,)/ | .~ =« BECEIVED BY DTIE .4} 14 197y OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.5. ATOMIC ENERGY COMMISSION ORNL- TM- 3229 COPY NO. - DATE - November 19, 1970 FLUID DYNAMIC STUDIES OF THE MOLTEN-SALT REACTOR EXPERIMENT (MSRE) CORE R. J. Kedl 5 ABSTRACT In the MSRE reactor wvessel, fluid fuel is circulated at 1200 gpm down through an annular region and up through 1140 passages in the grephite core. The core design was based on preliminary tests in a one-fifth scale model, followed by detailed measurements with water solutions in a full-scale mockup of the reactor vessel and internals. This report describes the models, the testing, and the date from which velocity, pressure drop and flow patterns are deduced. It also describes how the measurements were extrapolated to molten salt at 1200°F in the actual reactor. The few observations possible in the reactor were consistent with the predicted behavior. KEYWORDS : MOLTEN-SALT REACTORS, CORES, DESIGN, DEVELOPMENT, FLUID-FLOW, MSRE, REACTOR VESSEL, FLOW MEASUREMENT, MODELS NOTICE This document contains information of a preliminary nature ond was prepared primarily for internal use at the Ouk Ridge National Laboratory. It is subject to revision or correction and therefore does not represent a final report. DISTRIBUTION OF THIS DOCITMENT 1S UNLIMITED This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. iii i CONTENTS o | _ ' Page No. 3 ABSTRACT o | INTRODUCTION - . | | | 1 DESCRIPTION OF MSRE CORE AND TEST PROGRAM 1 One-Fifth Scale Model y Full Scale Model T DESCRIPTION AND ANALYSIS OF TEST RESULTS 9 Volute and Core Wall Cooling Annulus 9 Reactor Vessel Lower Head | ' 14 Graphit.é Moderator Assembly ' | 20 - Reactor Vessel Upper Head ' : ' . 25 o Miscellaneous Measurements | 26 EXPERIENCE WITH THE MSRE | o | 31 REFERENCES | | | 32 _LEGAL NOTICE- This report was prepared as an account of work © | sponsored by the United States Government, Neither | the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any : legal liability or responsibility for the accuracy, com- 1 pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its use | . would not infringe privately owned rights, : " ‘-Q;,_\ . N ~ T e ms'fBIBUT‘lON OF 'l"Hj_lS DOCUMENT 13 UNL‘I-M’]ED " - INTRODUCTION The MSRE (Molten-Salt Reactor Experiment) is a 7.3 MV fluid-fueled, graphite-moderated, single region nuclear resctor. The fuel consists of uranium fluoride dissolved.in a mixture of lithium, beryllium end zir- conium fluorides. A unique feature of this reactor concefitris that the power is generated in circulating fluid fuel rather than stationary solid fuel elements. The nominal operating temperature is 1200°F. A detailed description of the reactor_concept and its components is available in many sources, for instance References 'l, 2 and 3. A program was undertaken to determine the fluid dynamic and heat transfer characteristics of the core. This report presents the results of that effort. Most of the experimental results preséfited here were obtained in the early 1960's. This report was not _ N written, however, until after the MSRE:nuclear operations were terminated in December of 1969. DESCRIPTION OF MSRE CORE AND TEST PROGRAM Figure'l shows an isometrié view. of tfie MSRE core. 'The-fuel enters the reactor vessel at 1200_gpm through acénstant,flowarea volute near the top of the cylindrical section. Becausé of the varisble pressure gradient in evolute of this'type, orifices are used to obtain a uniform angular flcw'distribution tb the core wall cooling annulus. The fuel then swirls down the core wall cooling annulus and into the reactor vessel lower - head. Radial vanes are placed in the lower head to destroy the swirl gen- erated by the volute. The lower{head,then serves &s & plenum to direct the fuel uniformly to the_mbderatbf'région. The moderator region is composed of long graphite'core b1ocks, squéré-in cfossisection; and with grooves. cut longitudinally in the'h'faces; ‘When these stringers are assembled veftically together,'the_grooves_formgthe fuel passages. Figure 2 shows en ‘isometric and a planVViév ofra*small_clustef or core blocks. After u-paésifig through the mpdératbr,,the-fuellthen goes into the vessel upper. head vhich serves as a collection plenum end directs the fuel to the outlet pipe. Each of these various regions of the core will be described in more detail in the'appropriaté section of this report. GRAPHITE SAMPLE ACCESS PORT “ FUEL OUTLET =~ _ - CORE ROD THIMBLES - = " e LARGE GRAPHITE SAMPLES = CORE CENTERING GRID CORE SUPPORT FLANGE GRAPHITE ~MODERATOR Al CORE BLOCKS ' 5 = - i ! iz i F i 3 i, * N ll i i 1 - i FUEL INLET -/ " REACTOR CORE CAN —~ REACTOR VESSEL - ANTI-SWIRL VANES VESSEL DRAIN LINE FIGURE 1. 1668 ORNL-LR-DWG G1097RIA FLEXIBLE CONDUIT TO CONTROL ROD DRIVES COOLING AR LINES ACCESS PORT COOLING JACKETS REACTOR ACCESS PORT SMALL GRAPHITE SAMPLES HOLD-DOWN ROD OUTLET STRAINER FLOW DISTRIBUTOR VOLUTE ~\—_FLOW DISTRIBUTION ORIFICES T~ CORE WALL COOLING ANNULUS f 2 MODERATOR _ SUPPORT GRID MSRE REACTOR VESSEL 3 - ORNL-LR~-DWG 56874R PLAN VIEW NS | TYPICAL MODERATOR STRINGERS SAMPLE PIECE r NOTE: NOT TO SCALE » FIGURE 2. TYPICAL GRAPHITE CORE BLOCK ARRANGEMENT L The MSRE core development program was divided into two phases. The o (;; first phase consisted of building & 1/5 linearly scaled plastic model and | testing with water. This was considered to be a rough and preliminary - design checking device. The second phase consisted of bfiilding a full scale carbon steel and aluminum model and testing with water. The only data,pfesented in this report will be from the full scale model, however, g brief description of the 1/5 scale model and the way it was used is given in the next section. Early concepts of the MSRE called for a'variablé speéd pump. It was planned to operate the reactor at flow rates below the design flow of 1200 gpm. The lowest flow rate was undefined but could have been as low as 25% of design flow. Later in the design stage, this reduced flow spec- ification was dropped and design flow rate was fixed at 1200 gpm. This change occurred during'the\testing'of‘the full scale core model. 