Between Bearings - = = = = = = = « = = = = = = = = Lower Bearing Journal, 0D = = = = = = = = =« « « = « « Seal Washér Journal, OD = - = = = = = = = @ = = = = = 0D Below Seal . e e e e m e e e e me..m——m——- sk s s dobieailo. . ~122~ Thread for Impeller Retaining Nut - - = -~ = = = ~ - - = 1. 687 - 12 N. S. Thrust Bearing Height from Equator ' - - = = = = = = ~ = 47.812 | Lower Bearing Height from Equator = = = = = = = = - = = 35. 812 Distance Between Bearings « - = = - - = = = - =~ === 11.00 Distance Between Thrust Bearing and Impeller _ Locating Shoulder - - = = - = = = = = = = - - == - 25,437 Gver-all Shaft Length m - mmmm = - m— e = = = 32,437 ImPeller 7 ' SR TR o Fo et D ~<===- e e m e e et e i a e, 5.75 ”In_le't, e e e e e e e e e - -.—_T‘__ r,,,...,.‘.‘_., 3.5 'Dlscharge Passage Helght - e e mm === -e ===~ 1.0 _'No. of Vanes - = = = =« = = = - .- e e e e === D Axisl Distance from Top of Discharge Passage o .. Pop of Centrifuge - - = = = = ~ - _—— - - - - - - 225 "‘Shaft Locating Shoulder - = = = = = = = = = = = = = 87 Inlet Face - = = = === - ===~ -====== 2.00 Pump Boiy ) o G D e 0D of Locabing Journal Outside Lower Bearing - - - - - - 5.935 OD of Flange at TOp = = = = = = = = = = - = - = wome Q.37 No. of 1/2 - 20 Machine Screws in Flange - = - = = - - 8 Vertical Distances from Lower Face of Mbuntlng Flange Bottom of Thrust Bearing = = = = = = = = = =« - = - - == T50 Face of Shaft Seal Washer - - - - - - -~ = == === -13.3l2 Bellows Seal Mounting Pad = - = - = = - ~ = = = =~ = = = 15.34% Lower Face of Boron Jacket = = = = = = = = = =~ = == = = 20,750 . Top of Impeller Discharge Passage - = = ~ = -~ = = = = 26437 Maln Eeat‘Exchanger (in.) . S ~ Tube Centerline to Centerllne Spacing = = = = = = = = = = = 0.208 Tbe OD = = = = = = = = =& = = = = @ = = « =« = = = « = - - = 0.1875 Tube ID - = = = = = = = = - = e et m e - - 0.1375 Tube Wall Thickness - = = = = =« = =« = = = = = = = = = = = 0.025 Tybe Spacer Thickness - = = = = = = = = = = = = = = = = = = 0,020 Meson Tube Length « = = = = = =~ = = = = = = = = = = = = = = T2 Radius of Equatorial Reverse Bend -« - = = = = « = - = === 1.5 Inlet and Outlet Pipe ID = = =~ = = = = = = = = = = = = = = 1.65 Inlet and Qutlet Pipe 0D = = = = = - - m - e w we === 1.95 Header Sheet Thickness = - = ~ = = = = = = = = = = = = = = 0.25 .Hea&er Sheet Inner Radius - = =« = - - - - e e = e e = w 2,25 Moderator Circuit Heat Exchanger (1n.) | o Tube Centerline to Centerline Spaclng - m == - - 0.208 Tube 0D = = = = = = & = o = = = = = = = = = = = - - == = 0,875 Tube ID - = = = = = = = = = = = - - - - - - mm = === 0.1375 Tube Wall Thickness = - = - = = = - e r e e e - - - -~ 0.025 Tube Spacer Thickness = = = = =« = = = = = - === === 0.020 Mesn Tube length - = = = = = = - = == - === === - 28 : CL m o e SR I R e, SRRt kT L S R e ke t 1 -123- TABIE F.3, STRENGTH DATA FOR INCONEL TESTED IN A FLUORIDE MIXTURE Temperature, F | | 1200 | 1500 Ultimate tensile strenmgth, psi ) .;;: ‘”§5,OQOYhW ) | 21,000 Yield point, psi o - | '26,500ih o 13,000 Stress for rupture in 1000 hr, p61. | ;i7;660.t'- : 72100 VfiElongatlon, % 'fl”““jjv ’1,- '*.”i° "39 7 ’”7? L r”'ss | RN Stress Calculations | The principal element in +the reactor structure is the maln pressure shell. It is a spherical vessel 1 in. thick with an internal diameter of 52.5 in. During normal operating conditions the temperature of the pres- sure shell will be 1200°F. The stress in this shell for an internal pres- sure P is given by - ' R o o o s =L ELELEfiifiL = 12. 9 p . | . 2t .2 , The stress, S, and pres sure, P, are in p51 and the radzus, r, and the shell ~ thickness, %, are in inches., In re-examining the stress expression and ‘Table F.3 it is evident that the main pressure shell would begin te yield at a pressure of about 1000 psi at sitress concentrations such as those which will be present in the ligament between the heat exchanger inlet and outlet header pipes. It would be expected that progressive local ylelding . 