{t‘ | Y CENTRAL RESEARCH LIBRA R;f ¢ - ‘: DOCUMENT COLLECTION @ 3 ORNL 1615 N Reactors-Research and Power 20 a) ; e N L - " * r L d F.i §i i iy - v i — - —— - 8 % ] g», % b l'l i 3 yy5k 0349579 b 9 3 ““'""Hmmiuuumrnn!!;El:!ujiafiéi 4 1 | -t - W WO LR AES 4 Desald = e bk - . [ =i} L WAL=l - CRITICAL EXPERIMENTS ON DIRECT CYCLE AIRCRAFT REACTOR Dixon Callihan R. C. Keen CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON If you wish someone else to see this document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY OPERATED BY CARBIDE AND CARBON CHEMICALS COMPANNY A DIVISION OF UNION CARBIDE AND CARBON CORPORATION 1] m POST OFFICE BOX P OAK RIDGE. TENNESSEE Index No. ORNL-1615 This document contains 95 pages This is copy 20 of 154 Series A Subject Category: Reactors - Research and Power. CRITICAL EXPERIMENTS ON DIRECT CYCLE ATRCRAFT REACTOR Work by: Dixon Callihan E. V. Haake (Consolidated-Vultee) R. C. Keen (Louisiana State University) W. G. Kennedy (Pratt & Whitney Aircraft) J. J. Lynn Dunlap Scott D, V. P. Williams Preparation by: Dixon Callihan R. C. Keen ' 0CT 22 1953 DATE ISSUED PHYSICS DIVISION "A. H. Snell Director Contract No. W-7405, Eng 26 —_— OAK RIDGE NATIONAL LABORATORY |f MARTINMARIETTA ENERGY SYSTEMS Lig CARETE A RN CRRIONLS compATY IINHWIINI!HINI!H((H((NHlll(ffliIHMHINIHNI A Division of Union Carbide and Carbon Corporatio Post Office Box P Oak Ridge, Tennessee 3 4456 0349579 - \OCD-Q?\\J'I-F"L»I\)I—‘ 10. 12. 13. k. 15, 16. 17. 18. 19. 20. 21, 22, 23-26. 28. 29. 30-31. 32-36. 37. ORNL 1615 Reactors-Research and Power INTERQAL DISTRIBUTION C. E. Center C. E. Larson L. B. Emlet (§-12) W. B. Humes (M25) A. M. Weinberg A. H. Snell R. C. Btriant Dixon Callihan W. K. Ergen A. P. Fraas R. A. Charpie T. A. Welton i D. K. Holmes: 5 R. R. Coveyou A. J. Miller J. A. Swartoutj E. D. Shipley; J. A. Lane E. S. Bettis Dunlap Scot D. V. P. Williams E. L. ulr Mo ANP Reporjg 0ffice . Biology j@brary Health J¥ysics Library Reactojifixperimental Enginfering Library CentrglW Research Library Laboy¥ory Records Departme Labgitory Records Departmen$ ORNL R.C. 38. 100. - 101-10k. 105-107. 108. 109. 110. 111. 112. 113. 114-115. 116-117. 118. 119. . AF Plant Repre . ANP Project Of7% . Argonne Nations . Armed Forces Splcial . Atomic Energy Cgmmisglfon, Washington . Battelle Memorigl In@Ftitute . Bechtel Corporafond . Brookhaven Natidgad¥ Laboratory . California Reseall . Chicago Patent ¥ bup . Chief: of Naval gRc@earch . Commonwealth HNis . Detroit Edisge Congany . duPont Compaj . duPont Compglhy, Wigkington 3. Foster Wheeffer Co :'ration . Idaho Og} ERNAL DISTRIBUFION /4 entative, Bj¥fbank entative, @Pattle ‘ntative,/‘ood -Ridge ce, For, Worth Laborgory ¥opons Project (Sandia) AF Plant Repre AF Plant Repre (] Bureau of Ships and Development Company Carbide and Carbjh Chemicals Company (Y-12 Plant) ] Company Department ofjfkche Wavy - OP-362 , AuBlusta tric Copany (ANPP) (1 copy to J. A. Hunter) General Ejctric Cofgpany, Richland Hanford (ferations ¥ fice ations Offce Iowa Stgke College Knolls tomic Power @aboratory Los * ’mos Scientifi@ Laboratory (1 copy to H. c. Paxton) Massafhusetts Instit@e of Technology (Keufmann) Monsgiito Chemical Co:»any Mounll Laboratory Nat@bnal Advisory Confiittee for Aeronautics, Cleveland Najifonal Advisory Comilittee for Aeronautics, Washington Nglfal Research Laborafpry v York Operations Ofgice Orth American Aviatiol, Inc. General Ilg uclear Development Asdfpciates, Inc. Patent:Branch, Washingt 120. 121. 122. 123, 124, 125. 126. 127. 128-129. 130-131. 132. 133. 134-139, 140-154, Powerplant La Pratt and Whit USAF Headquarters U. 5. Naval Radiologii@lDefense Laboratory University of Calif ediation Laboratory, Berkeley University of Cal Wiation Laboratory, Livermore ABSTRACT The critical experiment program on the Direct Cycle Aircraft Propulsion Reactor was planned Jolntly by General Electric Company and the ANP group of Oak Ridge National Laboratory. The experiments were performed by the Laboratory group at the Oak Ridge Critical Experiment Facility. A cylindrical assembly having a core composed of alternate layers of hydrogeneous material and of an open lattice of stainless steel and metallic uraniuvm epproximately 36" long and 51" in diameter was constructed. The assembly was reflected by a 6" thick annulus of beryllium around its circum- ference and 6" of AGOT graphite on the ends which covered both the core and the beryllium jacket. The reactor required a critical mass of 41 kilograms of uranium enriched to 93.3% U-235, and used Plexiglas as & mock-up moderating material. Two mock-ups of the Direct Cycle Reactor are reported here. In one of these modifications the moderator, the fuel disks and the stainless steel supporting layers for the fuel disks were horizontal. The second mock-up con- sisted of a series of concentric rectangular shells formed by the rotation of an appropriate quantity of the core materials from a horizontal to a vertical position. Core tjpe fuel and moderator removal control rods were calibrated by both the stable period and the "rod drop" methods. Power and neutron flux distribution studies were made on microscoplc and macroscoplc bases for both these assembly mock-ups. Aluminum catcher foils in contract with the U-235 metal were used in the power distribution measurements and the flux dlstrlbutlons were determined by indium foil activation. Measurements were made of the changes in reactivity resulting from the removal of a section of the beryllium reflector and the substitution of other materials for it. Losses in reactivity were also measured in terms of the separation of the two halves of the assembly at midplane and compared with calculated values, treating the assembly as a bare thermal reactor. Reason- able agreement was obtained between experimental and computed values. Evaluations were made of certain experimental poison rods and. of sub- stituting limited amounts of molybdenum for stainless steel and of graphite for Plexiglas. The end reflector was removed from approximately one eighth of the aggembly and a stack of Plexiglas erected and other alterations made to further mock-up the Direct Cycle Reactor. The effects of these changes on the reactivity and neutron flux were determined. | TABLE OF CONTENTS Page AE’IRACT » . e . * ¢ . e [ L] . > * | - * . & . - ’ o - . * . e 5 IlIST OF FIGURES [ * e o ¢ & * o * - o ) ® * e . [ ) * . * - L] * ' 7 IIST OF TABIES . . ® . * . * * - - o o . * - * ® * a . * * ® . lo T. INTRODUCTION o « o « o o o o o « o o o o o o o o o o s wao. 11 II. IESCRIPTION OF EQUIPMENT AND ASSEMBLIES . . . . + . . . . .. 11 A. Permanent Equipment .. . . ¢« & ¢ ¢ 4 ¢ o o o o o o o » o o 11 B. Shelf: Type Mock-Up:of Direct Cycle Reactor . . .« . « « . . 1h C. Conversion from Shelf Type Assembly to Rectangular SHOLL TFPO = o v o o o o o o o o o o o o o o o e v o o4 18 D. Conversion from.Rectangular Shell to Water-Cone Reflocteod MOCK-UD +« « « o o o o« ¢ o o o« o o o o « o » o 18 III. EXPERIMENTAL STUDIES IN SHELF TYPE ASSEMBLY . . . « « . . .« & 22 A, Control Rod Calibrations « « « o o o« o o o« o o o o o o « o 22 B. Temperature Effect8 . .« « « ¢ o ¢ ¢ o o 0 o ¢ o o ¢ o o o« 27 C. Reflector Studles e e s o s o 4 e o s o s o s e e e e+ 29 D. Power and Neutron Flux Determinations . . « « « ¢« « » » « 36 IV. EXPERIMENTAL STUDIES Ofi RECTANGULAR SHELL TYPE ASSEMBLY . . . 56 A, Control Rod Calibratlions . « o o o o o « o ¢ o s o ¢« o« « « 56 B, Temperature Effects . . « ¢ ¢ ¢ ¢ ¢ ¢ ¢ 0 ¢ o o o @ C. Plexiglas Thicknees Reactivity Coefficlent . . . . «. « . & 59 D. Effect of Separation of Core at Midplane . . . . . . E. Rofloctor StUdle8 o « + o o o o o o o o o o o o o o o o o 63 F. Danger Coefficient Typs EvAluatlons . « « o+ « o o o « o+ o 66 G. Power and Neutron Flux Determination . . « ¢« « ¢ ¢« ¢ « & T2 V. EXPERIMENTAL STUDIES ON WATER-CONE REFLECTED ASSEMELY . . . . 86 A. Travefsés ianater-Cone Reflected Assembly . « « « . . . . 86 VI.comciusxoms......_,,.,............-.....9o VII. APPENDICES + & « « + v v o o o o oo v v o n oo v ueoon Ol A, Conversion from Shelf to Rectangular Shell Mock-Up . . . . 9l B. Analyses of Reactor Materials . . . . « ¢ ¢« ¢ ¢« o o o« « « = 93 C. Summary of Materials in Reactor Assemblies . . . . . . . . 9 mnliiy 6. 11. 12, 13. 1k, 15. 16. 17. 18. 19. 20. 21, 22, 23. - LIST OF FIGURES Photograph of Two Halves of Aluminum Matrix . o « + « . . Photograph of Fuel Element o o o o o o « o . . Loading Chart of Shelf Type Assembly . . . . Photograph of Interface of Shelf Type Assembly . . . . . . Loading Diagram for Rectangular Shell Type Assembly . . . Photografh of Midplane of Rectangular Shell Type Assembly Photograph of Small Area of Midplane of Recténgular Shell Type Assembly . * . . . . . o o . Q * * * . CalibratimnAof'Control.Rod A v v v e 6 o e s s e 0 e e Sengitivity .of Control RodiA e e e 4 e 4 e s Reactivity Change ve Temperature for Shelf Type Assembly Refleétor Value vs Void Width for Composite Reflector . . Sensitivity of Reflector Value vs Vold Width for Composite Reflector * . e L] * * . * * . » . . . . * . Reflector Value vs Composition for Composite Reflector of Stainless Steel and Plexiglas located at Bottom of Reactor Reflector Value vs Composition Reflector of Stainless Steel and Plexiglass located at Side of Reactor . . . « o « « « & Photograph of Materials used in Power and Neutron Flux Determinations . ¢« ¢ ¢ ¢ ¢ ¢ o o ¢ & o 4 e Radial Power Distribution in Shelf Type Assembly . . » . . . Axial Power Distribution in Shelf ije Agsembly . . . . Power Distribution 10 Cell Mel2 & v v v v o o o o o o o o & Radial Neutron Flux Distributlon In Shelf Type Assembly . . Axial Neutron Flux Distribution in Shelf Type Assembly . . Vertical Radial Neutfon Flux Distribution in Shelf Type Assembly Outline Drawing of a Unit Cell . . . o ¢« « + & Vertical Neutron Flux Distribution in Unit Cell Through Center Of Fue l Di Sk * -* - o e ° . o . * e * - o . o _7- Page 13 15 16 17 19 20 21 24 25 28 32 33 35 37 39 L0 41 L3 L 45 L8 L9 2L, 25. 26, 2To 28. 29. 30. 31, 2. 33. 3L, 35. 36. 37. 38. 39. h’Oo L1, el Vertical Neutron Flux Distribution in Unit Cell off Center . Of Fuel Disk * . @ ® » 0 o ® e . o ® * 0 ® a * o . L] *® ® - * * Vertical Neutron Flux Distribution in Unit Cell near Edge of Fuel DiSk . * . o e * e o e - o ° ® . . » » ® . ’ * o - * Vertical Neutron Flux Distribution Through Axis of Unit Cell Between Fuel Disks . & . & 4 v ¢ ¢ 2 « o o o o o o o o o » Vertical Neutron Flux Distribution off Axis of Unit Cell Betwoen Fuel Diske o o v o ¢ o ¢ o o o o o o o o o o o o Vertical Neutron Flux Distribution Near Edge of Unit Cell Between Fuel DiBKB o o o o « o o 6 o o a o « o o o o o o Neutron Flux Distribution Through a Unit Cell M-12 Having 2-Mi1l Disks Parallel to. PlexXiglas . « o o v o o o o o o o o Neutron Fiux Distribution Through a Unit Cell M-12. Having 2-Mil Disks Perpendicular to Plexiglas . . . . . . . . . . Bare and Cadmium Covered Indium Traverses Across a 10-Mil Fuel DiSk 2 L] o ? L] e N * o ° . * L] ® * . e L] L] o * . 