4s a result, data were taken at flow rates ranging from 1200 gp;n to 300 gpm, but the emphasis in this report is on the 1200 gpm data. Late in the operating history of the MSRE, the fuel pump was connected to & veriable frequency unit, and the reactor was operated at reduced flow rates. The purpose of these specisl runs was to study Xe-135 behavior. The "worst case" as far as lateral temperature gradients in the core is concerned, would be when it was operated at 5.5 MW (thermal) at half the design flow rate for a period of about 3 1/2 days. Figure 3 is a list of reactor parameters and physical properties of interest in this study. One-Fifth Scale Model A small transparent plastic model of the MSRE core, linearly scaled down by a factor of 1/5, was built and tested. The particular core com- ponents simulated in this model were the inlet volute, flow distribution orifices, core wall cooling annulus, lower vessel head with swirl killing venes, moderator support and the moderator fuel channels which were simu- lated with a tube bundle. A photograph of the model is shown in Figure L. The particular scale factor of 1/5 was chosen because a geometrically similar model of that size tested with water st sbout 95°F, and at a flow rate such that its fluid velocities are equal to those of the reactor, - will have the same Reynolds Number as that in the actusl reactor core when w " $# DESIGN CONDITIONS FUEL SALT FLOW RATE REACTOR POWER FUEL INLET TEMPERATURE TO CORE FUEL OUTLET TEMPERATURE FROM CORE REACTOR OPERATING POWER f FUEL SALT o CGMPOSITION LiF BEFz 'ZrFu UF,, LIQUIDUS TEMPERATURE PROPERTIES AT 1200°F DENSITY SPECIFIC HEAT THERMAL CONDUCTIVITY - VISCOSITY . PRANDTL NUMBER = (0 47)(19)/(0 83) HASTELLOY N - SPECIFIC GRAVITY _THERMAL CONDUCTIVITY AT 1200°F - . SPECIFIC HEAT AT 1200°F GRAPHITE “GRADE " POROSITY (ACCESSIBLE T0 KEROSENE) - WETTABILITY MEAN COEFFICIENT OF THERMAL EXPANSION (70-1200°F) FUEL SALT ABSORFTION AT 150 ps1 (CONFINED T0 SURFACE) DENSITY ~ SPECIFIC HEAT (1200°F) THERMAL CONDUCTIVITY WITH GRAIN AT 68°F NORMAL TO GRAIN AT 68°F WITH GRAIN AT 1200°F - NORMAL TO GRAIN AT 1200°F. COEFFICIENT OF THERMAL EXPANSION WITH GRAIN AT 68°F ~ NORMAL TO GRAIN AT 68°F - = ~ *ESTIMATED UNIRRADIATED 117 1bs/Ft3 1 0.42 Btu/1b °F 80-Btu/hr ft °F 45 Btu/hr ft °F - 0.56 x 10-6/°F 1.7 x 10°6/°F — **GRAPHITE NOT WET BY ‘FUEL SALT AT REACTOR CONDITIONS FIGURE 3. 1200 gpm 10 Mw?t) 1175°F 1225°F 7.3 Mw(t) 65.0 mole % 29.1 mole % 5.0 mole % 0.9 mole % 813°F 141 1bs/ft3 0.47 Btu/1b °F 0.83 Btu/hr ft °F 19 1bs/ft hr 10.7 8.79 “11.71 Btu/hr ft °F 0.139 Btu/1b °F - 7.81 x 10-8/°F CGB 6.2% *% 0.2% ~ IRRADIATED 117 1bs/ft3 35 Btu/hr ft °F 20 Btu/hr ft °F 23 Btu/hr ft °F* 13 Btu/hr ft °F* REACTOR PARAMETERS AND PHYSICAL PROPERTIES ONE-FIFTH SCALE CORE MODEL FIGURE k. " ) T circulating fuel. Measurable varlables of the model are then related to those of the reactor by the fOIIOW1ng proportlonalltles. (1inear dlmensions) '—'5 (llnear dimensions) MSRE Model (fluid veloc:.ty)MSRE = (fluid veloc1ty)M del _(Reynolds Number) —-(Reynolds Nuniber)Model (fluid age)MSRE (fluld age )y de1 . (relatlve fluid pressure gradlents in ft- fluld)MSRE _ (relatlve flu:.d pressure gradients in ft-fluid) p;(turbulent heat transfer coefflcients)MSRE = 0, 63 x (turbulent ‘Theat transfer coeffic:Lents)Mdel Model 'Since the model'was'so small not every surface andpchannelrof the reactor core exposed to salt was simulated As a8 result' the ‘model did not give exact comparlsons, and was used only as a rough design check1ng device to- establlsh early in the program the acceptability of major core concepts. For example it was used for studies of the volute de51gn ‘end Spaclng of flow dlstrlbutlon orlflces, design of swirl killing venes in lower head and prellmlnary measurements of solids settling characterlstics of the 1ower head. ' ' ' Full Scale Model _ The full scale MSRE core model is almost an exact dnpllcate of the actual reactor. Flgure 5 is 8 photograph of this model The model is con- structed of carbon steel with' the exception of the core blocks and part of the moderator support grld.whlch are of alumlnum The core blocks were ' ;made by extrudlng alumlnum approximately to shape 1ncluding the ‘longitudinal grooves ‘and then taklng only 8 finish cut on the b side surfaces. Most of - the tolerances of the reactor were 1ncreased (normally doubled) in the model q';for economic reasons. In addltlon other simpllflcatlons were made to reduce the cost if they were presumed to have a small effect on the fluid 'dynamlcs. The vessel was construeted‘w1th & large glrth flange just over . the volute so that the~internals could be removed with relative ease, Several transparent plastic windows wererplaced'in.the1VESSel heads, core ' wall cooling annulus and volute for viewing. Numerous holes were drilled into the vessel walls at various plaeces for fluid measuring probes. A carbon steel loop was built to operate this model and consisted of a pump, i FIGURE 5. s E3 - 1 Fs & FULL SCALE CORE MODEL & " »y *) 9 gate valve to control the flow orlflce flcwmeter loop cooler, 5400 gal 'surge tank and an 1on-exehange system for remov1ng salt 1njected durlng fluid age measurements., e All the initial data from the full scale model was taken Witn,the loop filled with water and operated between T5°F and 80°F. This results in a_Reyneids Number fer the model Ebeut'four times that of the reacter. Later in the progrem.a thickening-agent (Jaguar-508 by Stein, Hall & Co.) was added to the system in order to simulate Beynolds'Numbersg'end the measure- ments uhich were strong functions of the Reynolds Number, were repeated. AlthOugh7Jaguar—508 imparts-non;Newtonian characteristic to the water, in the low concentratlons that were used in these tests, this was a negligible con51deration. As data is presented in this report, it will be noted whether or not exact Reynolds Number similarity existed. Items to simulate the three eontrol rods and thelsurveiliance specimen holder were not included in the core model because their design was not sufficiently well known when the model was built. Rather, the regular graphite matrix was continued throughout this regions. | DESCRIPTION AND ANALYSIS OF TEST RESULTS Volute and Core Wall Coollng Annulus The main fuel loop piping'in the MSRE is 5 in. Schedule 4%0. Just prior to.entering the core vessel volute the pipe size is increased to 6 in. The cross-secfional flow area of the 6 in. pipe is approximately;equsl to that of the vblute 28.8 in.? and 26. 