'would contlnue to take place in these regions until failure occurredj some increase in pressure would occur in the process. Another major weak area would be the flat structural region at the top of the shell where the pumps and expansion tanks are 40 be located. While final detalled designs are - not yet avallable, it does seem certain that local yielding would begin in o this region at a pressure “of the order of 1000 psi. It also seems likely '>: f;!that one result of this yielding would be that the pump casings would dis- '*.Atort'and cause_the 1mpellers to seize in the casings and the pumps to stop. - Further distortion would tend to make this flat region sphérical until high v;ocal_sfiress}ConCentramlons had.produced large amounts of yleldlng to the 1y, that in the south‘heat exchanger”outlet nterline diameter of 36.4 in. The total cross aments in _he”main pressure shell wouldibe 60 in.% welded %o’ the pressure ‘shell between the heat exchanger outlet plpes.rfim;f " V’::W'w I R R T rer— e nd thls’would'be augmented.by approxlmately Lh in.2 of re1nforc1ng patches- ISR T T T T R R R T T Rk RN Ben . e . coblk R B gk g L omiD L @e i o oo & b L 3 v s B B e Bk _lgu_ The shear stress in this region would then be p.' . s _k2_ fT3 _pma® T2t 2—%— o iA | CEwa Px0786£3611-)__ . %o+ 1L | = 10,000 psi . This would glve a ten31le strength of 1k, 100 psi, which represents an \ average stress in the ligaments. The peak value around the edge of the holes would be spproximately twice as great so that yielding would begln there. The stress would then distribute itself uniformly across the liga- ment, and further plastic deformation would follow until a tear would‘begln and rupture would finally take placea The stresses in the outer core shell are of partlcular 1nterest be- cause expansion of that core shell.would increase the fuel volume in the reactor and as a consequence give an increase in reactivity, If the core shell were not supported by the moderator, the stress in the shell would be given by _ Pr P x 10.,5 - - S = = = Ze P - . ' loat lo X O .125 . 5 5 Note that a factor of 1.6 1nstead of 2 was used in the denomlnator because‘ the’ radius of curvature of the core shell at the equator in the vertical Plane is greater than it is in the horizontal plane. This core shell is. actually supported by flattened Inconel wires inserted betlween the core shell and the beryllium reflector and spaced in rows on l~in. centers. These strips are used primarily to maintain the spacing between the reflec- tor and the shell and hence to ensure the proper coolant flow passage open- ing. It is interesting to determine the radius of curvature of the shell when deflected between the supports by a pressure that would give the same stress in the core shell as would prevail in the main pressure shell. If | a simplified cylindrical geometry is assumed, the relation for that case is | g _Bx 25.7 Pr o T 2x 1. 7 0.d25 7 | | | 7 | | from which a radius of 1.6 in. is obtained. Thus the core shell would deflect to the point where it would have a radius of curvature of 1.6 in. between supports° The effect on the core volume of this distortion coupled" with that of the immer core shell would be relatively small, and would give -1 volume 1ncrease of about 1% for a lOOO-ps1 pressure. The stresses in the 1nner core shell would be similar to, but lower than, those in the outer “core shell because its dlameter would be roughly onewhalf that of the outer' core shell. T R AT TR W T A — N FREE NT PR R e S T W S U T ST T T Y g T i e e S e M follows: -325- The possihility of buckling and collapse of the tubes in the main - heat exchenger has also been consildered. The stress in the tube walls from an external pressure is '. ‘; Wi oRL D e e O 02 0.025 fi hP. Note that thlS stress 1s very mnch less than that in the main pressure shell so that it is very unlikely that the heat exchanger tubes Would col- . lapse 1n the event of a pressure surge._' T w3~~. The force reqnlred to blow the pump body out of the pump casing is | of interest. TP it is assumed that the main pressure shell would begin to yield locally ‘under & pressuré “of 1000 psiy the force actlng on the pump - body may-be calculated in