2 - * Change in Reactivity ve Plexiglas Thickness . . + « « « + . . Change in Reactivity vs Separation Distance at Midplane . . . Change 1in Reactivity ve Plexiglas Thickness for Stainless Steel-~-Plexiglas-Boral Composite Reflector . . . . . e e Change in Reactivity vs Plexiglas Thickness for Stainless Steel- Plexiglas-Cadmium Composite Reflector . « o s o o o o ¢ o @ Bare and Cadmium Covered Indium Traverses through Composite Reflector with Boral Strips Present . o ¢ ¢« ¢ ¢ ¢« o ¢ o « & Cadmium Fraction vé Distance from Bottom of Test Section for Composite Reflector with Boral Strips Present . . . . . . . Bare and Cadmium Covered Indium Traverses Through Composite Reflector with the Boral Strips Removed . « « ¢« ¢ ¢ o « o & Cadmium Fraction vg Distance from Bottom of Test Section for Composite Reflector with Boral Strips Removed « « &« ¢« + . Power Distribution Through Cells M-12, M-8, M.6 and M-4 in Rectangular Shell Type Agsembly . o o « o+ o ¢ o o o o o o & Radial Power Distribution at Midplane in Rectangular Shell Type Assembly * o o o o * . . . o e o o ° * . L] . o . & * e * -8- Page 50 51 52 o3 51 22 o7 58 60 62 6l &5 67 69 70 73 7> 42, 43. L, L5, L6, 47, 49. 50. 5_10 52, el Shelf-Shield.ing Of Fuel DiSk o’ * o - * 4 L} * * . . * . e * . . * Vertical Radilal Neufiron flux Distribution between Plexiglas in Rectangular Shell Type Assembly v « o « « « ¢ o & « » o & Horizontal Radial Flux Distribution Through Plexiglas in Rectangular Shell Type Assembly . o o ¢ ¢ o ¢ o o ¢ o ¢ o & Axial Flux Distributibn‘Through Cell M-12 in Rectangilar Shell TypeAssembly....A..._..._......-........ Horizontal Neutron Flux Distribution at Plexiglas Intersection in Rectangular Shell Type Assembly . . . « « + v ¢ + o o & Vertical Neutron Flux Distribution Across Centrial Rectangular She ll Type Cell’ M-12 -+ . ” . * * . ® ° . o * » . * * * - * Horizontai Neutron Flux Distribution Around General Electric Company Experimental Polson Rod + + & & o ¢ « ¢ & o o o & & Vertical Neutron Flux Distribution Around General Electric Company Experimental Polson Rod . « &+ « « ¢ o & v o & & © & Axial Neutron Flux Distribution Through Cell R-12 In Water-Cone Reflected. MOCk-UP . ¢ . * . o .. * . * . . LI . * ‘u e o . . Axial Neutron Flux Distribution Through Cell U-12 1n Water-Cone Reflected MOCk-UP * * . * * * * o ) * [ ] o e 4 * * - . * . . . Axial Neutron Flux Distribution Through Cell H-12 in Unaltered Segment of Water-Cone Reflected MoCk-Up . « « o« & ¢ & & « & Page _77 78 79 81 82 83 8 8 87 88 89 1I, IIT. XIV. Xv, Page LIST OF TABLES Comparison of Rod Calibratlons « « « « « v ¢ o ¢ o o o« o o o 26 Reactivity Change Introduced by Substituting Beryllium for Air * & * 9 e * * . * * ® ., * * 4 * * * .. o * * * - ’ 29 Reactivity Change Effected by Substlitution of Stainless Steel and Plexiglas for Beryllium . . « o o « o ¢« o & o & 30 Reactivity Losses for Varlable Void Width in Reflector . . 31 Reactivity Losses Effected by Substitution of Stainless Steel and Plexiglas for Beryllium . ¢ . ® & * . * * L ] ¢ * [} . * ) 31 Reactivity' Losses Effected by Varying Composition of the Bottom Reflector ® . . ) * [} . .. e v o . L} s e * - o e . L ] o ' 3'+ Reactivity Losses Lffected by Varying Compositlon of the Side Reflector . * * . e . .. & e & e * . & . * e L] * & . * . 36 Thermal Neutron Distribution in a Unit Cell . . « « o . . + . 56 Calibration Ihta"for Control Rod AT 4n Rectangular Shell Tneflssemblyoooo.olqo‘ooo--oonooooo.ooo 59 Summary of Control Rod Cai_l.ibz_'a'tion DRt o ¢ ¢ o ¢ o ¢ 0 0 o 59 Effect of Core G’ap on ReactiVity ¢ ¢ o & o 0 e ¢ e . o e 61 Polson Rod Reactivity Changes « + « « o ¢ ¢ s ¢ ¢ o ¢ ¢ « o T1 Compar‘ison of Molybdenum and Stainless Steel . ¢« « ¢ ¢« ¢+ o & 11 Relative AXial POWBI‘ s o @ .",.)...l... s 9o & @& o % o & & » & @ 72 Cadmium Fractions of Fuel Disks . « « o « + o s o o ¢ o o o 76 . «l0a I. INTRODUCTION This report is a summary of the Information contained in five progress reportsl, of limited circulation, published at intervals during the course of these experiments. Since these progress reports were of limited cir- culation and avallable only to those directly connected with the project, it seemed advisable to make the information obtained from the critical experiments available on a wider basis in the form of an overall report. A description of the program of the critical experiments to be performed by the Oak Ridge National Laboratory contributing to the design of the Direct Cycle Nuclear Reactor by the General Electric Company is given in their Doc- ument DC-51-11-6, dated November 3, 1951, entitled "9213 Critical Experiment Program", and supplements, dated December 14, 1951; June 12, 1952 and others subsequently lssued as work on the program progressed. The first mock-up of the reactor as actually constructed consisted of a right circular cylinder, with axis horizontal, and a core 36" long and 51" nominal diameter containing 41.2 kilograms of uranium or 38.5 kilograms of U-235, The H to U-235 atomic ratio was 225. The core composition by volume fractions was as follows: Plexiglas (to simulate water) 0.319; stainless steel 0.043; uranium 0.00185; aluminum 0.063 and voids 0.573. It was assembled in a matrix of square aluminum tubing which will be described later. This mock-up, in which the Plexiglas was in horizontal layers separating groups of six layers of an open structure of type 302 stainless steel sheets holding the uranium metal disks, was designated asg the "shelf type assembly”. This mock-up is also referred to as CA-5. | The second mock-up consisted of a series of concentric rectangular shells formed by rotating an appropriate number of Plexiglas strips and their accompany- Ing structure of stainless steel sheets and fuel disks from a horizontal to a vertical position. This rotation was supplemented by minor changes in some core and reflector materials but with no change in the amount of uranium present. This modification, which will also be described in detail later, was known as the "rectangular shell type assembly" and was used to obtain additional data pertin. ent to concentric annuli in the design of the Direct Cycle Reactor. Thls mock- up is referred to as CA-6. Additional minor changes were made to the rectangular shell type assembly during the course of the experiments and this third mock-up 1s designated as CA-T. " II. DESCRIPTION OF EQUIPMENT AND ASSEMBLIES A. Permanent Equipment: The critical experiment assembly has been adequately described in other reportse, and only a brief description of the components which are essential to understanding of current experimental. techniques will be included here. Such features as shielding, interlock system, safety systems and scrams. are omitted since they have very little bearing on these experimental results. 1 Callihan, D. '"Preliminary Direct Cycle Reactor Assembly Part I, Y-B23-1, February 26, 1952; Part II, Y-B23-2, May 21, 1952; Part III, Y-B23-5, June 18, 1952; Part IV, Y-B23-7, June 30, 1952, and Part V, ORNL 52-12-225, December 15, 1952. 2 Bly, F.T., et al, NEPA Critical Experiment Facility, NEPA-1769, April 15, 1951. -Xl- Assembly Structure ' The assembly apparatus consists basically of a matrix of square 2S aluminum tubes stacked horizontally which, when assembled, form a &' cube and into which the reactor materials may be placed. The &' cube 1s divided into. two identical halves, except that one i1s stationary and the other can be moved by remote control a dis- tance of 5' from the stationary half. Xach half consists of 576 tubes, stacked in a 2L x 24 cell array. These tubes are 36" in length, 3" x 3" outside, and 0.O4T" wall thickness. Part of the reactor materials are placed in each half and the assembly 1s made critical by control rod adjustment after the two halves are to- gether. Figure'l 1s a photogiaph of the two halves of the assembly. (The materials visible in the upper portion of the moveable half have no relation to the Direct Cycle Reactor.) The moveable half is driven by a power screw coupled by gears to a 1 HP. D.C. shunt wound motor. The speed of the motor and the reducing gears 1s so ad justed that the moveable half approaches the fixed half wilth a speed of 33" per minute between 5' and 8" separation, 6" per minute between 8" and 2", and 1" per minute between 2" and the stopping position of the moveable half. Positions of the moveable half relative to the fixed half are indicated to 0.01" over the entire 5' travel by a Selysn Veeder-Root mechanism coupled to the power screw. Separations less than one half Iinch are Indlcated to 0.001" by Starrett Micrometer dial indicators coupled through Autosyn 400 cycle Selsyns to points at the bottom of the matrix. The precision of the incremental separatlions of the two halves was the order of¥ 0.001". Due to a lack of parallelism of these facing halves, however, the lnaccuracy of the overall separation was estimated to be about 0.02", Instrumentation The radiation level of the reactor 1s monitored by eight in- dependent instruments: a scintillation counter, a fission chamber, and six BF3 filled neutron detectors, two of which are proportional counters and four are ionization chambers. Leakage radiation is measured by placing, the detectors at varying dis- tances from the reactor core. @f special interest in the present experimental procedures are the logarithmic meter and the plle period meter, both of which derive their incoming signals from a BF, ionization chamber, and are used in the calibration of the cofitrol rods. The logarithmic meter records power levels over a range of approximately 6-1/2 decades, and the period meter records instantaneous pile periods from infinity to plus 3 seconds and from infinity to minus 30 seconds. Stable pile periods may be observed directly from the slopes of the curves traced by the logarithmic recorder. -12- . 13 Fig. 1. Photograph of Two Halves of Aluminum Matrix. Bfl The Shelf Type Mock-Up of the Direct Cycle Reactor: 1. Core Construction The enriched uranium metal 1s in the form of circular disks approximately 0.01" thick with diameters of 2.860" and 1.430" and masses of 18,0 F 0.1 and 4.5 * 0,1 grams respectively. Since it was not possible to roll uranium metal foll of sufficlent uniformity for this emall mass tolerance, meny of the disks had small holes punched in them for mass adjustment. Each disk had a 0.196" hole drilled in the center. The fuel ele- ments were assembled by placing the large disks on 3.6" centers above an open array of three sheeth of stainless steel and cover- ing them with an inverted three plece open array. Because of insufficient large disks, 1t was necessary, In some cases, to substitute small ones. Quarter size fuel elements were also assembled and used near the circumference to approximate a cir- cular core from square elements. Photographs of the fuel elements are shown In Fig. 2. The moderating material, Plexiglas, which has been shown to simulate water in thermal neutron systems3, was in horizontal layers, one inch thick. The separate layers support the open structure of stainless steel and fuel disks as shown in Fig. &4, Some detalls of the core construction are shown in the loading chart, Fig. 3. The numbers and letters in the cells show the positions of the safety and control rods respectively. Safety rods ' 1, 2, '3, and 4 and control rods A and B are in one half of the assembly and an equal number are in the other. The assembly was critical with some of the control rods rémoved. Correction for the accompanying voids, based on an empirical calibration of the rods -in units of mass of uranium indicates the "clean™ critical mags to be Ll.1 kilograms of uranium or 38.4 kilograms of U-235. The structure of the control and safety rods 1s, of course, the pame as that of the fuel elements.