0 in.z, reSpectively. ‘The'uolute is a con- stant flow ares type and the tail end is hydraulically connected.to the head end 50 there is rec1rculat10n. One characteristlc of this type of volute is the varlsble static pressure around it therefore, in order to obtain ”a unlform angulaer flow dlstrlbution it 1is necessary to use oriflces between ~ the volute and core wall coollng annulus The orifices are 3/h 1n. and. occur -in stacks of 3. At the head end of the volute the orifice stacks are 5 deg apart. At the tail end of the volute because of the lower fuel veloc1ty -and resultlng higher statlc pressure the orifice stack spac1ng is increased to 22 1/2 deg. This orifice distribution was determined from the 1/5 scale o 10 model. The orifice holes are drilled at an angle of 30 deg with the tangent in order to maintain a tangential velocity component in the annulus. The resulting high heat transfer coefficients cool the core vessel wall and the reactor core can. | h Figure 6 is a plot of the exéerimentally measured centérline velocities in the volute as a function of angular position and at various flow rates. At 1200 gpm water approaches the core through the 5 in. pipe et a meen velocity of 19.2 ft/sec. Immediately inside the volute the velocity Jumps - to about 23 ft/seé because of recirculation around the volute. At the tail end of the volute the fielocity is about 10 ft/sec. The centerline velocities cannot be taken as absolute representations of flow rate, par- ticularly at the inlet where two fluid streams of different velocities merge. Nevertheless, the linear decrease in velocify is a good indication thét the wafier'is distributed uniformly to the corejwall cooiing annulus. At the head end of the volute the mean velocity through‘tpe'orifices is - about 4 ft/sec and at the tail end of the volute the mean velocity through the orifices is sbout 18 ft/sec. Figure T is & plot of the centerline velocity in the core wall cooling annulus as a function of elevation. Note that the velocity decreases as the water moves‘down.the ennulus, because the tangential component is attenusted. Figure 8 is a plot of the centerline velocity at the bottom of the core wall cooling annulus as a function of angular position around the core. Note that it is quite flat, indicating uniform flow to the reactor vessel lower head. Data for Figures 6, T and 8 was teken with water in the loop. The Reynolds Numbers involved in the volute and core wall cooling annulus are so high (over,lOh in all ceses at 1200 gpm) that exact Reynolds similarity is-notrimpbrtant, énd the reactor vessel containing fuel salt will have the same velocity pro- files. | At this point it would be informative to compute the tempersture dif- ference between the bulk selt in the core wall cooling ennulus and'the.vessel wall (Hastelloy N). At the midplane of the core (sbout 30 in. up in the annulus) and at 1200 gpm, the fluid veiocity in the annulus is 7.2 ft/sec (Figure 7). Now, molten salt behaves as a conventional Newtonian fluid so that standard. heat transfer relationships may be used. From the Dittus-Boelter equation and with physical properties from Figure 3 we can compute a heat n 11 ORNL-DWG 64-6T721Al 24 ,/'.\.‘ / / o / 7 o~ / 20 |4 % o’ N / % | T~ MEAN INLET '"?\\\\\ . \ / = PIPE VELOCITY ! < 46 '\o- / _ e - ' .\ \ / 3 * ol / S ' | ' .wo gplm II = 12 * ¥ q ~ . T ”,4-0..._.\ o~ e o Il ) - ~ S - | | g 8 * ] | 7 5 - | egogn |/ 3 . | L ) L e —— . o CTmees )/ / 4 .-—-—.'._“___—— -~ ——lg 300 gpm 7~ o o..__,___.____.___. - 0 0 45 90 . 135 180 225 270 315 360 - 8, ANGLE FROM INLET TANGENT (deg) FIGURE 6. ANGULAR DISTRIBUTION OF FLUID VELOCITY AT CENTER LINE OF VOLUTE e VERTICAL DISTANCE UP CORE WALL COOLING ANNULUS (in.) 12 ORNL-DWG 64-6723A1 70 ./ | ELEVATION OF VOLUTE 7 . fi f ® @ .;’,—O . 3009p'm/ | / /./ o - y | './600gpm ./ o [ / - / 1200 gpm IV 20 i — /l // | o | o ¢ 10 ' ’ | @ o ® oL 1] [ 0" 2 4 6 8 10 FLUID VELOCITY AT ANNULUS ¢ (fps) FIGURE T. AT CENTER OF CORE WALL COOLING ANNULUS . VERTICAL DISTRIBUTION OF FLUID VELOCITY 12 N » 4 FLUID VELOCITY IN ANNULUS ¢ (ft/sec) 13 ORNL-DWG 70-11789 O O ¢ 1200 gpm o v 5 4 3 d'\ : ) O ‘ A) i O 600 gpm O 2 | I ¢ O | O 300 gpm - 1 ' 0 "0 45 9 135 180 225 270 315 360 ~ ANGLE AROUND CORE (deg) FIGURE 8. ANGULAR DISTRIBUTION OF FLUID VELOCITY AROUND BOTTOM OF CORE WALL COOLING ANNULUS 1k transfer coefficient at this position in the core wall cooling anfiulus (1 in. thick) of 1370 Btufhr—ft2-°F. The heat generation rate in the vessel wall at.this point has been estimated to be about 0.2 watts/cm3. The vessel wall is 9/16 in. thick so the heat flux to the salt, assuming the outside surface is insulated, will be 905 Btu/hr-ftz. The tempersture drop across the fluid boundary layer will be only 0.66°F. Now the tempera- turé drop in the vessel wall with an internal heat source, and assuming - again that the outside surface is insulated, is represented by : 2 -9t AT = 2% where: t = wall thickness = 9/16 in. = internsl heat source = 0.2 w/cm3 k = thermal conductivity - 11.71 Btu/hr-ft-°F Evalueting gives a temperature drop in the metal wall of 1.81°F. Therefore the overall temperature drop from the outside surface of the veséel wall to the salt in the core wall cooling annulus is the sum of the above or " only 2.47°F. It is beyond the'scope of this report to include many detailed thermal analyses such as sbove. These analyses have been made and many are reported in References 8 and 9. It was felt however that one such computation would be worthwhile to give the reader an idea of the order of magnitude of these effects. Because of its rather lo# power density, lateral temperature gredients are quite low in the MSRE. Reactor Vessel Lower Head The lower plenum of the core vessél ié formed by a‘standard 60 in. OD. ASME flenged and dished