the follow;ng way ‘_Pump body blow-oub _fo Shaan 1000 x o 786 x 5.75 = 26,000 lb. The cross—sectional area of“the pump body attachlng screws is over 2 1n.2, and since "this is the vweakest point in the system restraining the pump ‘body from blow-out and since it is in a cold region where the tensile strength of the material is of the order of 90,000 psi, it is clear that this does not represent a weak point in the system. While the thrust bear- ing at the top of the pump impeller shaft might fail and allow the impeller shaft assembly to move upward several inches, the pump impeller itself "would.keep the shaft from moving on out through the pump body. If this failure took place there would be free communication from the reactor core through the pump 1mpeller bore 1nto the fuel expansxon tank. The volumetrlc stlffness of the main pressure shell is of interest from the standpoint of determining the pressure at various stages during a nuclear accident. The change in volume, AV, in cubic feet, is glven as where Ar is in inches and represents - er of the 1ns de :ngpof the pressure shell, Note that a modulus of elastlclty, E, of 21 x 10 'f'was‘“Séd’s1nceAthe pressure shell temperature should be a little below - 1200°F. From this it follows that “the increase in volume inside the pres- sure shell for a J00O-psi pressure surge would be 0.076 1. It should be noted that the pressure shell in some respects would be stiffer than it is ‘assumed to be here because of the presence of the liner in the heat ex- changer region and at the south end. On the other hand, the system would s e it s ol g T ey R R T T AR T e T T T e St v O T ST T T R i e b i il e i i g 106~ fAbe less stiff because of the presence of flat regions in the pump and 'Vexpan31on Tank region. It seems likely that a rough approx;matlon to the stiffness of the system has been given. Destructlveness of Pressure Shell Fragments - The velocity of the fluld escaping through & crack if rupture should occur at lOOO psi is of interest. Th1s can be cemputed as follows. «[ -J64 b x 2999§§5£&5- - \|u6 300 = 215 fps, YL e ‘ ‘wfieie H 1s the pressure (1n lb/fta) d1v1ded'by the densmty of the fluld (in Ib/ft ). This velocity of 215 fps is certainly a reasonable one and should not give any particular trouble. It is also important because it represents the maximum possible veloczty of a fragment that might be broken - out of the pressure shell in the event of a hydrostatic rupture. It is interesting to estimate the maximum kinetic energy that could exist in an escaping fragment for this case. The work or energy put into the fragment ~in the rupture process should equal double the area under the PV diagram for the pressure at rupture and the volumetric increase inside the pressure shell up to the rupture pressure. This triangular PV diagram area should . be multiplied by a factor of two to allow for the elastic overshoot or re-' bound of the shell accompanylng the pressure rellef- o Faei ARG Wbrk ‘RAV 1hh$ooa x 03 076 = 10,900 ft-lb | It is also of 1nterest to calculate the wezght W, of the mlnlmnm _ragment . requlred to carry off all the rupture energy.‘ ThlS is K. E. = 10,900 = 32 5 §215) W 15 1b. Thus it appears that the most destructive fragment from a pressure shell fallure at 1000-psi pressure would be a 35-1b fragment that would leave the reactor at a veloeity of 215 fps with a kinetic energy of about 11,000 ft-lb. If this fragment were of l-in.-thick Inconel, its area would be 260 in.2, If it were to strike the lead shield and all of its momentum were imparted to a lead fragment of the same area having a thickness of 7 in., the result- ing velocity of the lead fragment could not exceed 20 fps if no allowance were made for the energy absorbed in rupturing the lead. ’ In re-examining the above analysis it appears that a number of modes of failure might be encountered. If no special provision were made to re- lieve the pressure it seems likely that the north head would deflect plas- tically to the point where the pump impellers would seize in their casings. Thus whether the