.:”} = Y ». =) ORNL-TM-2987 Contract No. W-T4O5-eng-26 Reactor Division DEVELOPMENT OF FUEL- AND COOLANT-SALT CENTRIFUGAL PUMPS FOR THE MOLTEN~-SALT REACTOR EXPERIMENT P. G. Smith LEGAL NOTICE Thls 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- pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. ~ OCTOBER 1970 . OAK RIDGE NATIONAL LABORATORY ) Ok Ridge, Tennessee - operated by UNION CARBIDE CORPORATION for the U S. ATOMIC ENERGY COMMISSION SISTRIGUTION OF THIS DOCUMENT IS “ . 111 CONTENTS -~ o f | - | Page Abstract S 0 9 S 5 0 S8 P LB ISP I T O PP AP P RO N PSS LSS NI ESNES SN S l GeHEral Description Of_the MOlten-S< Pump 5 6 40 990 S Qe LR ST EER OSSO BSOS 2 Test Apparatus | Molten-Salt Pump Test Stahd secescescasssssssssecsencrssoss 5 Molten-Salt Properties ........;........................... Bench Tests and Cold Shakedown Tests _ Force-Deflection Characteristics of Shaft ..ccevcecesenasee 9 Critical Speed of Shaft Assembly .eeecocecceecencencncnsene 9 Room-Temperature Dry.Runs teessarsessasesessscasscssesse s 11 Molten-Salt TesSts cececicesocccrrsrsscccnnsensscsssssosonnsesscaess 11 Hydraulic Performance ..ccceceeceess teseeracrctrecensnanes 11 Cavitation PErfOrMANCE «eeeeeeesseesscossecesssnacennnesnes 1k % Effectiveness of Shaft Annulus Purge Against Back Diffusion ’k'r ofRadiOaCtive Gas .l."-l..............ll...q....Il....’.. lh < | Measurement of Undissolved Gas Content in Circulating Salt 17 Problems Encountered During Pump Fabrication and MOlten-Salt TEStS ® 0 2SO N PO PRI RIS RSLENEB SN Pe LS e 8 2 e 0w 21 Fabrication Problems .ceeeceecseccsscccssssosascassscssosss 22 Shaft Annulus Plugging seceecevescnssocssscasosossscsscssnsse 22 Insufficient Running ClEBIBNCES «evsesrssscssoancsccanossss ol 0il Leakage from the Catch Basin into the Pump Tank ....... 26 Pump Tank Off-Gas Line PLUEEINE wevevevevennessonnennaneass 28 Failure of We1d'Attachmén£of'Parts‘cf Flow-Straightening DEVICE. coevessonssocesesanssssasvocsosssccsccsoscnnasocsns 28 © Mark-2 Fuel-S8lt PUID euvevseneruesnenseneanesnasnsancssesessess 33 | *Déécri?tion of Pump...;}..;.;;;..;..;....,.....;;....;...., 33 Hydraulic PETTOTIIANICE s v e e vvnsasossncossenceneeanaonasaesos 36 Measurement of Undissolved Ges Content in Circulating. -' Salt o-uon-aoo;-ooJ--o-i¢Q--¢oooooooooo-.oooo}do(-;uoiob.i 36 , " Restrictions to Purge-Gas Flow ceereeienteecentsttertiaenes * 36 B ) . Performance Of MOlten-SBlt PumpS in MSRE'o.oooaoooocouoc-;;oouoo 38 *) i z . . ConCluSionS ....O...O..l‘....I..IICOCO...'...0..!'..l‘.....l.l.'.. ho Acknowledgments ......'.QOO.'....Ill.'l.l...l...'....0'!0!00'.0'0 ho References .....‘l....‘.j'.l..'...Q.l..‘l._l......ll...‘l:..l‘l‘OO APPendix-MSRE DraWings .....‘...'..'...C.VOOOO'.l....‘l..ll.l._c. 41 43 " ) a) “3 DEVELOPMENT OF FUEL- AND COOLANT-SALT CENTRIFUGAL PUMPS FOR THE MOLTEN-SALT REACTOR EXPERIMENT P. G. Smith Abstract The Molten-Salt Reactor Experiment (MSRE), a small nu- clear power reactor that produced about 7 Mw of heat while operating at approximately 1225°F and atmospheric pressure, requires & pump in each of two circulating molten-salt sys- tems. A vertical centrifugal sump-type pump was developed for each system through water and molten-salt tests of pro- totype pumps at temperatures to 1LOO°F and hot shakedown operation of the actual reactor pumps before they were quali- fied for reactor service. The development experience with the pumps in the molten-salt pump test stand and the perfor- mance of the reactor pumps in the MSRE are discussed here. The hydraulic performance of the pumps circulating molten salt corresponded closely with the performance obtained with water. The reactor pumps served well throughout the approxi- mately 30,000-hr operating life of the MSRE, which spanned the period August 1964 to.December 1969, during which the reactor produced 105,737 Mwhr(t) of nuclear energy. A back- up pump (Mark-2), which contains additional volume in the pump tank to accommodate thermally expanded salt, was also fabricated and tested for the MSRE. Keywords: pump, molten salt, Molten-Salt Reactor Ex- periment, centrifugal pump, sump-type pump, high temperature, nuclear reactors, pump hydraulic performance, water test puup. A centrlfugal pump was developed for c1rculating molten salt at ele- '_ vated temperatures in the Molten-Salt Reactor Experlment (MSRE) 1 Briefly, -_the MSRE 1s & salt- fueled graphite-moderated single—region nuclear reactor test facility with a heategenerstion rate of approximately 7 Mw(t). Two salt pumps are required,rone'in the fuel-salt loop and the other in the *coolant-salt loop.- Since the designs of the two pumps are essentially -ridentical, prlmary attentlon is. given here to the fuel-salt pump. ‘The pump is a vertical-shaft sump pump with an overhung impeller and an oil-lubricated face seal. It was developed through a series of bench tests, water tests,® cold shakedown operations, and high-temperature molten-salt tests. The problems encountered during development and testing and the results of pump'operatibn at afibient and elevated tem- peratures (up to 1400°F) are discussed in this report. Hydraulic per- formance data, priming conditions, and coastdown characteristics were obtained, and the effectiveness of spray devices for xenon removal was demonstrated. Thermal-stress and strain-fatigue analysesa-were made for the pump tanks of both the fuel- and coolant-salt pumps. They were made for an estimated operating history that included 100 heating cycles from room temperature to 1200°F and 500 reactor power change cyéles‘frdm zero to full power. The calculations indicated that = cooling air flow rate of 200 cfm was required for the fuel pump tank, while the coolant pump tank vas capable of the required service without air cooling. _ The pattern followed in the development of these pumps and the de- sign of a similar pump were discussed elsewhere in some detail.