head, containing anti-swirl veanes and a diain line configuration. The anti-swirl venes consist of 48 plates starting about 2 in. up in the core wall cooling annulus and extending along radial lines into the lower head for about 38% of the radial distance to the core center- line. They are slightly elevated off the core vessel wall, fihfis eliminating as much corner area as possible where settled solids could accumulate. The vessel drain consist of e short section of 1 1/2 in. pipe extending slightly | up into the vessel head at the centerline, and having a éonical umbrella over it. In addition it incorporates a secondary drain which is a 1/2 in. con- centric tube coming up the middle of the primary drain, penetrating through 4) o) 41 » 15 ‘the drain pipe wall just inside the vessel and wrapping around the drain pipe horizontally ebout 90°. The conicel umbrella will prevent gross settling of solid particles (should they exist) into the primary drain line, however, fuel containlng solids mey still drift in and out of the unbrellea and over a long period of time, solids could still plug this drain. The 1/2 in. line serves &s a safety drain since it is designed to prevent slow migration of fuel in and out. Because of its size, it would drain the ‘reactor too slow under normsl conditions. The secondary drain terminates inside the primary drain just below the freeze valve. The- object of the anti-swirl vanes is to prevent the swirl generated in the volute from penetrating 1nto the lower head end creating an excessive radial pressure gradient. The unlformlty of fuel flow through the graphite moderator region is a direct function of thls pressure gradient. Figure 9 is a plot of the experimentally observed radial pressure gradient as measured by wall pressure taps. Since the flow in the lower head is 3- dimensional, the static pressure at the wall is not an ebsolute measure of the pressure influencing flow through the moderator region, nevertheless it is & good indication. Note that the pressure is slightly higher at the center, therefore, one would expect a slightly higher flow tfirough the moder- ator near the center. This was measured to be the case as will be pointed out in the sectlon on the moderator. & As the water goes through the anti-swirl vanes and heads toward the vessel centerline, it is turned by the lower head and produces & high velocity Jet adjacent to the wallaerigure‘lo,is arprofile of this jet measured at & radius of 17 inches and at b poeitions‘90° apart The flow rate was lébcrgpm._ Again, since the Reynolds Number is so high, the same jet will exist with salt in the vessel This Jet does not per51st much farther towsards 7 the vessel centerllne when the flow character becomes gusty but still remains ~ turbulent. From this veloc1ty profile a heat transfer coefficient can be eStimated'by assuming parallel plate geometry. The equlvalent dlameter : would be L times the dlstance from the wall to the peak velocity. The cal- culation Wlth the Dittus-Boelter equation yields a heat transfer coefficient 7'_for fuel salt of 540 Btu/hr-ft ?-°F.' ‘Heat transfer coefficients were also measured with a locally developed "heat meter." Basically, it is an aluminum cube with a electric heater on one surface end a thermocouple in its interior. The surface opposite ORNL-DWG 64-6722A1 ANNULUS %ODERATOR c;/ ///// | // Y ///// ~ VANE l ——} — PRESSURE TAPS —— - o 04 — | i o g— I 2 - | 'gfi 02— FLOW RATE = 1200 gpm - > I = | - n 2 o 2 » g& K / 1 O o ® a -02 — . FIGURE 9. RADIAL PRESSURE GRADIENT AT WALL OF LOWER VESSEL HEAD 4) 2 AY " 2.5 ~ .N o o DISTANCE FROM WALL (in.) o 0.5 0 17 ORNL-DWG 64-6727Af DATA POINTS MEASURED 90° APART RATE = 1200 gpm qbwq::\\\\\\ AN | AA } oe : A Ao - 0 05 10 15 20 25 30 ~ FLUID VELOCITY (ft/sec) FIGURE 10. VELOCITY PROFILE AT WALL JET IN LOWER VESSEL HEAD AT RADIUS OF 1T IN. 18 the heater is then exposed to the fluid stream and mounted flush with the vessel surface. The heat meter is thermally insulated from the vessel walls. It is illustrated and described in more detail in.Ref. 10. By measuring the electrical power input and the temperatfire difference , between the meter and the fluid, the heat transfer coefficient can be cal- culated. The technique yields a heat transfer coefficient in a thermal entrance region and it is necessary to convert them to the "ror downstream" case. This can be done with correletions in References 4, 5 and 6. In addition, the measured coefficients must be converted from water to salt with the Dittus-Boelter equation. Heat transfer coefficients were measured at a radius of 1T in. end 4 in. with water in the loop and were checked with water thickened with Jeguar for Reynolds Number similarity and found equal. The data, after being converted;to the far downstream case and from water to salt, are shown in Figure 11. Also shown on this plot is the heat - transfer coefficient as calculated from the wall Jet. Note that the slope of the curve indicates turbulent flow. More confidence is pléced in the data at 17 in. than at 4 in. because the flow is well defined at the larger radius and less defined and gusty at the shorter radius. It is ex- pected that the predicted values of these heat transfer coefficients are conservative because they do not includethe effect of thermal convection which is expected to contribute roughly 100 Btu/hr—ft2-°F to the overall coefficient. | Consideration was given to the possibility of a éource of.nuclear power oscillations in the MSRE core existing because of fuel short cir- cuiting in the lower head. If a fraction of the fuel did short circuit, then a corresponding amount of fuel must reside a little longer in the head. This results in the short circuiting fraction entering the moderator and little cooler than the mean and the short-circuited fraction entering the moderator and-a little hotter than the mean. Power oscilletions result because of the fuel temperature coefficient of reactivity. Tests were conducted on the model by injecting a conductive saline solution at a constant rate into the lower head at wvarious locations and measfiring the oscillations in electrical conductivity at the wvessel outlet. These measure- ments indicete that the power oscillations originating from this mechanism- should be well below 0.01% of the mean power. (1] i " HEAT TRANSFER COEFFICIENT (Btu/hr ft2 °F) 19 ORNL-DWG 70-11790 1000 - ' : : O FOR DOWNSTREAM IN REACTOR AT RADIUS OF 17 in., DETERMINED WITH WATER IN LOOP / @® FOR DOWNSTREAM IN REACTOR AT RADIUS OF 17 in., . 