increase in volume associated with this deflection would be sufficient to relieve the pressure associated with the nu¢lear reaction would depend upon the particuler accident. Certainly, if carried far enough, rupture would occur. If rupture did occur it seems unlikely that TR TR . T I T T T e e e W T rFran'-z‘ - & LY i VT T T weT o f EA B RS B i i ok -127- a large fragment or, for that matter a small fragment, would be blown out of the wall of the main pressure shell, Rather it seems likely that the well would split and open up to relieve the excessive Pressure. It seems most likely that this splitting would occur in the ligaments between the heat exchanger outlet tubes at the south end of the reactor. In that event about a 5/&-1n. stretch in theee llgaments would prebebly'occur prior to failure. , , Pressure Relief Hecha.nisms ' ' Provision can be made in ‘the design to relieve a v1olent pressure surge from a nuclear accident in & number of ways. It would be possible, for example, to put a poppet valve in the fuel region just ashead of the fuel pump inlet so that it would open and vent to the expansion tank. The utility of such a move is doubtful because the volume available in the ex- pansion tank is probably not quite adequate to take care of an extreme nu- clear accident. While an overflow will be connected from the top of the expansion tank to the emergency fuel dump tank, and thls "line will be kept heated at all times, its diameter would have to be at least 4 in, if it were to be effective in relieving a pressure surge in a nuclear accldent. A second provision for pressure relief could take the form of a frangible diaphragm. If this frangible diaphragm were placed at the bottom of the reactor it would be in a high-pressure region and would be subjected to a - pressure of approximately TO psi under normel operating conditions. Thus ... it would have to be designed to rupture at a pressure of at least 400 psi because the ratio of the ultimate tensile strength to the 1000-hr rupture strength is apprOX1mately 4 at 1200°F. Thus, if a factor of safety of k were used in the design of the disk, under normal operatlng conditions the pressure at that point would have to be roughly 16 times as great to produce failure, or about 1000 psi. If, on the other hand, the diaphragm were placed at, or close to, the fuel pump inlet where the pressure would normelly be, perhaps 5 psi, it could be designed to rupture at a pressure of 60 psi and still have a factor of safety of 12, stress-wise, for normal operating con- ditions. Yet another device that might be used would be to employ a weak seam in the attachment of the whole north head or some part thereof. This .;‘,@fiagaln would have the dlsadvantage that it would operate under high stress - for normal”operatlng condltlone if 1t were designed for rupture at a pres- d_with 5 - , Add;tlonalfvelume could be ;and he eéd regzon and elnce thls zone is pnly partlally fllle ' prov1ded readlly w1th llbtle penalty;ln shleld'welght by leav1ng voids in | L‘fvarlous places in the 1rregular region eround the pumps and header tank.< ~ 'shielding placed in the re-entr e F " “this region and could be omltted., "haps'o;S 43 could be readily arranged. Yet another way in which expansion cormers is relatlvely_lneff?ctlve in® In this way an explos1on“volume of per- volume could be prov1ded would be to ‘place a blister on one “side or the top -or the bottom of the pressure shell., The side or bottom of the pressure shell is particularly attractive from this standpoint because a drain line T T R R [ T T T T ——— A R T F T ORIE T v = e E course, a large diameter pipe could be extended directly out through the ‘s1on relieved by boiling at the peak of the pdwer curve is of interest. tion of pressure shell fragments that could pierce the wall of the inner 'pressure of between 1000 and 2000 psi and will certainly rupture generally ~at a pressure not to exceed 4000 psi. The second point is that if a small il eos G R el Sk, o - tratlng, a fragment of such a shape seems quite out of the question. A - sphere has greater