* The problems and tests required for developing other specific elevated- temperature pumps have been reportedJ5-4° ‘ - Fuel- and coolant-salt pumps were instaelled in the appropriate salt circuits of the MSRE, where they each circulated salt or helium at ele- vated temperatures for & total of approximately 30,000 hr. The reactor was operated up to full power and was recently shut dowm permanently. During operation of the reactor, which produced 105,737 Mwhr(t) of mu- clear energy, the lubricants for the bearings and seals and the insu- lation for the drive motor were exposed to a nficlear radiation environ- ment., | ~ The numbers of the drawings for the fuel- and coolant-salt pumps, the drive motors, the lubrication stand, and the Mark-2 fuel-salt pump are listed in the Appendix. | ) | General Description of the Molten-Salt Pump The pump is of the centrifugal sump type with a vertical shaft. It consists of three main components: the pump tank, the rotary assembly, and the drive motor (see Fig. 1). The three main components are bolted ~ £ n ORNL-LR-DWG-36043-8R2 2 INNNY %] 2 SHAFT WATER COUPLI COOLED : ' ' ‘ MOTOR SHAFT SEAL , | ‘ (See Inset) = fiF——=mgoo—mie—-—o————rf g — - | ) i ¥ i | | e — N\ ' " LEAK DETE LUBE OiL IN- BE OIL. BREATHER BALL BEARINGS' (Face to Face) BEARING HOUSING " BALL BEARINGS GAS PURGE IN | . BASIN (Back to Back ) ' SHAFT SEAL (See Inset) SHIELD COOLANT PASSAGES (In Parafiel With Lube Oil) SHiELD PLUG GAS PURGE OUT (See inset) LUBE OIL OUT SEAL OiL LEAKAGE DRAIN LEAK DETECTOR SAMPLER ENRICHER (Out of Section) {See !nsefl GAS FILLED EXPANSION SPACE STRIPPER ~T BUBBLE TYPE {Spray Ring) “ LEVEL INDICATOR SPRAY - OPERATING LEVEL ST R TR TSN Ti) To Overflow Tfink Fig. 1. Cross Section of'Fuel-Salt Pump. together and sealed with oval ring-joint gasketed flanges. The ring- joint grooves are connected to a lesk-detection system. The motor and rotary assembly may be removed from the pump tank either as a unit or separately. All parts in contact with molten salt are constructed ofA Hastelloy N,* a nickel-molybdenum-chromium-iron alloy.. The pump tank, which provides volume to accommodate. the thermally expanded salt of the system, contains the pump volute or ca51ng, the Xenon-removal spray device, the salt level indicators, and various access nozzles. | The rotary assembly consists principally of the bearing housing; the pump shaft, which is mounted on commercially available conventional ball bearings; the shaft seals, which constrain the circulating lubricating oil from leaking out of the bearing housing; the shield plug, which is cooled with circulating oil; and the pump impeller. The shield plug and other parts of the rotary assembly that come in contact with the salt are suspended in the pump tank through a large flanged nozzle at the top of the tank. _ | The drive, which is housed in arhermeticaliy §ealed vessel, is a squirrel-cage induction-type motor ratéd for 75-hp.duty at 1200 rpm. The electrical inéulatidn system is resistant to a radiation dose of up to 10° rads and the grease lubricant is reported!! to be capable of with- standing a dose of more than 3 X 10° rads. The pump has some unusual features required by its application that .are not found in the conventional sump pump. The pump tank contains a salt-spraying device to remove 1335Xe (neutron absorber) from the circu- lating fuel salt and also has two gas-bubble sensors to indicate salt level. The spray device is connected to the volute discharge, frbm which it receives a proportioned flow of about 50 gpm of salt at pump‘design head and flow. This flow and other leakage fldws (bypésg flow) pass through the pump tank and return to the system at the pump‘inlet._.A split purge-gas flow in the shaft annulus keeps oil vapors from énfering the salt system and fission gases from entering the region of the shaft ¥Hastelloy N, known also as INOR-8 and Allvac N, has the basic com- position, by weight, 15-18% Mo, 6-8% Cr, 5% Fe (max), 0.4-0.8% C, balance Ni. il xy 3 2 lower seal. The down-the-shaft portlon of the purge-gas flow removes 136%e by contact w1th the salt spray and dilutes and transports 1t from the pump tank to the off—gas system. The pump is mounted on a flex1ble support de51gned to accommodate thermal expansions. A flow of nltrogen across the exterior of the upper nonwetted portion of the pump tank re- moves nuclear heat dep051ted 1n the tank wall The bolt exten31ons, which can be seen in Fig. 2 prov1de for remote installation and removal of the drive motor and rotary assembly in the MSRE.r Test Apparatus Molten-Salt Pump Test Stand A schematlc view of the ‘test stand components is shown in Fig. 3. The components include ‘the drive motor, the test pump, and the salt piping, which is 6-in. IPS sched 40, except