* Reflector Construction The core“of the reactor is enclosed in an end reflector of 6" ' of AGOT graphite and a nominal 6" thick side reflector of beryllium. The beryllium thickness varies slightly because the annular reflector is fabricated from square pleces. The graphite end reflector extends to the outside surface of the beryllium annulus. The control and safety rods contain the graphite reflector. A photograph of the assembled core at the midplane is shown in Fig. L. : P. Callihan, D., et al, "Critical Mass Studles, Part VI", Y-801, August 8, 1951 12, . A tabulation of the materials built into the assemblies. is given in - Appendix C pg. 95. -14. Gl Fig. 2. Photograph of Fuel Element. 16 owG. 21337 N A B CDEFGHI JKLMNOPQRSTUVWIX Be REFLECTOR, JACKET 18" LONG, EACH HALF BACKED BY 6" GRAPHITE (END). FUEL ELEMENT, EACH 18" LONG, EACH HALF BACKED BY 6" GRAPHITE. (TOP)—6 LAYERS STAINLESS WITH 5-10 mil x 3" U DISCS, HORIZONTAL; (BOTTOM)— I' PLEXIGLAS. TAINLESS STEEL, AIR, FUEL LEXIGLAS FIGURE 3. LOADING CHART 17 Y-12 Photo 11077 Fig. 4. Photograph of Shelf Type Assembly. Conversion from Shelf Type Assembly to the Rectangular Shell (RS) Type. The converslon from the shelf type criticel assembly, shown in Fig, 3 to the rectangular shell (RS) type shown in Fig. 5, consisted essen- tilally of the rotation of the planes of the fuel dlsks, the stainless steel sheets and the moderator strips from a horlzontal to a vertical position In approximately one half of the core cells. The effect on reactivity of the rotation of these core materlals has been calculated and is contained in Appendix A, Minor changes in the amounts of core materials.were made by substituting stainless steel sheets for Plexi- glas in the central cell, M-12, in which control rod A was located. It is recalled that the control:.and safety rods, all of which contaln fuel elements, extend from the end of the reactor to the midplane of the core. The only change in the reflector was the addition of a 2" layer of graphite reflector to one end, making it 8" thick, the other remaining 6". The lengths of control rods A and B were thereby in- creased 2". ' No changes were made in the locations of the control rods and the .safety rods nor in the amount of uranium in the core of the reactor. ‘Figures 6 and 7 are two photographs of the RS type assembly taken at the :midplane perpendicular to the axis of the reactor. Figure 6 shows the overall loading and Fig. 7 shows, somewhat more in detail, the arrangements of stainless steel and Plexiglas within the cells. The vacant cells shown in the photographs indicate the locatlons of either control or safety rods. The irregular boundary between those cells which have the planes of the fuel elements vertical and those in which they are horizontal is also visible. If the rotation of the fuel, stainless steel and moderator strips is in such a direction as to place the fuel adjacent to the beryllium reflector the assembly is more reactive than when the rotation is in the opposite direction since the latter puts a thick layer of hydro- geneous material between the core and the reflector. It was estimated that the RS type loading, as shown in Fig. 5, was less reactive than the former shelf type by about one percent in A k/Xeff. Converslon . from the Rectangular Shell. to the Water-Cone Reflected Mock-up. The water-cone reflected mock-up was constructed to better simulate water cooling channels of the reactor, ascertaining their effect on the flux distribution and reactivity. This information, céupled with the reflector studles, determined the severity of the penalty ' apsociated with the contemplated. substitution of the iron-water reflector for beryllium. Reflector changes made in the (RS) type mock-up of the Direct Cycle Reactor to simulate the water-cone reflector proposed for the end of the reactor were the following: .The six inch graphite end reflector was removed from one quadrant of dne half of the assembly consisting of the following cells: N-12; 0-11, 12; P-10 through P-1%, inclusive; Q-9 through Q-15, inclusive; R-8 through R-16, inclusive; S-7 through S-17, inclusive; T-8 through T-16, inclusive; U 10 through U-1k4, in- clusive. The Plexiglas in this group of 49 ¢ells was extended 10 /4", These alterations in the reflector made the reactor subcritical even ~18.- O D g O O H NN - 9 A BCDEFGHI JKLMNOPQRSTUVW X Be JACKET REFLECTOR, 36" LONG FUEL ELEMENT, 36" LONG STAINLESS STEEL, URANIUM, AIR SPACE, l%" THICK PLASTIC, I" THICK FIGURE 5. LOADING DIAGRAM Y-12 Photo 11452 Fig. 6. Photograph of Rectangular Shell Type Assembly. 20 Assembly . lar Shell Type Il Area at Midplane of Rectangula r i 'ograph of Sma Fig. 7. Photogra 12 with all the control rods in. Thie subcriticality was compensated for by the addition of 36" lengths of beryllium to the following cells: B-8 through B-16, inclusive; C-6, 7, 17, 18: D-5, 19; B-4, 20; F-3, 21; G-2, 22; H-1, 2, 22; I-1, 22; J-1, 23; K-1, 23; L-1, 23. The 6" graphite end reflector was removed from the portion of the beryllium Jacket which was adjacent to the quadrant contalning the 10 1/L4" Plexiglas extension and replaced by a 10 1/4" beryllium extension making the reflector the same length as the altered. core. The final change conslsted of erecting a stack of Plexiglas, approx- imately 12" thick, adjacent to one end of the reactor and starting at the table top and terminating at top of the following cells: Q-7, R-6, S-5, T-5, U-5, V-5 and W-6. The vertical edge of this Plexiglas stack, which simulated the water-cone, was 3/4" from the end of the aluminum matrix. The width of the simulated water-cone began at the first safety and control rod supporting column and extended to within about 1/2" of the edge of aluminum matrix. The change in reactivity produced by the removal of the graphlte and the extension of the Plexiglas in the 49 cells listed in the loading change was a decrease of 85.2 cents. The addition of the 36" extension to the beryllium reflector produced a gain in reactivity of 106.3 cents. Gains in reactivity were also observed due to the 10 1/4" extension of the beryllium reflector and to the erection of the simulated water-cone in the amounts of 17 cents and 8.4 cents, respectively. I1T. EXPERIMENTAL STUDIES ON SHELF TYPE ASSEMBLY A. Control Rod Calibrations: The displacement of a calibrated control rod serves as a measure of the changes in reactivity of a critical system. In the calibration the changes In reactivity introduced by the linear displacement of a control rod may be determined by either of two methods: 1) the measure- ment of the stable period resulting from the reactivity cliange, and 2) the "rod-drop" method'. 1. The Period of a Super Critical System A measurement is madé of the period of the super critical system resulting from the insertion of a control rod. When the changs in reactivity is small, it 1s related to the super critical period by the following expression: Q:E“____ - _L£24%§i_.= —__4éé___. (1) kKeff. )L /")4‘7_ { Where &k = change in multiplication or excess reactivity keff. -22- | T = stable plle periocd in seconds : _n = number of groups of delajed neutrons ')3_= decay .constant for neutron precursors of group 1 = fraction of fission neutrons in delayed group 1 or in units of cents: 7 / \ | 7. | =100 ST 77 (2) Korf E ]54 A\ TN Z,ga | ' / / . . where 100¢ is the reactivity equivalent to the effective fraction of delayed neutron. The symbols in equation (2) have the same significance as in eguation (1). Reactivity results are reported in cents values to avold the uncertainty in the effective value of the delayed neutron fraction nec- essary ‘in converting .to dk/kerr-" /2= 100 "Rod-drop" Method - Tn this msthod a measurement is made of the translent occuring when the safety rod is quickly removed from the assembly. The change in reactivity in cents (/) may be computed from the equation N. _ N - /o: 100 .._9_.__._1__ ' (3) Nl ’ where No = power level at critical | Nl_s extrapolated power level of the slow transient at the instant the rod is removed The change in reactivity produced by the linear displacement of each control rod was determined by these two methods. The supercritical period was measured for various increments of displacement until the rod had been displaced its full length and the corresponding changes in reactivity were computed from equation (2). Each of the safety rods,which was ¢colinear with'a control rod,was evaluated from the transient occuring in the flux as it was dropped. Evaluations of the safety rods were made from equation (3) and compared, respectively, with the integrated value of the symmetrically located control rod. Table I glves a summary of the total rod evaluations, in cents, for the four control rods and the values obtained by the rod drop method for the correspond- ing safety rods. Typical data obtalned in the incremental call- bration by the stable period method are shown in Figs. 8 and 9. The change in reactivity occuring when the rod is withdrawn from the reactor, plotted as a functlon of the rod position from the midplane, 1s shown in Fig. 8. In Fig. 9 1s plotted the change in reactivity per unit displacement, the rod gensitivity, as a function of the same rod position. The horizontal lines represent the aver- age sensltivity over the incremental displacement. The peaks . _23_‘ £ ROD VALVE Pl / s 7 // 7 P v P L’ /| // // /) 0 12 {6 20 249 ¢ ROD POSITION FROM MID PLANE 1IN INCHES FIGURE 8. CONTROL ROD A 6E€E1C "OMa be SENSITIVITY @/IN o o P o 10 1S ROD POSITION FROM MIDPLANE —IN INCHES FIGURE 9. CONTROL ROD A 20 25 — ovEld "OMd G¢ In Fig. 9 are attributed to the structure of the rod, but this result was not investigated. It should be noted however, that these peaks occur at the approximate positions where the channel at end of the reactor is opened first by the reflector removal and then by the removal of the complete element. TABIE T COMPARISON OF ROD CALIBRATIONS Control Rods Safety Rodse (by integrated periods) (by rod drop) A 18.2¢ 5 164 B 15.9 6 16 C 15.9 1 18 D 18.3 3 18 It is believed that the precision of the rod drop method 1is no greater than “$2 cents so that the agreement in some of the above cases 1s perhaps coincindental. In the course of the rod calibration experiments by the in- tegrated period method it was noted that the positions of the control-rods, when critical at an increased power level 1n a reactor which has been operated for several hours reproduced, within the limits of experimental error, the previous critical positions at a lower power, whereas the reverse was not true. That is, when the power was decreased by a factor of ten 1t was found necessary, in order to re-establish a critical system, to remove the control