6/ DETERMINED WITH JAGUOR IN LOOP : / A CALCULATED FROM VELOCITY PROFILE OF WALL JET AT RADIUS OF 17 in. | N A FOR DOWNSTREAM IN REACTOR AT RADIUS OF 4 in., A - DETERMINED WITH WATER IN LOOP — , SLOPE = 0.8~ T 200 - | B / / A 500 100 - . S _ o 100 _ 200 500 ' 1000 , 2000 FLOW RATE-(gpm) FIGURE 11. HEAT TRANSFER COEFFICIENTS IN REACTOR VESSEL LOWER HEAD 20 Graphite Moderator Assembly In flowing through the reactor moderator essembly, fuel first passes through a moderator support structure and then through the modefator core blocks. The support structure consists of two assemblies, the lower of which is a Hastelloy-N cross structure (Figure 1) which is the main support structure for the graphite. Resting on this'is_a grid cdnsisting of two layers of rectangular graphite bars, one 1ayer resting on the other and perpendicular to it. The purpose of the graphite assembly is to position and hold the graphite core blocks, and to_compensate'fOr a difference in thermal expafision between the Hastelloy N and graphite. The resulting square passages in the graphite grid are small and the fuel velocity is high (approx. 4 1/2 ft/sec), and therefore its pressure drop iélhigh. Comparatively, the salt velocity through the bulk moderator fuel channels is low (approk. 0.7 ft/sec) and its pressure drop is low. Figure 12 shows the experimentally determined head lcsSPacfbss{théamgder_ ator assembly as measured in the model. 'A_small correction has been applied to this data to account for small holes drilled through the support grid of the MSRE, as will be discussed later in this section. Also shown in this figure is the head loss across the support'grid as measured in & | separate model made for this purpose. The difference between these curves is the head loss in the moderator channels themselves including entrance and exit losses. For these fests, the loop wés filled with water. Note that the slope of the curves in Figure 12 is 2.0 as would be expected since the Réynolds Number with water is over L4000 in the fuel channels and the bulk of the head loss is due to form losses in the graphite grid. With fuel salt the bulk of the head loss would still be from form lbsses. The «ReynOIdstumber in the fuel channels would be.approximately 1000 so the flow character would theoretically be laminar, however, much of the tur- bulence generated by fuel passing through the tortuous inlet configuration would persist through the fuel channels. As a result one might expect a very slightly lower slope to the curves in Figure 12 with fuelvsalt in the system. | , . As pointed out, the bulk of the pressure drop through the moderator assembly is due to the graphite grid. There is also verf little room for cross flow in the space between the graphite'grid and the entrance to the #} 1" » HEAD LOSS (ft-fluid) 21 ORNL-DWG 64-6724Al 1.0 . SLOPE = 2.0 0.5 0.2 PRESSURE DROP RESULTING | FROM SUPPORT LATTICE o1 0.05 0.02 100 200 500 1000 2000 5000 10,000 * FLUID FLOW RATE (gpm) FIGURE 12. FUEL SALT PRESSURE DROP ACROSS MODERATOR ASSEMBLY 22 fuel chennels. Therefore the uniformity of flcwfdistribution 8cross the moderator assembly is controlled to a large extent by the uniformity of flow through the orifices in the graphite grid. The flow distribution ainong the core fuel channels was éxperimentally measured by two different techniques. The most successful technique was to simply inject a small amount of electrically conductive salt solution into the water flowing in a fuel chennel and measure the time required to pass two conductivity probes a given distance apart. The device used consisted of a length of 1/k in. tubing blanked off on the end, and a small hole drilled into the side.where the salt solution was injected. The tubing was inserted down into & fuel channel about 6 in. from the top. Access to the fuel channels “was availsble through the viewing ports in the upper head of the vessel. The assembly was made to readily slip into and out of the fuel channels, and was small enough in cross secfiion so that its presence would not sig- nificently change the flow rate. The conductivity probes were sbout 3 in. apart and sbout 2 in. up from the injection hole. The device was calibrated in e special test fixture. The flow rate was measured in 77 more or less randomly chosen passages and the data appear in Figure 13 as a function of the core radius. Note first that 2 separate regions are present which represents channels that are mutually perpendicular to each other. These regions are characterized by different inlet configurations beéause of their position over the graphite grid. A plsn view of the graphite grid with the position of the fuel channels superimposed is included in the figure. A separate' plastic model was built of a.fEW;adjacentrfuel channels with the graphite grid simulated to investigate methods of correcting this problem. Based on these tests, it was determined that 0.104 in. holes (no. 37 drill) drilled through the upper graphite grid bar and directly under the starved channel would equilibrate the flows. A typical location of one of these holes is shown in Figure 13. Note also from this figure that the flow rate - decreases slightly'with increasing radifis. This was expected and is due to the radial pressure gradient in the lower head as discussed earlier (see Figure 9). Actually this flow distribution is beneficial because - the flow rate is highest where the power density is highest, resulting\ in afbetter thermal utilizetion df the fuel. In a large power producing reactor, for instasnce, one would strive to match the flow distribution © 0 0 L) A FLOW RATE IN FUEL CHANNEL (gpm) 2.8 2.4 2.0 1.6 | 1.2 0.8 0.4 0 23 _ORNL-DWG 64-6T19A}. | | CHANNELS DESIGNATED BY @ CHANNELS DESIGNATED BY A | | ' PLAN VIEW OF CORE SUPPORT BARS WITH FUEL CHANNELS SUPERIMPOSED 7| | | ] | - TYPICAL HOLE THROUGH SUPPORT STRUCTURE TO GET MORE FUEL TO 0 THE STARVED CHANNEL T ? v REACTOR CORE CAN \é oo °—*8.