penetratzng power than any irregular shape likely to be ‘ejected from the pressure shell. The old Krupp formula for the penetration - of wrought iron plate by ¢ast iron cannon balls appears to be roughly appli- . "#i fih@proachlng the prdblem from a dlfferent stanipoint and examining the . pressureshell structure carefully, 1t appears most likely that a circular "_fuel from the core and that all of the pressure relief was obtained by 'nuclear pcwer surge caused by increasing kppp at a rate of 40% per second - were dbtalned, as shown in Fig. E.l. The fragment would be accelerated -128- o Y from the blister to the fuel dump tank could'be‘ea51ly installed. Of shield. However, since a pipe of at least 4 or 5 in. in dismeter would be reqnired this would constitute a serious gap in the shaeld'Whlch.would be dlfficult to block off W1th a Shleld patch. Whlle 1t is hard to see “how it could occur,'an extreme power excur-‘ At first glance, it appears that such an accident might lead to the ejec- tank. In re-examining the stress analysis results, together with the curves of Flg. E.l and the data in Tables E.2 to E, 6, two points are evident. The first is that the shell will probably rupture locally under an internal opening having a diameter of say'h in. appears, the rate of ejection will - . ST T T TR T T g . 3 4 il g s ot ki Jb s s L s R B T ‘not be high enough to keep the fluid pressure from rising so much that elther the hole would tear open further or a second fallure would occur. In elther case it would not be possible for all of the energy to be concen- trated in a ‘single small fragment travelling at a very high velocity. In E fact, from Bernoulli's equation, the limiting velocity for the fluid column (or any fragment) if the peak pressure were 4000 ps1 would be: _i Yi:e--{EgH 450 fps.flv- | The penetratlng power of progectlles varies w1th thelr shape, hardness and strength, mass, and velocity. While an ogival shape 1s the most pene- cable to thls case, i.e., where V is in fps and t and d are in 1nches° The velocity of & 6-in. sphere required to penetrate l/2--:1.n° plate is then 0.5 \2/3 5 V = 2000 (-'6'2) 380 fpS . ..1 section 36.4 in. in diameter would be sheared out through the heat ex- changer outlet belt. By using the numerical method outlined earlier and by assuming that there was no loss in reactivity until boiling began to expel shearlng ‘out the bottom of the pressure shell, the results for a very severe untll 1t had passed nearly through the lead reglon. It was assumed in the B et e A i i e RS ,_,fitank wall._ ~129- calculatlons that yielding 1n'the Inconel shell toak place until it had moved through the clearance for theérmal insulation and contacted the lea& gamma ‘shield, at which point rupture of the Incomel occurred. No allow- ance was mede for lateral escape of fluld into the space between the lead and the Inconel shells, nor was allowance made for the energy absorbed in shearing the lead. Note that the velocity of the Inconel and lead frag- ments at the end of the expension is only about 40 fps. While the failure might start first on one side so that the disk might tilt during the process of failure, this would.prdbably result in opening a flap in the pressure shell without tearing the flap loose from the shell and ejecting it as a progectlle. The worst case would occur if the flap were ejected and tilted in flight so that it would strike the tamk bottom edgewise. The velocity required for such a projectile to pierce the tank wall can be estimated from the relation for the energy required to punch holes in plate, i.e., 2 fl-—- ‘bEBS 2g If the Inconel disk is l in. thlck 1tS'welght, W, will be aboutjgfio 1b, and the periphery, B, of the hole 1t-w1ll punch will be sbout 76 in. The ultimate tensile strength, s, of the steel plate should be about 60,000 psi. The velocity, V, required to pierce the tank wall having a thickness t = 0.5 in. would then be !2t2§sg _ \J0~5 x 76 x 601900 X 3242 W - 3-2 5350 - 530,000 = n. v gsl £ps. This velocmty is comparable to the 380 fps found fer the 6-in. sphere from the