rod a few tenths of an inch beyond the po- sition where the system had been critical at the higher power level. ( This variation in rod position for criticality 1s attributed to an effective source of neutrons arising from the Be (¥,n) reaction which is relatively stronger at lower power levels following operation at higher power. In order to minimize errors arising from this source the following procedure was used in the calibration of, say, .control rod A. With the neutron flux constant at the lower level sufficlent excess reactivity was introduced by inserting some rod, not A, to produce a stable period of between 250 and 400 seconds. The power was allowed to increase, on this period, approximately ten fold and the assembly was made ceritical by withdrawing rod A. This displacement of A corresponded to the observed period. The power was then reduced and leveled by the appropriate manipulation of other rods and the procedure re- peated, thereby evaluating successive Increments of rod A. It is believed that, except for the first measurement, the rod displace- ment which produced the positive period is not exactly related to the period because, at the lower level, the assembly is not critical 6. due to the (¥, n) neutrons. In the Direct Cycle Reactor there was no beryllium in the core and this effect may not have been very large. An investigation was made of the linearity of response of the amplifier ion chamber circuit used In the perlod measurement by comparing it, at several lower levels, with proportional counters. A small deviation from linearity was found at power levels above those used in the control rod calibrations so the results are not significantly in error™. B. Temperature Effects At various times during the operation of reactor assemblies varl- atlons have been observed in the day-to-day settings of control rods required for criticality under constant loading conditions. In ‘the work reported here these varlations exceeded the sensitlvity required to detect the reactivity differences produced by some structural changes under study. Investigatlons have indicated the probable cause to be ambient temperature irregularities. The sta- bility of the room temperature has been improved so that the exlt alr temperature is constant to one half Fahrenheit degree at 80CF. An empirical relationshlp between reactor temperature and reactivity has been established. The temperature, measured at two points in the matrix with an iron-constantan thermocouple, has been varied from 63°F to 80OF by altering the room temperature and allowing sufficient time for the temperature throughout the reactor to equal- ize. The attending reactivity differences, from CELL M-12 URANIUM DISC DATA CELL M-I2 AL FOIL DATA CELL u-12 ° " " CELL M-4 " " " CELL R-7 3 6 9 12 15 18 21 24 INCHES FROM REACTOR MIDPLANE FIGURE |7. POWER DISTRIBUTIONS EXTENDING FROM REACTOR MIDPLANE TO GRAPHITE END REFLECTOR r disks. The results from the two methods are not distinguishable, validating the catcher foil technique, at least for this reactor. The best fit zero order Bessel functions of the flrst kind which represent these power distribution curves extrapolate to a core dlameter of 78" and a length of 49". Using these values of extrapolation dlstances, reasonably good agreement was ob- tained between the measured and calculated values of reactivity losses due to separation of the two halves of assembly at the midplane, which are tabulated in Section IV, Part D of this report. Power Distribution in Unit Cell Because of the inhomogeneity of both moderator and fuel iIn this assembly, the distrlbutlons of power and neutron flux inside the unit cell are of interest. Figure. 18:shows’ the vertical power distribution in Cell M-12 when 0.002" thick fuel dieks are in : Cells L-12, M-12 and N-12 and parallel to the planes of the Plexi- glas layers. This distribution was obtained by placing an aluminum catcher foll on each face of the thin fuel diske. The curve drawn through these relative activities assumes all the absorption be- twoen the fuel disks occurs in the stainless steel. The average activity over the cell is approximately 89% of the maximum value. The self shielding of the fuel disks and the fuel cadmium fractions are discussed in detall in Part IV, Section G of this report. Neutron Flux Distribution in Shelf-Type, Mock-Up Macroscoplc, or gross, neutron flux distributions were made by placing, successively, bare and cadmium covered indium folls at an equivalent position wilthin each of the cells In the reglon of interest. The position chosen was in the fuel plane between the uranium disks as shown in Fig. 15, The results of the hori- zontal traverses, taken radially at the mid-plane, and the axial traverses are shown in Figs. 19 and 20, respectively. In Fig. 19 the foll activity 1s plotted as & function of the dlistance from the axls of the reactor, parallel to the mid-plane, and in Fig. 20 the activity is plotted as a function of the distance from the mid- plane along the axis of the reactor in Cell M-12. The vertical radial traverse, Fig. 21, was also taken in the mid-plane and ex- tends from cell M 12 vertically dowvnward through the beryllium reflector. It is to be noted that in this lower region of the core the beryllium reflector is adjacent to a 1" layer of Plexiglas. No folls were placed in this layer of plastic during these traverses and the interpolation is based on exposures in a slngle cell which will be described later. A curve of a zero order Bessel function of the first kind 1s also included in this figure for comparison. Flux Distribution in A Unit ggl} Inhomogeneous loading of a reactor produces local variations in power and flux distributions which are of importance in re- actor calculations. Bare and cadmium covered traverses were made in & unit cell where both fuel thickness and -orientation could be varied. The unit cell used was always near the center of the re- “actor where the flux:;lis relatively.constant Over a 3".test.section. 42 RELATIVE ACTIVITY 1.0 o © o 0.90 0.85 0.80 PLEXIGLAS 1.O FIGURE 18. POWER Av.=893 % I7 mil ' PLEXIGLA STAINL STEEL LS ‘ 2.0 2.5 3.0 INCHES FROM BOTTOM OF UNIT CELL DISTRIBUTION IN UNIT CELL USING 2mil U-DISCS (STAINLESS STEEL AND U-DISC NOT DRAWN TO SCALE) 8yelZ "OMa ey ACTIVITY RELATIVE FIGURE I9. GROSS 9 12 BARE INDIUM ACTIVATION BE REFLECTOR CADMIUM COVERED INDIUM ACTIVATION BARE INDIUM ACTIVATION MINUS CADMIUM COVERED INDIUM ACTIVATION 15 18 21 24 27 30 INCHES FROM AXIS OF REACTOR INDIUM TRAVERSE HORIZONTAL IN PLANE OF |INTERFACE (M-12 THROUGH U-12) 33 6¥EILZ "OMA RELATIVE ACTIVITY BARE INDIUM ACTIVATION CADMIUM COVERED INDIUM ACTIVATION BARE INDIUM ACTIVATION MINUS CADMIUM COVERED INDIUM ACTIVATION 3 ' 6 9 12 15 INCHES FROM REACTOR MID FIGURE 20. GROSS INDIUM TRAVERSE ALONG DWG, 21350 GRAPHITE REFLECTOR I8 21 PLANE REACTOR AXIS (M-i2) 24 S ACTIVITY RELATIVE FIGURE 21 GROSS BARE x) CADMIUM COVERED BARE COVERED 9 12 15 INDIUM ACTIVATION CURVE INDIUM ACTIVATION INDIUM ACTIVATION MINUS CADMIUM INDIUM ACTIVATION 18 21 INCHES FROM AXIS OF -REACTOR INDIUM TRAVERSE VERTICALLY DOWN IN PLANE OF BE REFLECTOR - 24 2r 30 INTERFACE (M—I2 THROUGH M-22) 33 1612 "OMaQ In a normal unit cell the fuel disk is located 2" from the bottom of the test cell and 1" above the Plexiglas, producing a flux in the cell which is symmetrical above and below the fuel disks. The vertical flux traverses are, thersfore, symmestrical about the fuel plane. Figure 22/18 aidrawing of a:unit dellwilth: thetstainless steel removed to show the relative positions of the fuel disks and the outline of the Plexiglas in the bottom of the cell. The lines A-A, B-B, etc. show the locatlons of various traverses made through the cell. Figures 23, 24and 25! ghow thres vértical flux traverses, perpendicular to the fuel plane, taken through a unlt cell along the paths A-A in Cell M-12, B.B in Cell N-12 and C-C in L-12 respectively. The reductions in neutron flux which occur at the fuel disks are the results of shadow shielding by the fuel and the neutron sinks produced by lumped fuel. The reduction 1s less pronounced in the A-A traverse, Fig. 23, because the indium foil, in this traverse, was located over the center hole of the fuel disk and the shadow shleldling was greatly reduced. Three vertical neutron flux traverses were also made in a unit cell between the fuel disks and the results are shown in Fig. 26, 27 and 28. The increase in flux occurring at the fuel plane along traverge D-D, Fig. 26, was verified experimentally, although its cause is not fully known. Filgures 27 and 28 . indicate that the flux along E-E and F-F, between the Plexliglas layers, is essentially constant. - Determinations of neutron flux distributions in a unit cell vere also made with one 10-mil fuel disk replaced by five 2-mil thick disks. Only fifteen 2-mil disks were available so a large test volume was lmpossible. The thin disks were substituted in cells L.12, M-12 and N-12 near the reactor midplane with their planes parallel to the planes of the Plexiglas. The thin disks vere equally spaced between the Plexiglas layer and the top of the cell. The meutron flux distribution in cell M-12, along path A-A Fig., 22, 1s shown in Fig. 29. Thie- change caused the average neutron induced activity in cell M-12 to decrease by about 8 percent. In these traverses the indium foils were placed 0.25" from the center of the fuel disk to avold the effect of the hole. The ratioc of the average thermal neutron activity in the Plexiglas to that in the space between two adjacent layers of the plastic, with the Z2-mil disks loaded is 1.82 which 1s not significantly different from the ratio wheén a 10-mil disk is used., The results of an experiment - measuring the ratio of thermal neutron flux inside and outside the Plexiglas will be glven later. Another set of experiments was done with five 2-mil fuel dilsks each in Cells L-12, M-12 and N-12 placed perpendicular to the Plexi- glas layers and parallel to the axis of the core. The thin uranium disks were equally spaced between the vertical sides of the unit cells and extended into narrow slots cut into the Plexiglas. Vertical trav- erses, centrally located in the fuel disk array, were made in cell M-12 with the indium foils perpendicular to the Plexlglas and repeated wilth the foils parallel to it. The activity of the folls which were ‘ 7. 48 * \ * \ ‘. * “ \ II \ fa \ \ \ % \ * \ A \ . \ \ \ \ \ \ D \ 1 \ » \ 1 \ \ \ 1 \ \ A | \ . \ 1 Y \ . 1 . . \ i / N \ . .— \ . / 1 \ ! '\ \ / * ! \ \ * \ \ ] “ N\ — —— — — i, S e— — — p—— pp— — —— — w —_ ~ —— - - J \ . / / \ 11 \ 4 \ N ~ \ \\\ \\I.l.lll. a- \ oQ——— llllll.l'll\’fll. A IL/..II"}H'II'I - —_— 0 - g e o / N ’ A \ N 3 s\ 5 * / vl_d v 7 / lt . « . \ \ “ — 2! 1 . . o™~ A ) -a "\ - [ ¥ . '\ o i s \ \ . 3 w—_—— l[.ll.%lll._sl-.ll I..ll..__.lu..ll.lla'llll. e o \ . \ a -q If * at " — . “ of ao \ . _p 1’ -— / —/ \ \ / 1 * \ 4 \ \ \ i / . \ \ ! Y \ \ z- 7 Y // A \ ‘ \ v \ 3 ! \ N\ - '\ / . 1 \ L 7 * ' tl / / A a- S 4\\\ \\llll/ \ % I-lllll.l\\\ \ \\ . -— K A /7 N\ . . ) 7 x \ . Y ! \ " Y - ! \ . Xs / " s, 1y _- / * . 1 L) —§ f— ! pa / . i \ ! \ AN o — —— III|.