for a section of plpe at the pump inlet, which is 8-in. IPS sched 4LO. Other components include a venturi,flowmeter, the heat removal system, a drain tank for salt storage, the:preheating system, a flow straightener, high-temperature pressure and temperature senso’rs,"’the‘lubrication:system,12 and two salt freeze flanges, one of which provides a place to mount an orifice Plate torset the_system'resistancefto saltiflow, The system resistance was varied with four ,orii‘ice plates :having di_ff__erenthole diameters. The heat re- moval system,'which is arsalt;to;air heat exchanger, was used to control the salt temperature., The drain tank stored the fluoride salt in the molten state when the system was ‘not in operatiOn.' Commer01al diaphragm- 'rsealed NaK-filled pressure transmitters were used to indicate pressure | ‘at the pump discharge and at the 1nlet and throat of the venturi. Trans- | former controlled heaters were used,to preheat the salt plplng and com- ponents, and Chromel—Alumel thermocouples monitored the system tempera- ,rtures. Conventional 1nstrumentatlon was used to display and record tem— - peratures, salt flow, pump drive motor power, and salt level 1n the pump tank. A photograph of the. test facillty, Fig. b, shows the pump 1n the left upper foreground a portion of the salt piping beneath and to the rear of the pump, and a portion of the control cabinets. PHOTO T0901 a7 s R L4 and Bolting Rotary Assembly, Fuel-Salt Pump Drive Motor, Fig. 2. for Remote Maintenance. 4] o b ORNL-LR-DWG 72414 _THERMAL | | CWELL - - - AIR COOLER o IM - DISCHARGE ", | _1'— PRESSURE MSRE. _ ' PIPE HEATER : FREEZE -T "; ~—8-in. PIPE o - . FLANGE 4 6-in. PIPE ' I ‘ — -’ | | II--ll - . ‘ FLOW STRAIGHTENER INLET °~ FREEZE VALVE | . l MSRE FREEZE PRESSURE \ , - - FLANGE THROAT ||~ . : | PRESSURE— - MAND VALVE - DUMP TANK Fig; 3. Schematic Diagram of Molten-Salt Pump Test ‘Stand. PHOTO 70988 Fig. 4. Photograph of Molten~-Salt Pump Test Stand. W 'd n ) Iyl sk A Molten-Salt Pr0perties The molten salt used for pump tests was a mixture of lithium fluoride (LiF), beryllium fluoride (BeF,), zirconium fluoride (ZrF,), thorium fluo- ride (ThF,), and uranium fluoride (UF,) in proportions of 66.4k, 27.4k, 4.7, 0.9, and 0.7 mole %, respectively. The mixture is solid at room tempera- ture and melts at approximately 850°F. The densityl?® and viscosityl4 at three temperatures of interest are given below: Temperature Density Viscosity - (°F) (1 /ft2) (cps) 1100 - 132 11 1200 130 8 1300 129 ' 6 Bench Tests and Cold Shakedown Tests Force—Deflection‘Characteristics of Shaft The force-defléction characteristics of.the shaft were measured with the shaft assembly supported in the bearing housing and with the force applied at the .impeller. A typical curve of shaft deflection at the impeller versus force is shown in Fig. 5. This information and deflection data obtained during water tests were used to determine the . unbalanced force-vector acting on the 1mpeller at various head, flow,i - and speed operating condltions 18 The force values were used to analyze shaft bendlng stresses and bearing reactions and to specify the shaft 'support bearings. Crltical Speed of Shaft Assembly The crltlcal speed of each shaft assembly was determlned by V1brating rthe shaft assembly in the transverse direction.. The_shaft assembly 'was - supported in the bearing hous1ng and mounted vertically on a rugged-steel structure anchored to the building floor. The vibrating force was applied at the impeller and held constant over a range of applied frequencies. of Pump. 10 (X 10-3) ORNL-DWG 70-6823 40 | 3" < e 20 SHAFT DEFLECTION (in.) 10 0 100 200 300 400 500 FORCE (1b) Fig. 5. Shaft Deflection Versus Radial Force Applied to Impeller ¢ 1 ¥} N X} 11 Data were obtained for frequency versus vibration amplitude. The critical speed determined by the frequency at which the amplitude increased ‘sharply checked quite closely with the calculated value. TFor the fuel-salt pump the value was calculated to be 2850 rpm and was measured to be within 100 rpm of this value. Room-Temperature Dry Runs After assembly & pump rotary element is normally operated for about one week or longer in a cold shakedown stand before installation in the hot test stand; This operation-is conducted to verify that the element is free of mechanical problems and to test the performance of the shaft bearings and seals. Theuoil-leakage from a properly performing shaft seal is usually 10 cc/day or less. Molten-Salt Tests Tests with molten salt were conducted with the prototype and re- actor pumps to (1) verify the hydraulic performance observed in the water tests, (2) determine the effectiveness of the gas purge down the shaft annulus against the intrusion of radioactive gas, (3) measure the concentration of undissolved_gas in the circulating salt, (4) perform acceptance tests of the reactor pumps prior to their installation into the reactor system, ‘and (5) test the overall long-term relisbility of the pump and drive motor atudesign and off-design conditions. Table 1 presents & summary of the fiolten-aalt test operatioh of'the'prototype pump, the rotary elements for the reactor pumps, and the Mark-E fuel- ”:“-salt pump. Hydraulic Performance. Hydraulic performance'tests were conducted with 1l3-in.~ and 11 1/2-in.