- g oo o 2 b e At . L‘A ,‘A A LEAST SQUARE LINES 7 ] @ DATA TAKEN BY DIFFERENT TECHNIQUE IN ANNULUS Z BETWEEN GRAPHITE AND MODERATOR SHELL Z | 7 l , - Z 6 . 12 18 24 | RADIUS'OF;CHANNEL FROM CORE CENTER LINE ;(irn.) 'FIGURE 13. RADIAL FLOW DISTRIBUTION OF FUEL SALT IN FUEL CHANNELS 30 2h ‘with the pofier distribution so that the fuel outlet temperature across the core would be.approximately equal. Lastly note from Figure 13, the rather large scatter of data points around the least square lines. This is a result of inherent inaccuracies in the flow instrument and also because of tolerances in the orifices formed by the graphite grid. Recall thet the flow distribution through the moderator is controlled by these orifices because their pressure drop is high and cross flow sbove them to level out possible pertubations is low. The tolerénces in making the model were normally double those of the reactor, so that a greatér variation in orifice width was expected. The nominal orifice width is 0.375 in. In -the model we measured these orifice widths with go nogo gages as high as 0.400 in. and as low as 0.350 in. Estimates indicate that this variation is enough to account for a large fraction of the scatter in Figure 13. It was therefore concluded that the large scatter in flow distribution ex- perienced in the model will not be present in the reactor. | The second technique used to measure flow rates through the moderstor channels was to install pressure taps in the aluminum core blocks and meesure the pressure drop. Pressure taps were installed on 2 sides of 9 core blocks yielding information on 18 fuel channels. The core blocks were calibrated before installation in the model. Data from this technique was consistent with date obtained from the salt injection technique, although with more scatter. The data points at the extreme radius of the moderator are shown as squares in Figure 13. They represent the flow rate between the graphite moderator and reactor core can. The units are given as gpifi which is misleading, because the hydraulic parsmeters (e.g., cross sectional flow " grea) in this flow region are not accurately known. To be specific this is the flow rate corresponding to the measured pressure drop in a standard fuel channel. The significent observation is that these data points fall more or less on an extension of the least squares 1ine.. | | In the MSRE and in the model, the régular patternlof the graphite support grid is discontinued near the centerline of the core where the control rods and the surveillance speciment holder are located. This allows greeter salt velocities past these components for cooling. As noted previously however, the control rods and specimen holder were not similated in the model. Instead, the regular pattern of core blocks was continued \ (:ur‘ A 25 throughout this region..ln the model, sbout 16 regular fuel chemnnels were’ ‘directly affected snd a few more were indirectly affected by the discon- tinued support grid. The average flow rate measured through these channels by the conductivity probe technique was 2.3 gpm. Two values measured in this region by the calibrated fuel channel technique were 3.6 and 3.7 gpm. Fluid velocities in units of ft/sec would be sbout the same if control rods and the surveillance specimen holder were present. This is more than adequate for cooling so no attempt was made to resolve the difference in flow rate measured by the two different techniques.~ | - A suggested scheme for core power oscillations is due to changes in the character of flow in the centermost 16 fuel channels. The Reynolds Number of these channels is in the order of 3000 and is therefore in the transitional region. It was speculated that the character of flow could oscillate from laminar to turbulent between parallel channels. ‘To test this hypothesis, & flow visuelization study was made using an optically birefringent solution of Milling Yellow dye in water. An almost full scale model of h_parallel.fuel channels with inlet conditions similar to the reactor was built from transparent plastics. With two mutually perpendiculer polarized’plates,'we were able to observe clearly the transition between lamlnar and turbulent flow. The transition was smooth and contlnuous and no oscillatory action could be detected either in a single channel or between channels that could be regarded as significant. Many detailed temperature distribution calculations have been made .fer the MSRE core based_onfthe_results presented in this report. Results ~ of these eanalyses are presented in Refs. 8 and 9 and will not be presented “here. Reactor Vessel Upper Head The reactor vessel upper plenum is 31m11ar to the 1ower plenum in - f“that it is fbrmed from a standard ASME flenged and dlshed.head ‘The fuel , enters the upper head uniformly across the dlameter from the moderator fuel 'channels. The fuel then moves radially to the center and leaves the core through a 10 in. outlet plpe._ The‘model showed no tendency of the water to | rvbrtexpinto the outlet plpe., After the model was built, a strainer structure .