+]|I',| —_— e — - —_— 1 \ | v . \ \ 1 . \ \ \ L X \ \ \ A \ * DY \ v A QO — e ey c— — < — — —— —_— \ a2 » a FIGURE 22. TYPICAL CELL { WITHOUT STAINLESS STEEL) RELATIVE ACTIVITY 0.9 o8 Bare Indium Activstion 0.7 0. o.s ' Indium Activation 0.4 Mimus Cadmiwm Covered ' Tndium Activation wm Covered Indium Activation €GEIZ "OMa 0.2 0.1 Plexiglas Stainless Steel tainless Steel and Air and Air 0 I 5 Distance irom Bottom of Cell i-12 in Inches ‘FIGURE 23, INDIUM TRAVERSE IN UNIT CELL LOCATION: A-A IN M-I12 (SEE FIGURE 22.) &y 0.9 0.8 0.7 0.6 o4 RELATIVE ACTIVITY 0.3 0.2 0.l o 0o FIGURE 24. Bare Tndium Activation C um Cov Indium Activation e Indium Activation inus Cadmium Covered Indium Activation Plexiglas Stainless Steel tainless Steel lexiglas and Air and Air INDIUM TRAVERSE 2 3.0 | Distance from Bottom of Cell N-12 in Inches IN UNIT CELL LOCATION: B-B IN N-12 (SEE FIGURE 22) ¥SELE "OMd oS RELATIVE ACTIVITY . 0.9 0.8 Bare Indium Activation 0.7 0.6 0.5 O \ . . : 3 Bare Indium Activation Minus Cadmium Covered Indium Activation ) Cadmium Covered 0.3 Indium Activation o = © N 0|2 ; o 0.1 Stainless Steel Stainless Steel Plexiglas and Air and Air 0 o ! 2 3.0 1 Distance from Bottom of Cell 1-12 in Inches FIGURE 25. INDIUM TRAVERSE IN UNIT CELL LOCATION: C-C IN L—I2 (SEE FIGURE 22) n — RELATIVE ACTIVITY 009 0.8 0.7 0.6 05 0.4 0.3 0.2 0.l 0 o FIGURE 26. Tridiom Activation Indium Activation Minus Cadmium Covered Indium Activation Cadmium Covered Indium Activation Stainless Steel Alr 2 : 3.0 I Distance from Bottom of Cell M~12? in Inches INDIUM TRAVERSE IN UNIT CELL LOCATION: D—D IN M-I2 (SEE FIGURE 22.) 9SE1Z 'OMa A% RELATIVE ACTIVITY Bare Indium Activation Bare Indium Activation . Minus Cadmium Covered \, Indium Activation ~#3 admium Covered Indium Activa Stainless Steel Steinless Steel and Alr and Air { 2 3.0 | Distance from Bottom of Cell N-12 in Inches FIGURE 27. INDIUM TRAVERSE IN UNIT CELL LOCATION: E-E IN N-I2 (SEE FIGURE 22.) LSELT "OMa £S RELATIVE ACTIVITY FIGURE 28, Stainless Steel Stainless Steel and Air and Air Distance from Bottom of Cell I=-12 in Inches INDIUM TRAVERSE IN UNIT CELL LOCATION: F-F INL-I12 Indium Activation Indium Actl Minus Cadmium Covered Indium Activation Cadmium Covered ndium Activation (SEE FIGURE 22.) 8GE1Z "OMa RELATIVE ACTIVITY 0.8 0.7 0.6 0.5 0.4 um Covered Indium Activation 0.3 o ' = ® Indium Activati ;o - Minus Cadmium ;; 0.2 Indium Activation G °o|‘ Plexiglas 58 Steel, Fueland Air lexiglas 0 0 ! 2 3.0 i Distance from Bottom of Cell M=12 in Inches FIGURE 29. INDIUM TRAVERSES IN UNIT CELL M-12 WHICH HAS 5-2mil U-DISCS PARALLEL & TO PLEXIGLAS 1v, oriented perpendicular to the Plexiglas was about 6% less than that in those which were parallel. The two sets of data wvere normalized at a point 2" from the bottom of the cell and are shown in Fig. 30. The minimum in the flux distributlon oceurs at a point 1.5" from the bottom of the cell which 1s the location of the center of mass of the fuel disks. A measure of the flux distribution across the dlameter of a 10-mil fuel disk was obtained from bare and cadmium covered indium traverses across a disk located in cell M-12 and the results are shown In Fig.3l. No foils, however, were located adjacent to the edge of the disk. The ratios of the average thermal flux, i.e., below the cadmium cut-off, in the Plexlglas to that In the space between the Plexiglas layers for this group of unlt cell traverses are given . in Table VITIT, i, oG drois sl TABLE VIIT THERMAL NEUTRON DISTRIBUTION IN A UNIT CELL Av., Flux in Plexiglas Location of Traverse Average Flux Average Flux Between Av. Flux between in the Unit Cell in Plexiglas Plexiglas Layers Plexiglas layers A-A 4080 2275 1.79 B-B 3610 2345 1.54 C-C 3690 2300 1.60 D-D 3980 2650 1.50 E-E 3900 0655 1.47 F.F 3700 2750 1.35 EXFERIMENTAL STUDIES. DN RECTANGULAR/"SHELL! TYPE ASSEMBLY A. Control:Rod Calibrations In: The Rectangular Shell Type Assembly: The control rods in the rectangular shell type mock-up were call- brated by the super-critical period method as described previously. The rod drop method was not used for compariscon. The reactivity values obtailned for three of these rods are comparable to those obtalned when the planes of all the core materials were horizontal. The fourth, Rod A, is significantly more effective in the rectangular shell type mock- up due to the change on the neutron distribution resulting from the omigssion of the Plexiglas from this rod and also surrounding the rod on all four sides by one inch layers of the plastic. It is interesting to note that the rod calibration was not changed by a gap, in the axial direction at the midplane of the core, of thickneas up to 0.8": .. - Detailed data for rod A are given in Table IX and a summary for all four rods in Table X. -56- RELATIVE ACTIVITY 0.7 0.6 0.5 0.4 0.2 0.l o o - FIGURE 30. INDIUM LOCATION : A—A Stainless Steel, Fuel and Air um Bare Indium Activation Bare Indium Activation Cadmium Covered Indium Cadmium Covered Indium Indium Folls Parallel to Plexiglas @ Bare Indium Activation Bare Indium Activation Minus Cadmium Covered Indium Activation Distzance from Bottom of Cel]? M=12 in Inches TRAVERSE IN UNIT CELL PLACED PERPENDICULAR TO THE: PLEXIGLAS AND REPLACE |—I10 mil U-DISC IN M—-12 (SEE FIGURE 2. IN ) WHICH 5-2 mil U-DISCS ARE 09€LT "OMa S RELATIVE ACTIVITY 0.8 0.7 0.6 0.5 04 0.3 0.2 0.1 O 0o FIGURE 3l. | 2 3 Distance from Midplane in Inches INDIUM TRAVERSE ACROSS A 0.010" FUEL DISC Indium Acetivation um Covered Indium Activation Bare Indium Activation minus Cadmium Covered Activation, 4 LOCATION: M~— |2 19€1Z "OMQ 8¢ TABLE IX CONTROL ROD "A" CALIBRATION Position of Rod "A" from Core Mid-plane Integrated Average Sensitivity in Inches Value in Cents In Cents per Inch 0,00 _ 0.00 -- 3.23 3.42 | 1.06 4.93 6.80 1.98 6,43 9.81 2.01 8.35 13.08 1.79 10.30 16.95 1.99 13.73 ' 21,22 1.25 17.01 . 24,58 . 1.02 21.39 Lo 27.39 ‘ 0.64 26,09 | 28.79 , 0.30 TABIE X SUMMARY OF CONTROL ROD CALIBRATION DATA - Reactivity Value Control Rod - | Lengthe in Inches ~ in Cents A 26, 28.8 B 26, 17.1 C 254, ’ 15.6 D 254, 17.0 Temperatire Effects in Rectangular Shell Type Assembly: The effect of temperature on reactivity in the rectangular shell type assembly was measured by the same procedurs and over approxi- mately the same temperature range as was used in the shelf type assembly. The ihcreéase .in:.reactivity with increasing température ébeerved was 1.43 cents:per Fahrehheltndegree, compared.with the. sarlier value of 1.25 cents per Fahrenhelt degree. Plexiglas Thickness Reactivity Coefficient: The thicknese of the Plexiglas moderator strips in five cells, K-16 through 0-16, was varied and the changes in reactlvity produced were determined from the control rod positions. In Fig. 32 the changes in reactivity in cents referred to the reactivity with one indh‘strips,;are;plbttéd.as.affunCtion:dftthe-Plexiglasfithibkness. This curve shows an optimum thickness of approximately 0.9 inches for this region of the reactor at the operating temperature. Effect of Separation of Core at Midplane on Reactivity: It wag desired to determine the loss in reactivity due to the separation of the two halves of the reactor core, the limit of the experiment to be a gap of one inch or a reactivity decrease of 100 cents, whichever occurred first. =59~ IN REACTIVITY IN CENTS CHANGE -0 owc. 21362 (N ] L | L | | 0.5 FIGURE 32. CHANGE 0.6 0.7 0.8 09 1.0 1.1 PLEXIGLAS THICKNESS IN INCHES IN REACTIVITY VS. PLEXIGLAS THICKNESS IN TEST CELLS 09 The procedure used was to start with the two halves together and, as they were separated in an axial direction, compensate for the loss in reactivity by driving in control rods. The reactivity loss for each increment of separation was then determined from the control rod calibration curves. When all the control rods were in, extra beryllium reflector was added to strateglc areas and evaluated, allowing withdrawal of the control rods, and the process was repeated, with soms overlapping in distance, until the desired magnitude of separation had been achieved. Table XI shows the reactivity losses for various separation distances and the average values of the reactivity losses per unilt distance over successlve Ilncrements of separation, TABLE XI EFFECT OF CORE GAP ON REACTIVITY Separation Dilstance Reactivity Loss Average Sensitivity at Mid-plane in Inches in Cents in Cents per Inch 0.03 1.92 61.9 0.06 L.37 Th.25 0,10 7T.39 81.6 0.1k 11.20 90.5 0.19 16.41 . 99.8 0.26 23.88 114.6 0.34 33.75 124 .2 0.43 45.83 133.2 0.46 50.00 138.1 0.26 25.30 99.4 0.43 48.17 131.4 0.51 2907 139.0 0.59 T1.43 148.2 0.63 77.13 _ 150.1 0.29 26,61 9l.1l 0.4k 16.05 . 134.1 0.55 62.50 . 150.9 0.60 71.23 . 143.9 0.6k T7.95 158.1 0.70 86.22 159.0 0.75 94 .23 . 160.2 0.79 102.2 & . - 162.3 0.81 105.7 164 .8 When separated, the facing sides of the core halves are scmewhat non-parallel introducing an inaccuracy in the separation distance estimated. to be about 0.02". In Fig. 33 the solid curve shows the values of the measured reactivity losses as a function of the distance between the two halves of the core when separated at the midplane:. The broken line in this figure represents the calculated . ‘~values of the reactivity losses using the method suggested by'Tamor7l 7 Tamor, S.,"The Effects of Gaps on Pile Reactivity; ORNL-1320, July 1k, 1952, pages 11, 13. ~61- REACTIVITY LOSS IN CENTS DWG. 21363 _ 120 100 80 60 40 20 EXPERIMENTAL CURVE 0.1 0.2 0.3 0.4 0.5 0.6 0.7 SEPARATION DISTANCE AT MIDPLANE IN INCHES FIGURE 33. LOSS OF REACTIVITY VS. SEPARATION OF TWO HALVES 0.8 0.9 1.0 R OF CORE AT MIDPLANE and treating the assembly as a bare homogeneous reactor. The reactor constants used were those recorded by Leverett®. Reflector Studies on The Rectangular Shell Type Assembly: 1. Stainless Steel, Plexiglas and Boral Composite Reflector The test cells used in these reflector studies were K-21 through 0-21 and K-22 through 0-22. The beryllium reflector- was removed from these cells, on both halves of the assembly, and replaced by stainless steel, except for a 5/16" layer adJacent to the core which was filled with Boral* strips 36" long. The system was then made critical by the addition of beryllium reflector at locations remote from the test section and by control rod adjustment, thereby measuring the loss in reactivity. Keeping the overall dimensiocns of the test section constant, the thickness of the stainless steel was reduced and Plexiglas was added between the core and the Boral in a step- wise menner until all the steel had been removed. The changes in reactivity, referred to the all beryllium reflector, which were incurred by these reflector alteratlons are shown in Fig. 3k4. An additional experimental point in this figure, designated as 2.9/16" Plexiglas and 2-7/8" stainless steel shows the effect of removing the Boral from a Plexiglas-stainless steel composite. Another point, measured with the Boral replaced by 5/16" of Plexi- glas shows the effect of a 5/16" vold between the Plexlglas and the steel to be small. The effect of the removal of all materials from the test section is also shown on the graph. ‘ Stéinless Steel, Plexlglas and Cadmium;Composite Reflector The experiment described