Krupp formula. While the mass of the lead piece sheared out of the gamma shield would be large, it would be soft so that most of its energy upon impact would go into deforming itself rather than 1nto piercing the examlnatlon of the results of the numerlcal analyses dis- 1figcloses fihat'the reactor should not be damaged by a nuclear accident of the _‘type shown in Fig. E.2 if the initial rate of increase in kepp did not ex- ceed 10%‘Ak/k per second. More severe accidents that did not result in - boiling of the fuel would probably cause distortion in the pressure shell that would interfere with pump operation and possibly cause leaks. While o .ea llght frangible d;a@hragm could be used to relieve such accidents so that . the pressure shell structure would not be damaged, any such incident would . involve SO severe a temperature overshoot that further operation of the - reactor would not seem to be prudent. Thus there seems to be little point - ;;w;§1n maklng the design compromises necessary for a light frangible diasphragm. rjfThis is particularly true in view of the very remote probability of acci- dents in which it would be of benefit and the much greater likelihood of troubles that it might give in the course of routine operation. For this reason the analysis in this report and the curves in Figs. E.l and E.2 ‘have been prepared by presuming that up to pressures of 1000 psi the only T T e e T I T AT ) - i e e e ke g T T e TR T T T T T e e S - - . L ‘. : A . T i R . L . u s an i Dl e i i e s b i . G it e o BRRRE L i s ic L RE b sk ok AR i d -130- avenue of escape of fuel from the core circuit is through the clearances between th.e pump impellers and their casings and through the centrifuge d:l.scha.rge holes into the fuel expansion tank. It should also be noted . that the worst casé in Fig. E.2 that appea¥s, on this basis, to leave the reactor undamaged is that for an initial rate of increase in keff of 10% ak/k per second for which the volume of fuel expelled is about equal to the free space normally available in the expansion tank. For cases tending - to give pressures above 1000 psi, it was deemed best to :.ncorporated. a circular groove in the bottom of the pressure shell so that the pressure would be rel:teved by failure in that region. The fuel would be egected | d.ownward and. a mim.mum of damage to the system would result. : 3 e g o B e e -131- - Appendix G EXPOSURE HAZARD CALCULATTOfis' -~ Criteria - The caleulations of the exposure hazard were based on a total internal exposure of not more than 25 rem to any internal organs (pone, thyroid, lungs, G. I. tract, kidneys) over the lifetime of the people exposed. Since the general population is being considered, the lifetime 1s taken to be . 70 years. The group of people to be protected includes children, the preg- nant, those especially radiosensitive, those with large previous exposure Tecords, and those occupationally exposed. If D is the exposure rate to the organ, then t=70 Yy 25rem 2 \ Dat , where time, %, begins with the intake of the isotope mixture. The intake may be either by inhalation or by ingestion. \ Basic Formulas The exposure rate s Dy to an drga.n is the sum of the':'é;'x;édsure rate from material a, D, the exposure rate from material b, Dps; etec. Therefore, D = 21;' D; (rem/day) , _ where D, is the exposure ‘rate toa bod,y bréé’.fi resulting from its content of - isotope i. The exposure rate from isotope i at time t then is ' D:=J Q; (ue) x 3.7 x 1% (dis/sec. mc) x 8.64 x 10 (sec/day) Y & 5B (ev/ais)[REE]N (vem/red) < 1.6 % 1070 (erg/iev) .e- _7\17[111 (g) x 100 (erg/g-radfl ‘ o % BEEN Az R T e o (rem/day) , 1. Thisl--:seef.ién prepared by T o Burngtt,;« ORNL Health Physics Division.. T I g pmeee g e PSSR T T T T o e e e nie e i L s b R b L e e - 132~ at tlme, *, where Q = amount of 1sotope i in organ Cuc), m mass of organ containlng Ql (g), | ZE(BBE)N energy, E , dissipated in organ of mass m :E‘rom each disintegration of isotope i, weighted for biological effectiveness 2 BBE » and nonumformity of d:.strl”oution s N, " .;; (BBE l for X, fi ’ x: a'nd' e” b lO for a -4 i‘ " G E e B AR TYToeT ny Tem oF '- = 20 for reco:u.ls 1for Y and X = I 5 for a, 3 gt, e”, and atom recoils in bone 1 for all body organs except bone) s | x‘j_; = elimination rate constant for isotope i from the body organ = 0.69%/74, where T; = effective half life of isotope i from body organ T m ; Where Ty is the b:i.olog:.ca.l ha.lf life and T, is " the rada.oactlve ha.lf life. (In the expression of the exposure rate the unit "rad" is the dose from deposn.t:.on of 100 ergs per grem of tn.ssue) The ’cotal exposure +to the body organ from isotope 1 over any perlod t following a sn.ngle inta.ke is % t N [ e Di= S D d.'t-—Dl 0.695 l-e ), ‘b=0 ' where 1):L is the initial dose rate from isotepe i over a period of 70 yea.rs (= 25,568 days); thus - - - ‘DTOy - 1° I3 [1_8-7\1: (=70y in days)‘] i 1 0.693 e A m—————— e 2 ¢ o RRREL L L e ‘ent doses are then summed for the organs affected. Based on the sum of ;'1nhaled to give a total dose to any orgen of 25 rem in 70 years is calcu- ¢ -133- The. total exposure to the body organ from a mixture of isotopes is then Z D:L , in which the amounts of the components i which reach the organ following intake are considered, For a mixture, the values of Qi will depend on the composition of the mixture and the uptake of components to the organ.l Method A.unlt 1ntake (1nhalatlon) of l,yc of the mixture of interest is con- sidered. The total dose of each component of this mixture is calculated by using its fraction of the total of the mixture cons1dered. The compon- these total doses, the number of microcuries of the mixture which can be lated. This nuflber of m¢crocur1es 1s the total permissible inhalation 25 rem, 0 y 1ntake for a s:.ngle ., exposure 'bo t.he miz.ture considered (MPI sin gle exposure) | Meteorologzcal calculatl sfiof 1nhalax10n exposure (based on Sutton s equations) yield total integrated dose values in curie~geconds per cubic meter as the integral of concentration with time : . TID = J~x (c/ma) at (sec) =_curiesosec/m3 The TID tlmes the breathlng ra e in m5/sec gzves the inhalation intake in curies. R T T T R e S The maximum permissible t@tal‘integfated dose in curies-sec/'m3 is - 25 /70 E : MPISQ./ i o - mm S TTER ¢ E . _where’ BR is the breathlng rat Varlous ‘values for breathing rates may be % ,.?fjgconsidered and are approprlate to the conditions of rest exercise, excite- azvalue of : & _ iters per’mlnute, ‘which corresponds 1o moderate exerelse and possible excltement {since the appllca$1on lS to 1nstances of a reactor accldent) x(fik’mn) x 1073 @Sil) _ r. ' 3 i A =5 0™ e ¢ MPI_..- curies) e (' =2 x 10 5 MPI 5/7 curles-sec/m . S 5=x 0 | ; g ror e - P 1 1 1 i P L . R i T S A T S S S B R i e S R e v S S dianias « B R & BEERRE o s vk B R Gl S R e e . Lot bbeilioh & i B ek o B e A ARG K R R R i b R o B Y k B 5 ‘ o s, . e - . . . ‘ where % is the fraction ef the mlxture that is 1sotope i and fa¥.1s the - -134- If a unit 1nhalation intake of lluc is con51dered, % x faz Mc in organ per 1nhaled,uc, ¥ ¥ 3 vn-'vv o fraction of the inhaled isotope i that is retained in the critical organ. The critical organ is that organ of the body which receives the isotope '_lthat results in the greatest to the body. In most cases it is that organ which receives the greatest damage. However, some organs are less essen- tial to the well-being of the entire body. Ususlly the critical organ is . that which receives the greatest concentration, but there is considerable varlatlon in semsitivity. The critical organ depends also on the mode of _intake and.may change W1th tlme after 1ntake. ' Calculational Procednre | The com@onents of a given mlxture of mixed fi351on products are ~ ‘grouped by the body organs.affected. A tabulation of %, Taqs and ::Z: E(BBE)N values is then made and the corresponding values of D are cale jculated.by using the m of the organ. These can be summed to glve the ini- tial. dose rates to the organ. - The values of D; are then tabulated, together with those for T;. By combining these with values of (l-e }&t) for various times, t, follow1ng