(Figure 1)'Was addeduto the reactor ofitlet'design which penetrated down - into the Upper plenum. Its purpose was to catch graphite chips (graphite floats in fuel salt) should they break loose from the moderator. This 26 device ‘was not installed in the model but it would not be expected to influence the flow significantly. Velocity profiles in the model were determlned in the upper: head,by & fluid age technique. Electrically conductive salt’ solutions were injected as & step function at the core inlet. Conductivity probes were placed &t selected locatiqns in the upper head presumebly along the fluid étreamlines. : By measuring the time required for these conductive pulses te_pass between the probes, the average velocity'between the probes can be calcuiated. There are, of course, inaccuracies in the calculation because they assume knowledge of the streamllnes and they neglect lateral mixing between streamlines, Velocity profiles estimated by this technlque appear in Figure 14. Plotted is the average velocity between the prdbes indicated (see sketch) as a function of distance from the vessel wall,.assfiming that the streamlihes were parallel to the wall. At the rated flow rate (1200 gpm), the Reynolds Number in the upper head is high for both water and fuel salt and the flow is-turbfilent. Heat transfer coefficients to the vessel walls can therefore be estimated with Dittus-Boelter type equations. Assuming parallel plate geometry where the equivalent diameter is twice the channel width, the estimated heat trensfer coefficient is a little over 100 Btu[hr-ft2—°F neglecting thermal convection. This value hes been generally confirmed with heat meter measurements teken at approxi- mately the same location. The calculation is asétmed conservetive because the entire upper head would be a thermal entrance region (low length/diameter ratio). The region near the knee of the upper head will have:a lower heat transfer coefficient but the heat generation in the MSRE vessel wall in this region is very low. In addition, some small channels have been milled in the MSRE core support flange to allow & small flow of cool fuel to short circuit directly from the core wall coeling annulus to this.regien of the upper head. 'The flow rate in this'circuit has been estimated to be between 25 and 50 gpm depending on certain manufacturxng and assembly tolerances. Estimated temperatures in this region are presented in Reference 9. Miscellaneous Méasurements The overall head loss through the core vessel from the 5 in; inlet line to the 5 in. outlet line was measured and is shown in Figure 15. The curve has & slope of 2.0 indicating that the head loss is primarily due to. form drag as expected, and very little due to skin friction. " wy wr DISTANCE FROM WALL {in.) DISTANCE FROM WALL (in.) 2.5 2.0 1.5 1.0 . 0.5 2.5 2.0 1.5 1.0 0.5 27 ORNL-DWG 70-11791 OUTLET PIPE UH1 PROBE LOCATIONS MODERATOR CORE BLOCKS I ] | | MEAN VELOCITY BETWEEN UH1 AND UH2 § I Y | 300 gpm ///500 gpm /1 1200 gpm 75 Ny 5 i/ | Ss/ LA §/ !/ - I/ 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 MEAN VELOCITY (ft/sec) — T | | | MEAN VELOCITY BETWEEN UH2 AND UH3 N | =-0.79 \ 300 gpm " N6oogpm | 1200 gpm - 5\\ : — i N T (GUSTY FLOW) 1\, 1 NN ¢ L Y | \\- b S )\ { . | l 0.2 - 0.3 0.4 . 0.5 - 0.6 - 0.7 © MEAN VELOCITY (ft/sec) FIGURE,lh. FLUID VELOCITY PROFILES IN REACTOR VESSEL UPPER HEAD FLUID FRICTIONAL HEAD LOSS (ft-fluid) 20 S 28 v ORNL -DWG 64-6720A1 ~ 100 200 500 {000 2000 5000 10,000 FLOW RATE (gpm) FIGURE 15. OVERALL FLUID HEAD LOSS ACROSS MSRE CORE FROM 5 INCH INLET PIPE TO 5 INCH OUTLET PIPE @) . 29 Experiments were conducted to obtain information on the behavior of heavy solid particles in the fuel salt. The question arose because if sufficient oxygen or water vapor come in contact with fuel salt, zir- conium oxide would precipitete. A considerable effort was made in the MSRE to preveht this, and indeed it never did happen. Nevertheless, during the design etage it wes desirable to know the disposition of heavy solid particleS'shOuld'they occur. One of the most likely places for particles to settle out is in the lower head of the reactor. This is also one of the more critical areas because they could potenfially develop into a siéhificant heat source on the primary containmeht'surface. There was also concern that if enough partlcles developed they may plug up the drain line conflguratlon. , The experiments consisted of adding 2 or 3 pounds of sized iron filings to the water at the core inlet while the pump was running. They were added slowly over a specified period of time, generally about 20 minutes. The pump was then turned off either immediately’efter the addition or left running for an additional period up to 8 hours. After the pumpwasturhed off, the loop wes drained and the quantity of solids that re- mained in the.reactor vessel lower head was determined. The solids passed throughrthe core only once because downstream of the core model was the 5400 gallon surge tank in which the particles would most certainly settle out. All dats were taken at 1200 gpm, mostly with water but some rums with thickening agent added for Reynolds Number simulation. A summary of the data is shown- in,Flgure l6 A partlcle diemeter of 200-300 microns seems to be'critical that is;-partlcles larger than this settled out in 'signlficant quantitles, and partlcles smaller than this passed through the '}system. The gusty nature of the flow around the drain line assembly also - seems to ‘heve some ablllty to re-suspend particles less than 300 microns -in size if the pump is kept runnlng after the addition phase. The solids "that diad depos1t in the lower head had the appearance of an annuler ring around the drain pipe but did not touch it. In no case was significant _ . quantities of solids found in the 1 1/2 in. drain pipe. Because of the . way the model was_built,_itjwaS-nOt possible to check the 1/2 in. emergency drain tube after each run. After all runs were completed, however, this tube was examined and no partieles were found in it. After all the runs 30 TIME NOMINAL INTERNAL PUMP RUN MESH | LOOP USED TIME AFTER LOOP SIZE OF SCREEN LIQUID TO ADD SOLIDS WERE FLOW IRON APERTURE VISCOSITY SOLIDS = ADDED RATE FILINGS {microns) (1b/ft hr) (min) - (hr) {gpm) 100 147 2.0 120 0 - 1180 40 370 2.0 120 0 1180 20 833 2.0 120 0 1180 150 104 2.0 35 0 1180 150 104 2.0 32 8 1180 60 248 2.0 20 0 1180 60 248 2.0 20 8 1180 150 104 2.0 20 0 1180 50-30 290-490 9 1/2* 20 0 1180 50-30 290-430 9 1/2* 20 8 1180 50-100 290-147 9 1/2* 20 0 1180 50-100 290-147 9 1/2* 20 7 1180 WT SOLIDS ADDED (1b) w N NN NN W W W Ww w w Ww “THICKENING AGENT ADDED TO WATER FOR REYNOLDS NUMBER SIMULATION FIGURE 16. SUMMARY OF SOLIDS ADDITION DATA % OF ADDED SOLIDS THAT . WERE RECOVERED FROM VESSEL LOWER LEAD (%) ~0% 45% 552 9% 0% 612 55% 5.6% 562 243 30% 0% T U “r » nt 31 were completed, the top head was rembved from the vessel and it was found.that & large amount of filings had settled fairly uniformly on the top of the core blocks. 