above was repeated with a layer of cadmium substituted for the Boral. Since the cadmium sheets, which were 0.02" thick and sandwiched between 5-mil aluminum, were thinner than the Boral, the Plexiglas-steel composite was correspondingly thicker. Changes in reactivity were measured as the thickness of the Plexiglas, separating the core from the cadmium shéets, was increased from zero to 2-3/16". As in the preceding experiment, the total reflector thickness remained constent. .The results are shown graphically-in Fig. 35, where, for reference, the zero reactlvity change 1s again taken as that with the normal beryllium reflector. Neutron Flux Distribution through Stainless Steel, Plexiglas and Boral Composite Reflector . Bare and cadmium covered indium traverses were made through this composite reflector starting at a point G" inside the core Leverett, M. C. "The Direct Cycle Aircraft Reactor", Reactor Science and Technology, 3, 7 (1953). The Boral was prepared from a mixture of 35 weight percent B)C and aluminum. This mixture is held between two aluminum sheets; each about 0.04" thick, forming a 1/4" sandwich containing 250 mg of boron per square centimeter. The Boresl etripe were wrapped in masgking tape to mlnimize boron contamination. -63- REACTIVITY LOSS IN CENTS DWG. 21364 Y O I [ l \I [ BERYLLIUM REFLECTOR 50}— — 60— . . L REFLECTOR COMPOSED OF 2-9/16" PLEXIGLAS AND 70— 2-7/8" STAINLESS STEEL | 2-7/8" PLEXIGLAS AND 5 '/2-7/8" STAINLESS STEEL 90— 100— | {Op— 120— - | — 150 : —_— REACTIVITY LOSS FOR /_vom IN TEST SECTION | | I I I I 60 0 | 2 3 4 5 6 PLEXIGLAS THICKNESS IN INCHES FIGURE 34. REACTIVITY LOSSES VS. PLEXIGLAS THICKNESS FOR STAINLESS STEEL-PLEXIGLAS-BORAL COMPOSITE REFLECTOR REACTIVITY LOSS IN CENTS — DWG. 21365 SECURITY INFORMATION 65 T | A | BERYLLIUM REFLECTOR 715p— : — 2- 7/8" PLEXIGLAS AND ' 2-7/8" STAINLESS STEEL 80|— — 85— —_— 90— , ' — S ' ‘ | — | 00— —_ 05— —_ 50— - /vom IN TEST CELL 0o I 2 3 4 5 PLEXIGLAS THICKNESS IN INCHES FIGURE 35. REACTIVITY LOSS VS. PLEXIGLAS THICKNESS FOR STAINLESS STEEL -PLEXIGLAS - CADMIUM -COMPOSITE REFLECTOF and terminating at the bottom edge of the reflector. For these traverses the composite reflector consisted of an out- side layer of 2-7/8" of stainless steel separated from a 2-9/16" Plexiglas inner layer by the Boral sheet. Figure 36 shows the results of these traverses, in which the various activities are ‘plotted as a function of the distance from the bottom of the test section. Figure 37 shows the cadmium fraction computed from the relation An Cadmium Fraction = B - AC Ag (where A refers to the activity induced in an indium foil and the subscripts B and C indicate it to have.been bare or cadmium covered, respectively) and plotted as a function of the same distance. : - Figure 38 shows the results obtained along the traverses reported in Fig. 36, except that the Boral strips were omitted and the space left void. Cadmium fractions were again computed and are plotted in Fig, 39. No flux traverses notr cadmium fraction determinations were made with the stainless steel, Plexiglas and cadmium composite reflector. F. Danger Coefficlent Type Evaluations in Rectangular Shell Type Asgembly: 1. Poison Rods The changes in reactivity caused by prototypes of various control rods in design by General Electric were obtained from - the positions of calibrated control rods-when the assembly was critieal. The solid strip of Plexiglas in the bottom of a test cell was replaced by a plece through which a hole 5/8" in diameter had been bored. The axis of the hole wae parallel to the long axls of the Plexiglas strip and equally spaced between the top and bottom faces. A hole was also bored through the graphite end reflector to match the one in the Plexiglas. The poison rod wvas inserted into the hole with one end of the rod flush with the midplane. The reactor was then made critical. The rod was dis- placed 18", the end of the rod being at the end reflector-core interface, the solid plece of Plexiglas replaced and the system again made critical. Table XII gives a description of the rods for identification, 'the matrix cells in which the evaluation was made and the decreases in reactivity, in cents, Incurred when the rodsg replaced the plastic. " The three values recorded in Table XII and evaluated In cell I-12 were made after the conversion.of: one guadrant: of:erne. halfiof . the assembly:fromthe rectdngular. shell. .type, graphite reflected, to: the water-cone' reflec¢ted mock-upy .The.other four values were obtained in the rectangular shell type assembly before the conversion. ~66- RELATIVE ACTIVITY DWG. 21366 Iy " L 67 ' | | | | | 0.9— — BARE INDIUM ACTIVATION 0.8}— - 0.7 — 0.6— CADMIUM COVERED N ] INDIUM ACTIVATION 0.5}— — N 0.4}— — BORAL 0.3p— — BARE INDIUM ACTIVATION MINUS CADMIUM COVERED INDIUM ACTIVATION 0.2}— —_ STAINLESS STEEL PLEXIGLAS 0.1 _ CORE ———» 0 | | | 15 12 9 6 3 0 DISTANCE FROM BOTTOM OF TEST CELL M-22 IN INCHES INDIUM TRAVERSES FIGURE 36. BARE AND CADMIUM COVERED THROUGH STAINLESS STEEL- PLEXIGLAS - BORAL COMPOSITE REFLECTOR CADMIUM FRACTION 0.8 0.6 0.4 0.2 DWG. 21367 BN ss I I I I I I BORAL STAINLESS STEE PLEXIGLAS: CORE — I I I | S | 15 12 9 6 3 0 DISTANCE FROM BOTTOM OF TEST CELL M-22 IN INCHES FIGURE 37. CADMIUM FRACTION VS. DISTANCE FROM BOTTOM OF TEST CELL FOR STAINLESS STEEL -PLEXIGLAS - BORAL COMPOSITE REFLECTOR RELATIVE ACTIVITY 1.0 - : I l | o | | 0.9 BARE INDIUM ACTIVATION . 0.8{— \ — 0.7}— — 0.6}— \ ] \ CALMIUM COVERED I INDIUM ACTIVATION I | 0.5— I _ I | | 04— { — ‘ VOID 0.3}— S ———— \ ] BARE INDIUM ACTIVATION \ MINUS CADMIUM COVERED \ INDIUM ACTIVATION \ | STAINLESS __| 0.2 STEEL ' PLEXIGLAS g O.l}— \ E — - g CORE —= N \ N I | I I \ 15 12 9 6 3 0 DISTANCE FROM BOTTOM OF TEST CELL M-22 IN INCHES FIGURE 38. BARE AND CADMIUM COVERED INDIUM TRAVERSES THROUGH STAINLESS STEEL- PLEXIGLAS COMPOSITE REFLECTOR CADMIUM FRACTION 1.0 0.8 0.6 04 0.2 DWG. 21369 _ l l I | | [ VOID STAINLESS STEEL PLEXIGLAS-\- CORE — l I I ] R I 15 12 9 6 3 0 DISTANCE FROM BOTTOM OF TEST CELL M-22 IN INCHES FIGURE 395. CADMIUM FRACTION VS. DISTANCE FROM BOTTOM OF TEST CELL FOR STAINLESS STEEL-PLEXIGLAS COMPOSITE REFLECTOR 74 TABLE XII POISON ROD REACTIVITY CHANGES Reactivity Loss Rod Description Cell Number in Cents GE Rod #1* I-12 ‘18.5 Rod Labelled "Silver" I-12 12.9 Rod Containing 85.6% Ag I-12 18.2 and lh 4% Boron GE Rod #1 - M-8 17.5 " GE Rod #1* M-11 19.9 Rod Detailed in GE Print M-11 23.2 #B4098083-55" Cadmium Rod Containing .8.56 M-11" 13.2 grams Cd/cm., 34-1/2" long and 1/2" 0.D. 2. Comparison of Molybdenum and Stainless Steel The stainleas steel sheets in three adjacent cells in one half of the core were replaced by ones of molybdenum having the game dimensions. The resulting change in reactivity was de- terminéd, in terms of previously calibrated contrel rod settings, by bringing the reactor to criticality before and after the molyb- denum sheets were introduced. The experiment was repeated in three other positions in the reactor, using three adjacent cells each tims, Table XIII gives the group of cells used and the corres- ponding gains in reactivity resulting from the substitution of the molybdenun for stainless steel. TABLE XTIIT COMPARISON OF MOLYBDENUM AND STAINLESS STEEL Cells Used Gain in Reactivity in Cents L-12, M-11, N.12 | 11.9 L-9, M-9, N-9 Te9 L-7, M-7, -7 - 5.6 L-k, M-k, N-L 1.9 3. Comparison of Plexiglas and Graphite Removal of the Plexiglas from an array of nine matrix cells, K through M-16, 17 and 18, in one half of the assembly incurred a loss in reactivity of h? 8 cents. When graphite was substituted for Plexiglas in these same nine cells the loss amounted to only 26.7 cents. * Specifications for: this'rod are given:in Appendix B. - . ~71- L, Effect of Arco Soil Sample on Reactivity The loss in.reactivity incurred when a small sample of Arco, Idaho, s0o1l® was introduced into the reactor was determined by comparing the reactivity of the assembly when the soll sample was placed in a stainless steel box 1" x 2-7/8" x 2.7/8", having a wall thickness of 8 mils, located at the mid-plane of the reactor core in cell M-12, to that of a void in a box of equal volume at the same location. The results show that with 16l.L4 grams of soil in the box there was a loss of reactivity of 0.83 cents. - G. Power Digtribution and Neutron Flux Detérminations:in Rectangular : Shell Type ASsomblys Axial and radial power traverses were made through the rectangular shell type assembly employing the aluminum catcher foil technique which has already been described. Since no fuel disks were located at the center of the core, the axial traverses were extrapolated to the mid- plane for normalization. Relative power wae then plotted as a function of the distance from the core midplane. The one radial power dilstri- bution which 1s reported was taken with all catcher foils 1.8" from the midplane. Neutron flux distributions in the rectangular shell type assembly were determined using indium foils. In each case the bare indium activation the cadmium covered .indium activation and thelr differ- ence are shown on each graph, plotted as a function of the appropriate distance. 1. Axlal Power Traverses Axial power traverses were made through cells M-12, M-8, M-6 and M-L. The small variations in relative power through cells M-12, M-8 and M-6, do not Justify plotting separate curves for each of these cells. There.is, however, an appreclable decrease through cell M-4 and a separate curve is plotted. A summary of the results 1s given in Fig. 40 and Table XIV. TABIE XTIV REIATIVE AXTAL POWER Distance from Midplane In Inches Cell Number . 0.0 . LB .. 5.k .. 9.0 . 12.6 ... 16.2 M-12 1.00 0.99 0.99 0.87 0.84 0.71 M-12 1.00 0.99 0.98 0.88 0.81 0.69 M-8 1.00 0.99 0.97 0.88 0.82 0.73 M-6 1.00 0.99 0.68 0.87 0.82 0.72 M-k 1.00 0.99 ~ 0.94 0.83 - 0,75 0.62 M-4 ~1.00 0.99 0.93 0.88 0.7k 0.65 * An analysis of this soll sample is given in Appendix B. :-72- RELATIVE ACTIVITY DWG. 21370 _ I | T I 7 I I | I M—12 O @ A A 0.4}— o.2t— 0 I I I I I I | I I 2 4 6 8 10 12 14 16 18 DISTANCE FROM MIDPLANE IN INCHES FIGURE 40. AXIAL POWER TRAVERSES THROUGH CELLS M-{2, M-8, M-6 AND M-4 Radial Power Traverse One vertical radial power traverse was made 1.8" from and parallel to the midplane starting at the axls of the reactor and extending through the fuel plane of cell M=k, The results of this traverse are shown graphically in Fig. kl, The large decrease in relative power which occurs along the filrst few inches of thls traverse 1s due to cell M-12 being surrounded on all four sides by & one inch layer of Plexiglas which forms the innermost: v of the series of concentric shells making up the agsgembly. The data represented by the last two points on this curve were obtained from aluminum catcher foils placed in contact with the uranium fuel disk in cell M-4, one foil being adjacent to the top .and onecionithe bottoin of this fuel.dlsk. - The curve is. drawn through points repréesenting data obtained from the upper surfaces of the fuel disks. Factors Affecting Power Production by Fuel Disks a. Self Shielding If the fissionable material in a reactor is of high density the neutron flux producing fissions is depressed, a phenomenon known as fuel self shielding. The self shield- ing of a 0.01" fuel disk was measured before the converslon from the shelf type to the rectangular shell type mock-up. by forming a sandwich composed of five 2-mil fuel disks separated by aluminum catcher folls. The results are sum- marized in Fig. L2 where the experimental points are the averages of catcher foil counting rates which varied less than 1%. The distribution is assumed to be symmetrical about the central disk and the experimental points have been reflected about the plane of symmetry. The ratio of the aver- age activity to that at the surface of the 10-mil disk is 88%. b, Cadmium Fraction of Fuel Disk The cadmium fraction for a fuel disk 1s defined as the fraction of fissions which are produced by neutrons having energles below that of the cadmium cut-off. It is obtained by the relation: Cadmium Fraction = 1BC Fp where F refers to the fission fragment actlvity on the aluminum catcher foll and the subscripts B and C indicate the uranium disk as having been bare or covered with cadmium. Cadmium fractions for fuel disks were obtalned both before and after the conversion from the shelf to rectangular shell type mock-up from aluminum catcher foll measurements. All cadmium fractions for fuel disks were mesasured at a distance of 1.8" from the midplane of the assembly. The results are sum- marized in Table XV. . The RELATIVE ACTIVITY 0.2 DWG. 21371 [ I I I l | | I l | 1 l l I l 6 9 12 15 I8 21! 