- OD fuel punp impellers in the molten-salt pump test stand with the fuel- salt pump tank and volute installed. ‘The head- capacity performance of the 13-in.~-0D impeller with both water and molten salt is shown in Fig. 6, in which the head is plotted against flow for three test speeds. At 1030 rpm, 12 Mark-2 pump < : ‘ Poble 1. Sumary of Tests of MSRE Pumps in _the Molten-Salt Pump !l'estVStand Tost -, Molten-Balt Puvp Shaft © Molten-5alt Impeller Test : . Reason o Temperature Speed Flow Diameter Duration Primery Purposes of Test Tor ‘ (' (rpm) (ewm) (in.) (br) . Termination 1 80-1200 1150 13 118 Ghaxedown test of prototype fuel- galt pimp e 1l 1200 7002030 T50-1600 13 21T Hydraulic performance test of Shaft seizure prototype fuel-salt pump T - 2 1200 700-2030 T50-1200 13 96 Same a3 above Bcheduled 3 1200 600-1150 R0O-1500 11 1/2 1,968 Bame as above Variable frequency ‘ ’ 7 motor-generator set fallure b 1200 185 1100 1 1/2 1,848 Back-diffusicn teste of prototype Variable frequency - fuel-galt pump motor-generator set fallure 5 12001320 1185 1100 1 1/2 792 Gas-concentration tests in circu- Scheduled : ) lating salt and back-diffusion tests of prototype fuel-salt pump 6 1200 1150 1070 u 1/ 335 " Gas-concentration tests of proto- type fuel-salt pump in circu- lating salt ’ 6 1000-140C 600-1150 S00=-10T0 11 1/2 368 Hydrsulic performance of prototype Shaft annulus plugged : tuel-sglt amp T 1200 1150 1070 1 1/2 120 Proof test of lubrication stand - and fuel pump supports 7 1100-1300 T00-1150 600-1070 1n 1/2 409 Scheduled 8 11001300 1159 1070 1 1f2 168 Proof test of coolant pump lubri- Scheduled . cation stand and fuel pump sup- ports g 120 o800 11 1/2 Reactor fuel pump hot shakedown Shaft rubbed at startup test 10 1200 1750 750 10 1/3 90 Reactor coolant pump hot shake- Scheduled down test n 1200 u7s 1200 1 1f2 100 Reactor fuel pump hot shakedown Scheduled test 12 12004515 175 1200 13 452 Reactor spare fuel pump hot shake- Impeller rubbed volute . down test 13 1200 s 540 10 19/32 1,000 Reactor spare cooclant pump hot Scheduled 7 shakedown test 14 1200 uT7s 1200 13 2,654 Reactor spare fuel pump hot shake-~ Loop flw straightening down, back-diffusion, gas concen- vanes became detached tration tests ' 15 1200 1175 1200 13 155 Reactor spare fuel-punmp impeller Bcheduled shakedown tests 16 1200 1175 1200 13 166 Reactor spare coolant drive motor Scheduled ‘shakedown tests 17 1200 175 1200 nifz 2,631 Prototype fuel pump test Scheduled 18 1200 u7rs 1200 nifz 100 Reactor spare fuel pump hot shake- Scheduled i down tests 19 1200 1750 90 10 19/32 100 Reactor spare coolent pump hot Scheduled shakedown tests . 20 1000-1325 17 1350 1 if2 1k,000 Perfarmance and endurance tests of Continuing n 1) 13 ORNL-DWG 65-6666 H 1 60 . \ ® ® . \ .! 50 \ {030 rpm 40 g = : _ ™~ 860 rpm S = 30 tw I. * \.\ . - >~ e x 1\ 700 rpm | | 20 ALLIS-CHALMcRS MFG, CO. IMPELLER AND VOLUTE DESIGN SIZE 8x6; TYPE E; IMPELLER P482, 13-in. OD ORNL WATER PERFORMANCE 10 DATA POINTS: ORNL MOLTEN SALT PERFORMANCE AT i200°F o - - e , — 0 200 400 600. 800 1000 1200 1400 1600 1800 - @, FLOW (gal/min} ‘F:Lg. 6. Hydraullc Perf‘ormance of Prototype Fuel—-Salt Punmp with Water and with Molten Salt. i o o 1k the molten-salt head veries as much as 1 1/2 £t from that of water, and at 860 and 700 rpm the head for the molten salt is low by approximately 1 ft. Cavitation Performance The NPSH requirements for the pumps, as reported by the manufacturer of the hydraulic components (volutes and impellers), are 5.5 and 11 ft, respectively, for the fuel- and coolant-salt pumps at their operating. conditions. Converted to pressure of the circulating salt the NPSH values aré 5 and 10 psia, respectively. These pressures are below atmospheric, and evacuation of the pump tank gas space would be required to investigate the suction pressure at which cavitation sets in. Since1¢vacuation is required to cause cavitation and the operating pressuré of the pump'tanks in the MSRE is 5 psig, experimental investigations of cavitation inception ‘were considered unnecessary and therefore were not performed. Effectiveness of Shaft Annulus Purge Against Back Diffusion of Radioactive Gas ' The migration of radioactive gases from the pump tank to the region of the shaft lower seal via the shaft annulus could result in polymeriza- tion of the oil that leaks past the seal and lead to plugging of the drain line from the oil catch basin. To minimize this migration, a flow of helium purge gas was introduced down the shaft annulus. Figure 7 is a schematic diagram of the shaft annulus configuration and the purge-gas flow paths. The upper end of the flow path for the down-the-shaft purge contains a lsbyrinth seal (not shown), which has a 0.005-in. diametral clearance. | In back-diffusion tests, 85Kr and nitrogen were injected-into the pump tank as shown. The concentrations of 85Kr in the pump tank off-gas line and in the line from the leakage-oil catch basin were determined with count-rate meters for variocus flow rates of purge gas. The data shown in Table 2 were obtained during the tests. The 8F5Kr was not de- tected in the purge gas from the catch basin even with a count rate -10 meter capable of detecting a concentration as small as 0.95 X 10 Ci/cnfi. 15 \\\\\\\\\\\\\\\\\\\\A Tz 0 .rfl ( =< - : . 