'They had rusted togethér, otherwise, a large fraétion of them may have fallen back down into the fuel channels when the pump wes stoppéd.-Note that the top of the core blocks is in the shape of a pyramid (Figure 2). Calculations have been made for the MSRE to determine the effect of precipitated fuel particles on the con- tainment wall tempersture near the drain pipe. They are not serious and some of the results are-fePOrted in Reference 9. | | Experiments were performed to determine the behavior of gas bubbles in the core. Bubbles between 1/8.and 1/4% in. were injecfied into the core inlet. At 1200 gpm virtualiy'all,the bubbles went down the core wall coolihg annulus and essentially none short circuited through the slots in the core support flange to the uPper,head. In contrast to this, at about 350 gpm, all the bubbles went through the core support flange slots and none went.down the core wall cooling annulus. The bubbles injected at this lower flow rate were considefably larger, however, than at the higher flow rate. At 1200 gpm, no pockets could be noted anywhere in the core vessel where bubbles tended to accumulate. ‘In the MSRE, bubbles in the fuel loop that have been generated by the xenon stripper spray ring are estimated to be in the order of 0.010 in. Certainly they will travel with the salt at 1200 gpm afid not accumulate in pockets or short circuit through the core support flange. - EXPEEIENCE WITH THE MSRE . A brief summary of the MSRE operatlng hlstory is a8 follows: | - First crltlcality - June 1, 1965 Nuclear operatlon termlnaxed - Dec, 12, 1969 ~ Time reactor was crltlcal - 17,655 hours Nuclear heat productlon - 13 172 equlvalent full power hours 'ISalt c1rculat1ng in fuel loop - 21,788 hours . 32 There were no direct fluid measuring probes of any kind (veloecity, pressure drop, etec.) installed in the MSRE core. Even the flow rate was - an inferred quantity from measured AT's in the fuel and coolant loop and the coolant salt flow rate from a venturi. The only instrumentation that could yield fluid dynemic performence information in the core was 64 thermocouples spotted at various locations on the outside of the core vessel (not in wells). These thermocouples yielded no indications of any adverse conditions inside the core. Another rather indirect source of information would be by the reactivity balance, which was kept in detail for this reactor because of its experimental nature. Again, no effects were noted that could be attributed to fluid.dynamics in the core. The conclusion then is that, based on lack of evidence to the contrary, the MSRE core did behave as was expected from testing the model, It should be noted however, that because of the conservative design of the core, a rather large malfunction would have been necessary in order to see it with available instrumentation. REFERENCES Robertson, R. C., MSRE Design and Operations Report, Part I, Descrip- tion of the Reactor Design, USAEC Report ORNL-TM-T28, January 1965. Entire Issue of Nuclear Applications and Technology, Vol. 8, No. 2, - February 1970. Molten-Salt Reactor Program Semiannual Progress‘Report for Period Ending July 31, 196L, USAEC Report ORNL-3708, November 196L.. A adyev, I. T., Experimental Determination of Local and Mean Coefficients of Heat Transfer for Turbulent Flow in Pipes, Tech. Memo 1356, NACA, Sparrow, E. M., et al., Turbulent Heat Transfer in a Thermal Entrance Region of a Pipe “with Uniform Heat Flux, Appl. Sci. Res., Sec. A, Vol. T. Abbrecht, P. H., The Thermal Entrance Region in Fully Developed - Turbulent Flow, AIChE Journal, Vol. 6, No. 2, June 1960. Lindeuer, R. B., Revisions to MSRE Design Data Sheets, Issue No. 9, ORNL Publication ORNL-CF-6L4-6-L43, June 24, 196L4, Internal Distribution Only. Engel, J. R. and Haubenreich, P. N., Temperatures in the MSRE Core During Steady-State Power Operation, USAEC Report ORNL-TM-378, November 5, 1962. 4 w} o 10. 33 Bettis, E. S., et al., MSRE Component Design Report, ORNL Report 'MSR-61-67, June 20, 1961, Internal Distribution Only. Mott, J. E., Hydrodynamic end Heat Transfer Tests of a Full Scale Re-Entrant Core, ORNL Publication ORNL-CF-58-8-5L4, Aug. 8, 1958, Internal Distribution Only. ~ » - [N ok 13. 15, 17, 18. 19. 20. 21. 22. 23. 2k, 25, . 26, 27.. 28. 29. 30. 31. P. 33. 3. 35. - 6. 37. 3. 3. k0. u1L e 3. oLl hs =48, h9. ~_'50 51. 52 53. 5L, 55. we»mmzzwflzwm?mswwbsozgyuq MSRP Dlrector s Offlce ) LL Anderson H. F. Bauman S. E., Beall H. R. Beatty M. J. Bell M. Bender E. S. Bettis R. E. Blanco F. F. Blankenship R. Blumberg . G. Bohlmenn ‘I. Bowers Boyd Briggs Burke Cardwell Carter. Caton Collins - Coock Cooke Crowley Culler Eatherly Engel Ferguson Fraas Frye . Furlong ‘Gabbard Grimes = Grindell Guymon Harley ‘Harms . fl”-fl"—!c-lfloc-l:&'UO:UE.'):.:fltl:l Helms Hoffman - Holz Huntley Kasten Kedl ‘Kelley - Kelly - Keyes _ Kirslis Krskoviak Kress . Lindauver L . » . - .- . e - - . - - . [ ] * - » - . WOHNGLUHLUY YO MmN NN NI T S NS M S s W ‘Haubenreich 35 s 58. - 29. 60. 61. 62. 63. 6L, 65. 66. 67, 68. 69. T0. T1. T2. T3.-T6. TT. T8. T9. . 80. 8l1. 82. 83. 8L, 85. 86. 87. 88. 89. 90. . 91- 92, - 93. -9k, 95. 196.-97. 198.-99. - 100.-102. 103. tlj'bU‘C-imQU_.':flO'JU:.fiBU EWQEQPFPWQ?mHO?ZQUngflb - INTERNAL DISTRIBUTION Lundin Lyon MacPherson MacPherson Martin McCoy _ McElroy McGlothlan McLain McWherter . Moore Nicholson _ Perry W. Pickel Richardson C. Robertson . P. Sanders unlap Scott L. Scott J. Skinner N. Smith L. Smith Spiewak . H. Stome Stulting Tallackson . Thoma, Trauger Watts Weinberg . Weir . Whatley White Wichner Wilson Gale Young H. C. Young . * - - * .z ' Ho = » - 2R EY P s . a8 - o 4"601?»1%3?‘&1?150 Central Research Library Document Reference Section Laboratory Records | Leboratory Records (RC) '10L.-105. 106. 107. 108. - 109.-110. 111. 112, 113. 11k, 115. 116. 117.-131. 36 EXTERNAL DISTRIBUTION D. F. Cope, AEC-ORO David Elias, AEC~Washington Kermit Laughon, AEC-OSR C. L. Matthews, AEC-OSR T. W. McIntosh, AEC-Washington D. R. Riley, AEC-Washirigton H. M. Roth, AEC-ORO - Jd. J. Shreiber, AEC-Washington R. M. Scroggins, AEC-Washington M. Shaw, AEC-Washington M., J. Whltman AEC-Washington Division of Technical Information Exten51on (DTIE)