24 27 VERTICAL DISTANCE FROM AXIS OF REACTOR IN INCHES FIGURE 4I. VERTICAL RADIAL POWER TRAVERSE AT MIDPLANE 6L 6. TABLE XV éADMIUM.FRACTIONS OF FUEL DISKS Rectangular Shelf Type Assembly Shell Type Assembly Cell Cadmium Fraction Cell ‘Cadmium Fraction M-12 0.92 M-12 0.94 U-12 0.92 M-8 0.92 M-L4 0.90 M- 0.90 The plastic encloging Cell M-12 in the RS mock-up appears to increase the fraction of thermal fissions there. Absolute Power Determination Absolute power in the rectangular shell type assembly was determined by normalizing the power traverses to absolute fission rates by comparing & one-mil gold foll activation in the critlcal aggembly with that induced by the Oak Ridge National Iaboratory Standard Pile. From this normalization the absolute power during foll exposures was found to be approximately two watts. Radial Flux Distributions a. Traverses not in Plexiglas Bare and cadmium covered indium activatlions were measured along a vertical radius starting at the axis of the reactor in. Cell M-12 and extending downward through the beryllium reflector, terminating at the bottom of cell M.22. Data for these traverses were taken with the indium folls located as near the mid-plane as poseible and not in the Plexiglas. Results are shown graph- ically in Fig. u43. These curves indicate that the flux in cell M-12 1is high. This flux distribution agrees with the peak found in the relative radial power distribution shown in Fig. 41. The flux peak in the beryllium reflector also corresponds to that found for the shelf type assembly as shown in Fig, 21. b. Traverses Through Plexiglas Bare and cadmium covered indium traverses were made through the Plexiglas starting at the axls of the reactor in cell M-12 and extending through the beryllium reflector along a horizontal radius. The results are shown graphically in Fig. L4. The rela- tively high values of the flux in cell M-12, which 1s surrounded on four sides by one inch thick layers of Plexiglas, and the peak in the beryllium reflector, which have been noted previously, are again observed. Axial Flux Distribution Bare and cadmium covered indium traverses were made through cell M-12 along the axls of the reactor, starting at the midplane and =76-" DWG. 21372 7 1,00 O Experimental Points D Reflected Points 0095 0.88 Relative Activity o 0 o Q . (@] AVa ¥ 0.80 0 2 L 6 8 10 Uraniuwm Thickness, Mils FIGURE 42. POWER GENERATION THROUGH FUEL RELATIVE ACTIVITY DWG. 21373 -0 I 0.8}— 0.6p— X MINUS CADMIUM COVERED *\\/—— INDIUM ACTIVATION BARE INDIUM ACTIVATION O BARE [INDIUM ACTIVATION CADMIUM COVERED ™~ _ J*BE REFLECTOR | fis_ o INDIUM ACTIVATION 04— ~ 0.2}— NN 0 | | | | I I N >SS S JJ 4 8 12 16 20 24 28 32 DISTANCE FROM AXIS OF REACTOR IN INCHES FIGURE 43. VERTICAL BARE AND CADMIUM COVERED INDIUM TRAVERSE AT MIDPLANE FROM CELL M-12 THROUGH M-22 8L 0.8 RELATIVE ACTIVITY DWG. 21374 Oy | | M~ NI~ 1 ~ N> NN DY ~N ~ > N> - ~ NN BARE INDIUM ACTIVATION ~ ™~ Y <> h ~ N ~ - \ ~ ~N ™ _ ~ O e < D | REFLECTOR ™~ ~ > ~ ~ ~ ~ q \ \ \ N e~ BARE INDIUM ACTIVATION Y \\\ “w¢———MINUS CADMIUM COVERED ~ NN N\ INDIUM ACTIVATION ~ N ~ N ~N M AN - ~ ~ ~ — ~ ~N ~—_ SO~ CADMIUM COVERED T~ ~ ~ /_INDIUM ACTIVATION =~ ~ N~ n ~w ~ O ~ ~ \\ ~N N “~ porte— \ e \ ~ \ NG 0 NN | | | 1 | l IS 0 4 8 12 16 20 24 28 32 36 DISTANCE FROM AXIS OF REACTOR IN INCHES FIGURE 44. RADIAL BARE AND CADMIUM COVERED INDIUM TRAVERSES TAKEN HORIZONTALLY | AT MIDPLANE THROUGH PLEXIGLAS extending through the graphite end reflector. The results are shown graphically in Fig. u45. 7. Neutron Flux Dlstribution at Plexiglas Intersection The conversion from the shelf type to the rectangular shell type of mock-up produced intersectlons of the ver- tical and horizontal layers of the Plexiglas strips ex- tending radially from the four corners of cell M-12 at azimuthal angles of about 45° both above and below the horizontal. The obJect of the measurements noted here is to study the flux distributlon in the vicinity of one of these intersections. The region selected 1s 1llustrated in Fig. 46. The vertical colums of the matrix are indicated by the notation at the bottom of the figure and the row se- lected 1s noted on the right of the figure. The arrangement of the Plexiglas may be clearly seen from the loading dlagram, Fig. 5. Bare and cadmium covered indium traverses were measured along a horizontal in the midplane with the indium folls occupying the positions indicated by the small triangles in Fig. 46. This figure also shows graphically the results of the traverses. 8. Unit Cell Traverses Bare and cadmium covered indium traverses were made vertically across cell M-12 and extending through the one inch layers of Plexiglas adjacent to the top and bottom of this cell., This central cell contains stainless steel and uranium fuel disks but no Plexiglas. It 1is, however, sur- rounded by a one inch layer of Plexiglas on all four sides. The results of these traverses are shown graphically in Fig. L7, G, ' Neutron Flux Distribution Around a Poison Rod Two flux distributions,'each mitually perpendicular to the axis of one of the experimental poison rods supplied by the General Electric Company, and designated as Rod.#l*y were determined. One traverse was in the plane of the Plexiglas layer and the other perpendicular to 1it. The poison rod occupied the position in cell M-8 which had been used during the reactivity evaluations and which has been previously described in this report. This position is 1llustrated in Fig. 49. The foil positions for the hori- zontal and vertical traverses are illustrated in Figs. 48 and 49 respectively, and the diagonally cross-hatched lines in ° these two filgures represent sections through the Plexiglas. The horizontal flux distribution is represented graph- ically in Fig. 48 and the vertical distribution in Fig. 49. All the flux distribution curves taken in the vicinity of the poison rod show a discontinulty at the surface of the rod since it was not possible to make neutron flux determin- ations inside the rod. In Fig.. 48 those foils whose distances ¥ The specifications for this rod are glven in Appendix B. =80~ RELATIVE ACTIVITY 0.6 0.5 0.4 0.3 0.2 0.1 DISTANCE FROM MIDPLANE IN INCHES FIGURE 45. BARE AND CADMIUM COVERED INDIUM TRAVERSE ALONG AXIS OF THE REACTOR THROUGH CELL M-12 DWG. 21375 - T T <~ | | | | ~ J\ S N | | | NN N NN S NN N \\ \ ™~ - ~ NN ~ SN ~ 0 ~_ \[~ GRAPHITE REFLECTOR SO O O ~ — N OO O N SO N Y U ~ N ~ N e ™~ ~ ~ ™~ ~N \\ ~ — ~ o ~ N ~ @ NG ~ ~ o OIS ~ BARE INDIUM ACTIVATION ~ WU — MINUS CADMIUM COVERED NN Yo / INDIUM ACTIVATION N N LN Q \\ ~ T -~ - , 0 ~ = - -—.__.-_. ™~ ~ O o "‘\\ > ~N d J \\\ ) ~ CADMIUM COVERED S BARE INDIUM ACTIVATION INDIUM ACTIVATION I (H e . U ~ \ \ 9 N ™~ ™~ @ Q ~N 0 I I _ l I ™~ \1\\ ™~ \ 1 I I 4 8 12 16 20 24 28 32 36 18 RELATIVE ACTIVITY DWG. 21376 BARE INDIUM ACTIVATION A FOIL POSITIONS BARE INDIUM ACTIVATION f\ MINUS CADMIUM COVERED —O——0— INDIUM ACTIVATION ——————————— - O ROW 9 0.3} — = T ~\ O — lA”'Z/A A A A A alla alala afa l _ { | | l /0 . P 0.2f=© O O o —4ol,_ __ o ! / ] CADMIUM COVERED _0_0_’/0~ INDIUM ACTIVATION L LNM _‘-—-—J O.lp— —_— t— CELL N —>t®— CELL O -—>=but— CELL P —3}je— CELL Q —%4<— CELL R —»j«— CELL S —j I I . | L] | ] 2 3 | 2 3 | 2 3 3 0o I 2 3 I 2 3 I CELL DIMENSIONS IN INCHES FIGURE 46. HORIZONTAL BARE AND CADMIUM COVERED PLEXIGLAS INDIUM TRAVERSES AT INTERSECTION. TAKEN AT MIDPLANE é8 RELATIVE ACTIVITY 0.6 0.5 04 0.3 0.2 0.1 DWG, 21377 — 83 , Eg INDIUM ACTIVATION ___™ | I I I I I BARE INDIUM ACTIVATION BARE INDIUM ACTIVATION - MINUS CADMIUM COVERED _ /7 \/ INDIUM ACTIVATION // \ / \ T e . m— T —— — T — — —— — o— | CADMIUM COVERED PPLEXIGLAS%STAINLESS STEEL AND FUEL—*}-PLEXIGLAS-l l I | I I 1 2 3 4 5 6 7 DISTANCE FROM BOTTOM OF CELL M-i3 IN INCHES FIGURE 47 VERTICAL BARE AND CADMIUM COVERED INDIUM TRAVERSES ACROSS CELL M-12 AT MIDPLANE RELATIVE ACTIVITY 0.5 o D o w 0.2 0.1 DWG., 21378 - | I [ I | I I AIIRIITIRRIRZRIRIRINRINRNNY hyp O I BARE INDIUM ACTIVATION BARE INDIUM ACTIVATION MINUS CADMIUM COVERED / INDIUM ACTIVATION / / / i CADMIUM COVERED / INDIUM ACTIVATION IN REGION OF POISON ROD # |. ROD LOCATED IN CELL M-8 / - / — / I | I [ | I I | I 8 6 4 2 0 2 4 6 8 DISTANCE FROM AXIS OF ROD IN iNCHES FIGURE 48. BARE AND CADMIUM COVERED INDIUM TRAVERSES THROUGH PLEXIGLAS g DWG. 21379 Ay 1 l [ | [ NN BARE INDIUM ACTIVATION MINUS CADMIUM COVERED INDIUM ACTIVATION %\\\\\\\\\\ P S 0N - / n / { / CADMIUM COVERED M-8 , — ,’ ?j* INDIUM ACTIVATION | )I O\ \ g e (N N NN Z 1 I S $ 0 P ——— M - § — BARE INDIUM ACTIVATION h-(—-|———b-l:—— A l I I | I | 0 0.1 0.2 0.3 0.4 0.5 RELATIVE ACTIVITY FIGURE 49. VERTICAL BARE AND CADMIUM COVERED INDIUM TRAVERSES IN MIDPLANE THROUGH POISON ROD #1|. ROD LOCATED IN CELL M-8 ¢8 from the axis are indicated as zero were actually attached to the circumference of the rod by a very narrow strip of tape. V. EXPERIMENTAL STUDIES ON WATER-CONE REFLECTED ASSEMBLY A, Traversese in The Water-Cone Reflected Assembly: Bare and cadmium covered indium neutron flux traverses were made through cells R-12 and U-12 parallel to the axis, in the quadrant of the reactor in which the reflector had been altered to represent a mock-up of the water-cone reflected assembly. The traverses began at the mid-plane and extended through the simu- lated water cone. Both the bare and the cadmium covered folls were located in the Plexiglas strips one-half inch from the bottom of the cells. The results of these traverses are shown graphically in Figes. 