2T ‘ awn . Fig. 7, fSchem&tic,Diégiam of_Purge,Ggs:FlowiPassages'in Fuel-Salt 16 These date indicate the capability of the purge gas at flow rates of approximately 0.25 liter/min to reduce the concentration of 85Kr in the catch basin by a factor as much as 35,000 times below the concentration in the pump tank. Table 2. Back-Diffusion Test Results. - Diametral clearance; 0.005 in. Shaft Purge = Catch Basin Purgea 85Kr Concentration'in Pump Tank (Liters/day) (liters/day) (Ci/cmB). | " - X110 406 - k87 0.89 406 | L87 - 0.89 %30 497 ' 2.06 382 481 2.77 366 | 351 | 3.28 215 360 1.85 108 - 360 1.21 173 673 1.65 109 360 ‘ 0.75 109 681 0.90 10k | 681 3.5k o | 85Kr concentration in catch basin was less than 0.95 X 1071° gt g1l purge flows. The meximum permissible concentration of radioactive gases in the catch basin to avoid polymerization of the leakage oil was calculated to be 3.1 x 10~ Ci/en®. Calculations based on this limitation and the data of Tsble 2 indicate & maximum permissible concentration or radio- active gases of 2.4 to 11.0 Ci/cn® in the pump tank. The results of additional calculations relating the flow rate of pfirge gas in the down- the-shaft annulus to the concentration of radioactive gases in the pump 17 tank for a conservatlve assumption of the MSRE nuclear operating condi - tions are presented in Flg 8. This figure indicates that a flow rate of purge gas of less than 1 liter/mln suffices to protect the seal oil leakage in the catch basin'from polymerization by radiocactive gases. A satisfactory supply of purge gas was available at the MSRE. Additional back-diffusion tests were made with 8SKr injection in which the shaft annulus labyrinth seal clearance was 0.010 in., and the data obtained are given in Table 3. With a detector sensitivity of 1.5 X 10~ ? Ci/cnP 885Kr could only be detected in the catch basin at low down-the-shaft purge flow rates. ' ' Table 3. Back-Diffusion Test Results Diametral clearance: 0.010 in. Catch Basin Kr Concentration Shaft Purge " purge (ci/cmB) (1iters/day). (1iters/day) — : , - In Pump Tank In Catch Basin - . -8 . alo 212 173 - 2.66 x 10 <1.5 X 10 : 8 . -10 232 © 16k . 5.50 x 10 <1.5x 10 ' - , D -8 - - 12 105 h.2h x 10 <8.17 X 10 Measurement of Undlssolved Gas Content in Clrculating Salt f o Undesirable quantities of gas bubbles in the pump tank liquid can be entrained in the c1rculating salt by the return of salt that has '_leaked into the pump tank: from the high—pressure side of the pump 1 The high velocity of these salt leaks from the internal spray and other '_ - sources within the pump tank can carry gas under the salt surface. The m:downward velocity of the returning salt may then carry the bubbles to | the impeller inlet and on into the circulating salt. ' To obtain insight into this the concentration of undissolved gas in the c1rculat1ng salt was measured by.u51ng gamma, radiation densitometry. 18 ORNL-DWG €3-6483 20 - T | = | 10 Mw POWER | E * 100 % STRIPPING EFFICIENCY 216 50 gpm BYPASS FLOW ] T \ 5 psig PUMP TANK PRESSURE L3 Z e 2 42 NG o * = \\\\\\\ = O Z 8 S g \Q s 4 » 0 TR 0 : | 0 1000 2000 3000 4000 5000 PUMP TANK PURGE (liters /day) Fig. 8. Concentration of Radiocactive Gas Versus Purge-Gas Flow Rate in Prototype Fuel-Salt Pump. » 1 A LO-Ci 187(Cs source of 0.622-Mev gamma rays was placed on one side of the salt piping at the pump inlet, a region of low static salt pressure, and a radiation detector was placed on the other side. The detector consisted of a cylinder of plastic phosphor (3 in. in diameter and 6 in. long) and an electron multiplier phototube. The plastic phosphor absorbs photons, which are transmitted from the source through the salt. Light is emitted by the plastic-phosphor to a phototube that produces a current. The current is a function of the photofi transmission through the Salt, which in turn is an inverse function of the density of the salt. The signal current from the detector is fed to a suppression circuit that indicates only variations in the current. The current indication is recorded on Visicorder tape. Thus an increase in the concentration of undissolved gas, which causes. a decrease in éalt density, is indicated by an increase in the éurrent outpfit of the densitometer. ‘ The densitometer was calibrated over a small rénge of thermal change in salt density correspondinéfto-themsalt temperature rangerlloo to 1400°F. During calibration the salt contained no entrained gas. The pump was not bperating,_sbfgas bubbles could leave the densitometer and enter the higher elevations_in‘the salt system. Thermal-convection velocities in the salt were too small to entrain fresh gaé from the pump tank. | | o The two Visicorder traces shown in Fig. 9 give a measure of helium concentration in the salt loop at two different salt levels in the pump tank. The pump vas operated with a 13-in.