intake, the total doses Dt aré found. The values of Dt are then summed, and the totals are used to calculate MP125/70 : ‘ Values of % will vary'W1th the mixture of 1sotopes consldered, Whlch in turn depends on the time of reactor operation. This is also sensitive to the eholce of isotopes for the mixture to be considered. " The components of the selected mixture were chosen because-af their jknown hazard and for half lives generally in the range of 20 hr to 20 years. Shorter~lived isotopes would have Df values small enough to ignore, and longer~-lived. isotopes would have small % values for the operational times of interest with high-power-density reactors where the reactor is operated with 8 maxX. '1rredlat10n time of the order of 1000 hr.. Buildup of longer lived materials is small for ‘these times. Parent-daughter relations also influenced the choice of some isotopes and their inclusion. Sinee the - composition of the mixture chosen will vary with time following a reactor incident because of the differing decay rates of its components, there would , be correspondlng different MP125/70 values at different points sufflclently far downW1nd for the airborne tran51t time to 1ntroduce these decay dlffer- ences. As a first spproximation the decay differences can be ignored, since the mixture chosen undergoes small decay for the relatively short times of e '1nterest for most reactor accldents for which the radius of hazard is small. ' N ST TR e R TIE e w pr » ™ 5 - o e s i e o R el iy e M | i s, diGn i [ Lk : - . . ~135- For this fzrst approximamion, the MPIEB/TO based on initial concentrataon can be used. It is prcbable that the effect of decay may be compensated by increases of the relative %y values and, since the longer lyved mater~ ials are more hazardous, the net effect could be smaller MPI25 values at more distant p01nts where dec&y'would be 51gniflcant. The E: E(RBE)H values used are-either those given by the Internatzonal Committee on Badiologlcal Protection (ICRP) or are those caleulated as pre- scribed in the ICRP Internsal Dose Subcommitiee Report of December T, 1953 (Ko Za Mcrgan, chairman). The prescribed formule is < 1/2 /2 ZE(BBE)N = Z_ £¢Ey (Lwe” ) + 0.33 £, B, (1.. o ( Efi )Nfik Jsk,m J 4 A whlch conslders the decay scheme of ‘the 1sotope 1nvolved. 'In this formula REE and N are as prev1ously defj_nea_ and o - - fE’ = fractlon of the dlslntegratlons of the gfih type that result 1n the em1551on of a gamma‘or X photon of energy:Ea/, ad , fB fractlon of the dlSlntegratlons of the th'type that o §.H7result 1n the emlssion of a beta ray of maxnmum energy Efik » ' fé_ = fractlon of the;dl 1ntegrat10ns ‘of the m' i type that -m iresult in the em18310n of a conversion electron of energy'E., ’ 53 ==total coefflcient of absorptlon mifius Compton scaitering coefficient in em™* for tissue for photons of energy85a2, Jvly 25, 191*7-» W, 5. Tewmer et al., Preliuinsry Design and Performnce of Sodimn-to-\ Air Radla:tors, OREL-15()9, Deca 26 1953.:. | P. L. m1L, AJlmli':Metals Aren Sefety Guide, Y-Bll, Deco 13, 1951, R TR TR 6T TR I Ty - v e : o TEtETETE T T g e t T T T R e PR T b e bR R T .-1&6-”5'.m~ o \g | L. Wilson, Building Codes and Other Criteria, GM-la'T (A,EC), (no date) - A. P. Fra.as s Three Reactor-Heat Ebccha.nger-Shield mements for Use o with Fused Fluoride Circulating Fuel ORNL Y-FlB-lO, June 30 1952. = O. G. Sutton, Weather, Apr. 1911-7, p. 108 0. G. Sutton, Proc. ROy. Soc. ‘London 135A, 155 (1932) W. F. Davidson, The Disperston and Spreading of Gases and Dusts from Chimneys , Ind. Hyg. Foundation Amer. Trans. Bull. 72013 (l9h9 'F, W Thouas, TVA, Wilson Dai, Ala. Plume Observation Watts Bar Steam”" Plant (1952) unpubl:.shed manuscript 'P. B. Stockdale, Geologic Conditions at ‘the Osk R:I.dge National (X-lO) " Area. Relevant to the Disposal of Radioactive waste, ORO-58 (Aug. 1, 1951) © ANP Project Qnarterly Progress Reports, ORNL 1816 1771 1729, 1692, | 1649, 1609, 1556, 15155 1539 Ce Es Win't;ers a.nd A. M. Welnberg s A Report on the Safely Aspects of the HomOgeneous Reactor Experimen‘b ORNL-'BI (June 20 ’ 9505 '