50 and 51. ' Bare and cadmium covered indium traverses were made through cell H-12, in a quadrant of the assembly in which the reflector was unaltered, for comparison with those of R-12 and U-12. Cell HE-12 is the same horizontal radial distance from the central rec- tangular shell cell, M-12, as is R-12, Alsc safety rod No. 8 occuples the same relative position with respect to cells H-12 and M-12 as No. 7 occuples relative to celle R-12 and M-12. The bare and cadmium covered folls were again located in the Plexiglas one-half inch from bottom of cell. The graphic results of these traverses are shown in Fig. 52. "No satisfactory explanation for the neutron flux peak observed in the vicinity of the midplane in cell H-12 is immediately available. Most other axlal flux curves taken through this assembly follow the general shape of the zero order Bessel function of the first kind, especially near the midplane. The dotted portion of the bare indlum activation curve shows this relation. -86- RELATIVE ACTIVITY DWG. 21380 Ay 0.3 0.2— O.4p— | ! I | | I I I I BARE INDIUM ACTIVATION —— END OF STAINLESS STEEL AND FUEL " BARE INDIUM ACTIVATION MINUS CADMIUM COVERED INDIUM ACTIVATION CADMIUM COVERED INDIUM ACTIVATION /END OF PLEXIG|LAS EXTENSION pt—— WATER CONE — 3 5 10 15 20 25 30 35 a0 45 DISTANCE FROM MIDPLANE THROQUGH CELL U-IZ2 IN INCHES FIGURE 50. BARE AND CADMIUM COVERED INDIUM TRAVERSES THROUGH CELL U-I2 50 8 RELATIVE ACTIVITY 0.6}— 0.5|— BARE INDIUM ACTIVATION . D) 0.4 <—— END OF STAINLESS STEEL AND FUEL J 0.3 BARE INDIUM ACTIVATION ~ Q MINUS CADMIUM COVERED INDIUM ACTIVATION —-—.___\ S ™~ o~ | 0.2 ™ ——END OF PLE |XIGLAS EXTENSION . \\\ /‘- Q N 4 S 0 f<—— WATER CONE ——»= @ N\ 01— > ' CADMIUM COVERED A o INDIUM ACTIVATION e J A e . __7 S U J I | I I | 1 = O====0 0 5 10 15 20 25 30 35 40 45 DISTANCE FROM MIDPLANE THROUGH CELL R-12 IN INCHES FIGURE S5I. BARE AND CADMIUM COVERED INDIUM TRAVERSES THROUGH CELL R-12 DWG. 21381 | l I I 50 RELATIVE ACTIVITY DWG. 21382 Ay | | l | | N NN | \ N\ NN\ NN YN . 2p— \ AN NN \\ N \ \ N N NN N N\ S8ARE INDIUM ACTIVATION \ N N\ GRAPHITE - \ \\ \\\ ““REFLECTOR BARE INDIUM ACTIVATION 0.6 p— MINUS CADMIUM COVERED — / INDIUM ACTIVATION 04 T CADMIUM COVERED INDIUM ACTIVATION 0.2}— I 1 l I | | 0 3 6 9 12 15 27 DISTANCE FROM MIDPLANE THROUGH CELL H-I12 IN INCHES FIGURE 52. BARE AND CADMIUM COVERED INDIUM TRAVERSES THROUGH CELL H-12 68 VI CONCLUSIONS Results of the Critical Experiments on the various mock-ups of the Direct Cycle Aircraft Reactor indicate satisfactory agreement between ‘the theoretical and experimental values desplite the complexity of the assembly.9 The agreement obtalned between the measured and computed reactivity losses due to separation of the two halves of the core at the midplane, based on a homogeneous thermal reactor, gives added confidence to the application of methods of reactor calculations to complicated assemblies. The simllarity of the results obtained for power distributions from catcher foll activities and the counting of fission product radiations directly from the fuel disks indicates that this method of measuring fission rates and for normalizing indium foll exposures between different runs is satisfactory for the neutron spectrum studied here. Thle eliminates the necesslity for counting the fuel element back-ground prior to exposure. The fair agreement between the results obtained for rod calibrations by the stable period and the rod-drop methods indicates that the latter may be used with confidence. Since the rod-drop method is much less laborious, it may be used to check the integrated value of control rods and to quickly evaluate safety rods. Some effects observed during the course of the critical experiments have not been thoroughly studied. Among these are the peak observed on the control rod sensitivity curve; the peak 1n the reactivity vs Plexiglas thickness curve; the increase in reactivity with an Ilncrease in temperature of the assembly. Some of the changes in reactivity which accompany alter- ations of the reflector in different localities have not been explained. 9 1Ibid. Table 8, p. 70. . G0 - APPENDIX A By E. L. Zimmerman CONVERSION FROM SHELF TO RECTANGUIAR SHELL MOCK-UP During the early stages of the critical experiments on the Direct Cycle Aircraft Reactor, considerable controversy arose over the proper method of including neutron streaming in the calculation of age. One of the purposes of the comparison of the shelf and rectangular shell con- figurations was to estimate this streaming effect. Whatever the magni- tude of the streaming effect, it was anticipated that reducing the radial straight line void distences would .result in an Increase in multipllication. Contrary to the expected effect, converting from the shelf to the rectangu- lar shell configuration caused a substantlal negative change in multlpli- cation. While the original question was not answered, a partlal explanation for the negative change is given on the basis of a net shift of fuel away from the region of greatest importance in the assembly. To affect the change from the shelf to the rectangular shell type assembly, approximately half the cells were rotated through 90° in such a way that, in these cells, the fuel was moved 1/2" or 1.27 cm away from vertical longitudinal midplane of the assembly, Call-this dlsplacement a. The component of the displacement a along a radius at an angle © 1s given by a cos 83 b represents the distance of a particular cell from the axils. Averaging the radial displacement over interval 0= © < 17 T /’Tr/h gives - o P CZB: a8 | . cos8” ‘(-—" X , 1 - L - ' bde : . Y2 o "cos2 © | Approximately half of the elements were rotated, 80 the overall average:dis- placement amounted to A - as 0.56 cm | 2 For a bare reactor of radius R having the aéymptotic flux distribution of the actual assembly, the radial part of the flux is given by ¢ (r) = J5(By ), where B, = 2.405 , .1s the radial bucking component. Since, in the asymptotic R approximation, the flux is self ajoint, the lmportance function is [¢ (fl]e. The change in miltiplication due to increasing the fuel denslty by a small . amount &m at v is c &m @2 (r). Integrating over the entire radius one has R _ Skzzfcéhjg ¢2(flrdf . (2) This change in multiplication may also be found by adding the perturbation -91- (l*'éé) to the fuél cross sections in the critical equation: ‘ m Lrox VB 148 pin TE ST o(14 G2 Y4 o DB m Wherqzza refers to the thermal absoprtion cross-section in materials other than the fuel and) Y is the thermal absorption cross-section in the fuel. Solving (3) for 6k and neglecting second order term glves | Korp = (3) . m 2 - ‘ DB ' 8}{ = Za+ §I_D._ - 0.)4'7)4_ éE o (’-l-) Shonw | The term in the brackets is evaluated from the unperturbed reactor con- stants. The change in multiplication caused by shifting an element of fuel a distance = 18 ag C[¢2 (r+d) -¢2(’f'):' m=cmn2/ ¢ (r) drly (5) Integrating over the entire radius and substituting the value of ¢ found by comparing equations (2) and (4) gives: ’ R . troN g, b = 2 |t D" -le”(*““")*d* (e Sathat ™ | [y Forirar Substituting J, (Bg? for ¢ (r), the values from (L), and By = 0,0355 cm™t J glves the result §x = - 0.00938, ; or & loss in reactlvity of about 130 cents ;92; APPENDIX B ANALYSES OF REACTOR MATERIALS 1, Arco Soll Sample Principal Constituents Silicon, Aluminum, and Iron Present 1n Spaller Quantities Sodium, Calcium, and Potassium Quantitatively Measured Boron 0.05% Lithium 0.19% Magnesium 0.02% Maganese 0.05% Elements not Detected Cadmium and Copper 2. General Electric Company Experimental Poison Rod #1. Volume of By,C as Packed 50.77 cmd *Length of Rod Packed with ByC 28.8 inches Weight of B)C . .. . 81.0 grams Packed Density of B) 1.595 grams/cm3 *¥Boron per cc 1.248 grams Length of Lavite Plug 3/8 inch *¥Ineslde Measurement **Agsuming Pure ByC 3. General Electric Comcany Experimental Poison Rod Detalled in GE Print No. B 4098083 - 55. This poison rod consisted of a cadmium portion 20" long whose crogs-sectional area was a Maltese cross, having arms 0.082" thick and a total length of 3/4", formed from O.04L1l" cadmium sheet stock. This cadmium rod was inserted into an aluminum tube 30" long, 0.835" inside diameter and 0.083" wall thickness. The cadmium cross was held in position by a sultable plug. 4, Stainless Steel, Type 302 Maganese 1.11% Chromium 17.3 Nickel 7.6 Boron 0.001 Note: The 7.6 percent of nickel 1s below the specifications for type 302 stainless steel. This value was determined six times by three methods. -93- Spectrographic Analysis of Stalnless Steel Type 310 5. Ag < 0.04% Cd < 0.04% Hf < 0.5% Ni - 10.0%4 Ti < 0.005% Al < 0.04% Co 0.63 In < 0.15 Pb < 0.08 v 0.08 B < 0.004 Cr = 10.0 Mg< 0.02T Si 0.63 W <« 1.3 Ba < 0,02 . Cu- 0.15 Mn ~ 0.63 Sn < O0,04T Zn < 0.31 Be <« 0.0003 . Fe =10, Mo 0.31 Sr < 0.06 Zr < 0,15 Ca < 0.80 Ga < 0,04 Na <« 0.63 Ta < 1.3 6. Analysis of 28 Aluminum Used as Catcher Foils QUANTITIES PRESENT IN PERCENT Ag< 0.04 Co< 0.08 Mn - 0.04 - Sn <0.04 Al7 5.0 Cr 0.08. Mo« 0.08 Sr< 0.1 B « 0.01 Cu 0.0k Na <« 0.01 Ti< 004 T Ba<=0.04 Fe 0.3 Ni=< 0.08 V < 0.08 Be =< 0.001 K. 0.02 Pb< 0.08 Zn< 0,3 Ca < 0.08. Mg< 0.02T Si. 0.3 " Zr«< 0.15 7. Analysis of Square Aluminum Tubing Used in Matrix QUANTITIES PRESENT IN PERCENT Ag< 0.04 Co<0.08 Mn 0.0k Sn < 0.0h4 ° AlZ 5.0 Cr 0.08 Mo=0.08 Sr<0.1 B< 0.0l Cu -0.15 Na<=0.01 Ti<€0.04T Bg < 0,04 Fe 0.6 Ni< 0.08 V =< 0.08 Be< 0.00l K < 0.02 Pb< 0.08 Zn< 0.3 - Ca<= 0.08 Mg 0.02. S1 0.2 Zr=< 0,15, 8. Analysis of Plexiglas Used as Moderator Carbon 58.7% : Hydrogen 8.15% F; Ignited Oxides 0.0053% Spectrographic Analysis of Ignited Oxldes QUANTITIES PRESENT IN PERCENT Ag < 0.04 Co. € 0.08 Mo £ 0.08. Sr< 0.1 Al 0.2 cr 0.15 .Na< 1.0. Ti 0.2 B £ 0.01. Cu 0.15. Ni 0.3 V'j( 0.08 Ba <& 0.0h4 Fe» 0.5 P <0.08 2Zn< 0.3 Be 0.001 Mg 0.2 Si 2.5 Zr< 0.15 Ca 0.08 Mn 0.3 Sn < 0.08‘ T 9. Chemical Analysis of Stainless Steel Type 310 Fe Ni Cr 52 .8% 19,1% 25 0% ';9#; APPENDIX C SUMMARY of MATERIALS in REACTCR ASSEMBLIES A. BSBhelf Type Assembly 1. Core Density Mass Mass Volume | Volume™ Materlal gm/cc kg Fraction liters | Fraction Urenium (total) 18.7 b2 0.038 2.2 | 0.002 U-235 -- (38.1 )% -- - 'Stainless Steel ' : Type 302 7.87 384.5 0.357 48.9 0.040 Plexiglas 1.18 460.9 0.k26 391.5 0.320 Aluminum Tubes Type 2S 2.72 193.0 0.179 70.5 0.058 Void -- - - 708.9 0.580 1079.6 1222.0 2. Reflector Boryliium ~T1.86 1 1096.1 0.586 589.3 0.565 Graphite (AGOT) 1.72 957.8 0.425 556.8 | 0.440 Aluminum Tubes Type 2S 2.72 200.0 0.089 73.5 0.058 Void -- - ‘ -— 46.9 0. 037 ‘ 2253.9 _ 12865 : . B. Rectangular Shell Type Asgembly 1. Core Uranium (Total) 18.7 L1.07% 0.038 2.2 0.002 1U-235 - (38.3)** -- -- -- Stainless Steel ' . . Type 302 CT7.87 383.4 0.357 48.6 0.040 | Plexiglas 1.18 4L59.2 0.426 389.8 0.319 Aluminum Tubes Type 2S 2.72 192.9 0.179 70.6 0.058 Void -- -- - 710.2 0.581 1076.7 I25T.4 2. Reflector Beryllium 1,86 1096.1 . 0.451 589.3 0.429 Graphite (AGOT) 1,72 1119.0 0.460 650.6 0.475 Aluminum Tubes Type 2S 2.72 216.6 0.089 79.2 0.058 Void -- - - 52.0 0.038 PI31T 1371.1 ¥ These values are averages over the entire core or reflector and differ slightly from those given in the text, page 11, because of peripheral irregularities. *% The mass of uranium reported here .1s the amount loaded in the matrix; the mass of U-235 is that required to make the system critical with all control rods completely inserted. 9%