-OD impeller at & flow of 1650 gpm through the 6-in,-diam pipe and a salt temperature of 1185°F; approximately 85 gpm of salt was bypassed through the spray ring. The _curves are plotted with the flow density at zero deflection for compari- ‘son and to illustrate:the transient behavior of the gas concentration when the pump is first stopped (flow reduction to zero)'énd then started '_(flcw_increase frdm'01t011650 gpm); Trace I presents the density data for the normal level of salt in the pump tank (3 3/8 in. above center _line of the volute). Trace II relates to a higher salt level in the pump tank (4 7/16 in. sbove the center line of the volute). The gas concentration in the circulating salt when operating at the normal level was 4.6 vol %, and at the other higher level it was 1.7 vol %. The 20 ORNL-DWG 65-11798 10 o2 @ SILICA , 8- BEAM & FILTER Tz j 6= [ 46% n 5+5 I NORMAL PUMP = a BOWL LEVEL / 34 Sk 5 | © I UPPER PUMP o2 BOWL LEVEL | ui 1 i o -1 SALT TEMPERATURE: H85°F VOLUME SENSITIVITY: 0.64%/div i ke ek e Lt 0 20 40 60 80 {00 120 {40 160 - 480 200 TIME (sec) Fig. 9. Concentration of Undissolved Gas Versus Time as Determined by Gamma Radiation Densitometry of Prototype Fuel-Salt Pump with 13-in. Impeller and 1650-gpm Salt Flow. b ¥ » 21 gensitivity of the Visicorder wvas 0.64 vol % per division. At the higher operating level, the discharge ports of the xenon removal spray ring were covered W1th salt. Since the high-velocity spray from the xenon-removal spray ring is a magor contrlbutor to the agitation that causes gas bubbles, it is understandable that the gas concentration was lower when operating at the higher level in the pump tank. In s separate test, the same densitometer arrangement was used to measure gas concentrations in the fuel salt circuit at the MSRE. The principal conditions inclndedgthe impeller diameter of 11 l/e,in., a salt Tlow of 1150 gpmrat l200°F; normal operating level' and a spray-ring flow of 50 gpm. The v01d fraction in the salt was not detectable at these conditions. At & lower level’ of salt in the pump tank, a void fraction of 2 to 3 vol % was indicated Subsequent measurements with the same densitometer on the molten- salt test loop and with a 11 l/2-1n -diam impeller, a flow of 1200 gpm at 1200° F, the salt at the’ normal operating level, and a spray-ring flow of 50 gpm gave a void fraction of 0.1 vol %. Comparing the gas concentrations for operation with the 13-in.- and ‘the 11 l/2-1n.-diam 1mpellers at the normel operating level shows that the content is less with the smaller 1mpeller by a factor of 46. This is attributable to the xenon-removal spray flow, which is approximately 50 gpm for the smaller impeller, compared W1th 85 gpm w1th the 13-in.-diam impeller. The jet velocity from the spray ring is cons1derably less with the SO-gpm flow; thus there is less agltatlon of the liquid level surface upon impingement of the jet streams. Problems Encountered Durlng Pump Fabrication and ' Molten-Salt Tests _Fabrication{problems were”encountered with the dished heads for the ~ pump tanks, the1impellerLand'rolute_castings, and the hermetic vessels for enclosing the drive motors. Pump operating prdblems-were'encountered _with the shaft purge flow, shaft and impeller running clearances, ‘shaft seals, and the flow straightener in the test stand salt piping during testing and operation of the pumps at elevated temperatures. 22 " Fabrication Problems The Hastelloy N dished heads for the pump tanks were originally fab- _ricated by hot spinning of plate stock. The resulting heads contained 'many crack-like defects on the knuckle radius that were readily detectable by dye-penetrant inspection. However, the heads were repaired by removal of the cracks by grinding, and this rendered them ussble forjtesttpurposes. Later, during fabrication of the reactor pumps, the problem was solved by hot pressing the plate stock. Considerable effort and time were expended to obtain satisfactory Hastelloy N castings for the impellers and volutes. The initlal castings were unsatisfactory due to defects such as cracks‘resulting from shrinkage, inclusions, and porosity. Castings of improved quality, whieh vere accept- able after repairs, were obtained from a second foundry. The castings were improved by modifying the chemistry of the casting melts and by the founder giving close supervision and attention to the details of the casting pro- cedures. | ' Two problems were encountered during the fabrication of the hermeti- cally sealed vessels for the drive motors. The design”originally'specified brazing a cooling coil of stainless steel pipe to the outside of the cylin- drical carbon steel vessel. However, because of the difference.in thermal | expansion between the two materials, continuous attachment of the coil to the vessel was not achieved ny'brazing. The problem was solved by welding the coil in place. A more serious problem was encountered during the-build- up of weld metal, "buttering,” on the inner surface of the vessel to accom- modate the attachment of a flat head. Large laminar defects appeared in the vessel wall in the region of the buildup. Removal of the defects by grinding and weld repair resolved the problem. Satisfactory vessels were obtained but at the expense of delays in delivery of the vessels and additional expense to the fabricator. ' Shaft Annulus Plugging During test 6 (see Table 1) the recorded trace of drive motor power began fluctuating after 70O hr of operation with molten salt at 1200°F. An anomalous value of gas pressure, which exceeded the supply pressure » " " ¥ 23 by approximately 5 psi;'was,observed in the pump tank. The test was ~therefore terminated,%and inspection of the rotary assembly revealed the presence of solidified salt in the annulus between pump shaft and shield plug. L oy Subsequent analysis shoued‘that the gas flow in the lower portion of the shaft annulus had been reversed and was upward instead of down- ward, the proper direction. The flow reversed because the gas flow from the liguld level indicators, although thought to be small, was actually significant. The excess gas exceeded the throttled capacity of the-off-gas flow line from the pump tank; consequently the excess gas flowed upward, Jjoined the annulus purge gas, and left the pump through the shaft seal oil leakage drain line. Small droplets of molten salt were carried by the reversed purge flow into the lower end of the - shaft annulus where they solidified, accumulated, and finally acted as a brake on the shaft.‘ The chemical composition of a sample of the mate- rial'taken from the shaft annulus vas very close to that of the salt being circulated. The results of x-ray and petrographic examinations of the sample indicated that the material vas carried into the annulus as an aerosol. | - To prevent the recurrence of this incident and to provide protection against back diffusion, the flow rates of the gas to and from the pump tank were monitored more closely to assure that purge gas did flow down the'shaft annulus. The flow rate of gas from the pump tank must equal “the flow rate of the purge down the shaft annulus plus the flow rate of '7gas associated with the operation of the bubble—type level indicators. Another case of shaft annulus plugging was encountered during test | ;17 (see Table l) The plugging was indicated by a buildup of pressure in the seal oil leakage catch basin compared with the pump tank pressure. The pressure differential (0 to 5 p51) thus created caused 0il to lesk '"rout of the catch ba51n and down the outer surface of the shield plug and .dinto the pump tank.- This leakage was detected by a hydrocarbon analyzer sampllng the pump tank off-gas. The plug was located at the lower end of the shaft annulus, and it was of such a nature it could be temporarily removed by heating the system circulating salt from 1200 to 1250°F or by 2k stopping the pump briefly and then restarting it. The pressure differ- ential between the catch basin and pump tank would disappear, and leakage of oil would stop as indicated by the hydrocarbon analyzer. After termi- | nation of the test, material from the plug was analyzed and found to be of a composition similar to the test salt. | The original salt deposit is believed to have resulted from a filling operation of the systéem wherein the supply of salt in the dump tank was depleted before reaching the normal level in the pump tank. Gas then bubbled up through-thg dip-line connecting the dump tank to the | salt piping and rose into the pump tank, where it erupted from the surface and splashed salt against the lower end of the shield_plug._ The tempera- ture of the sfirface was low enough to freeze the salt and retain it. Insufficient Running Clearances It is normal practice to try to operate centrifugal pufips at or near their highest efficiency. During operation along the constant sys- tem flow-resistance curve that passes through this best-efficiency pdint, also called the balance line (see Fig. iO), the various losses in the impeller and volute are minimum, and the pressure distribution in the volute is nearly uniform. This uniform distribution gives rfise torthe minimum pet radial force on the impeller. Dfiring operation on higher or lower lines of system flow resistance (see Fig.llo),'thé volute pres- sure distribution becomes nonuniform, and a net radial force is exerted on the impeller that increases_as the operating condition departs farther from the balance line. Thus operation at off-design conditions produces radial forces on the impeller that defléct the overhung shaft. Three incidents of insufficient running clearance to accommodate this deflection were encountered during operation of the MSRE'pumpé with molten salt in the test facility. During the initial molten-salt test_(test 1, Table 1), the prototype pump shaft seized in the shield plug. 0pération at off- design conditions deflected the shaft sufficiently to cause rubbing of the hot, dry shaft against the bore of the shield plug at its lower end. The shaft was friction-welded to the plug over a length of sbout k4 in. The radial running clearance between the shaft and the shield plug was L3 25 ORNL-LR-DWG 72413R HEAD —> CONSTANT FLOW- ® | RESISTANCE LINES CONSTANT . SPEED INCREASING LOAD MSRE FUEL CIRCUIT (DESIGN) INCREASING LOAD oy /<\1PUMP HYDRAULIC BALANCE LINE - MINIMUM RADIAL FORCE ON IMPELLER PROTOTYPE PUMP OPERATION WITH 13-in. IMPELLER | FLOW —> ~ Fig. 10. Relationship Bgtweén Lines of Constant Flow rRes'isrtance ’ Pump Characteristic Curves, and Impeller Radial Load for Centrifugal 26 therefore increased from 0.010 to 0.045 in. to avoid shaft seizure at anticipated off-design operating conditions. During test 9, the shaft of the MSRE fuel-salt pump rubbed against the helium purge labyrinth seal. The rubbing was detected and the pump was stopped before seizure could occur. The diametral clearance was 0.005 in. The rubbing evidently occurred as a result of the buildup during assembly of eccentricities in the mechanical structure of the pump, shaft deflection due to dynamic loading, and shaft run-out. The diametral clearance was increased to 0.010 in. | During test 12, the impeller of the spare reactor fuel-salt pump rubbed against the volute. Rubbing occurred when a flow of cooling air was supplied to the exterior of the nonwetted portion of the pump tank while the pump was being operated at 1200°F. It was found thét the mounting of the volute was slightly skewed in the pump tank, and the axial rumming clearance between the volute and the inlet shroud of the impeller was less than required. Additional properly delineated axial running clearance between the impeller inlet shroud and volute was there- fore provided. 0il Leskage from the Catch Basin into the Pump Tank 0il leakage from the catch basin down the qutside of the shield Plug and into the pump tank was observed in some of the molten-salt ~pump tests and during initial operation of the fuel-salt pump with barren salt in the MSRE. The leakage path traversed a joint that was sealed with a soft-annealed solid-copper O-ring compressed between adjacent flat horizontal surfaces oh the bearing housing and shield plug (see Fig. 11). Modifications were made on the spare rotary elements for the fuel and coolant pumps for the MSRE. | A weld arrangement was devised to seal the joint between the bearing housing and the shield plug in a positive manner. The relationships be- tween the pump shaft, the shaft lower seal, bearing housing, catch basin, shield plug, and the pump tank are shown in the larger section in Fig. 11. The insets show the portions affected before and after modification. The modification was made on both the fuel- and coolant-salt pump spare rotary "elements for the MSRE. 27 _ U , co e 'ORNL-DWG 66-2049 _ SOLID COPPER O-RING B ‘ BUNA N O-RING - LOWER SEAL CATCH BASIN " s , LOWER SEAL CATCH BASIN - - SEAL WELD BEFORE MODIFICATION " AFTER MODIFICATION . SHAFT LOWER SEAL L BEARING HOUSING SEAL OIL. LEAKAGE OUT | ' \,UBE' o # PUMP TANK- SHIELD PLUG- SHAFT § ‘Fig. 11. Cross Sections of Catch Basin.Region of Rotary Element Be- fore and After Modlflcatlon to Seal the 011 Leak Passage Between Shield | = _, Plug and Bearlng Housing.r T jf 3 28 Pump Tank Off-Gas Line Plugging During test 17 (see Table 1) the purge~gas flow rate down the shaft annulus was set at b4 1iters/min to investigate plugging that had been experienced in fhe fuel-salt pump off-gas line at the MSRE and to test filters for possible use to prevent the plugging. The purge flow rate on all previous tests was considersbly lower (factor ofllO or more). At the high test purge rate some partial plugging in the pump tank off- gas line was experienced. The plugging was caused by the solidification of salt aerosol swépt out of the pump tank by the purge‘gas. The aerosol is presumably generated by agitation of the salt in the pump tank by the high-velocity salt Jets that issue from the xenon-removal spray ring. The problem was a nuisance but did not interfere with pump operation. Failure of Weld Attachment of Parts of Flow-Straightening Device Test 12 (see Table 1) was terminated after the operator reported a strange noise that he had heard only one time. A slight roughening of the trace on the pump power recorder was also observed. Inspection indicated that one of the three parts of a flow-straightening device . had become detached and lodged in the impeller inlet. This caused some damage to the inlet edges of the impeller vanes and deformed & swirl preventer of cruciform configuiation that was located édjacent to the impeller inlet. Figure 12 shows the posttest configuration of one of the three parts from the straightener. Figure 13 shows the rubbing damage to the leading edges of the impeller vanes, and Fig. 1k shows the damage done to the swirl preventer. Figure 15 indicates the locations in the flow straightener from which the three parts were detached. Other than the barely perceptible roughening of the power trace, no effects of the lodged parts on the hydraulic performance of the pump could be de- tected. - ' The salt flow had carried the detached parts downstream through the venturi meter and the interconnecting salt piping and upward into the impeller inlet. The straightener parts had been welded in intermittent fashion to the carrier pieces, as shown in Fig. 15. The straightener itself was located just upstream of the venturi meter to straighten the salt flow before entry into the meter. 29 Fig. 12. Part of Flow Straightener After Detachment and Lodging in Impeller Inlet. S pE S A . n 30 o Fig. 13. Rubbing Damage to Inlet Edges of Impeller Blades. Damage was caused by rubbing of impeller against detached part of flow straightener that lodged in impeller inlet. 31 i let. in Damage was caused by lodging of ller - . impe in Preventer. mm&f ghtener 1 ir strai = w Q3 ‘ _ 3 m G e S i D o . 4 ~ M —~ o . . e - o~ QO ) , L . o o ® , s REE | kA o , _ Fig. 15. Locations are 32 Failure of Weld Attachment of Parts in Flow Stralghtener. shown from which parts were detached. 33 At this time it was decided to modify the salt piping as shown in Fig. 16 to locate the venturi meter in the upper leg of the salt piping and the straightener upstream of it near the outlet end of the air cooler. The designs of the straightener and the welds joining its parts were also modified to reduce the vibration-induced fatigue thought to be the reason for the detachment of the three parts. The impeller and the swirl pre- venter were replaced also. Mark-2 Fuel-Salt Pump - The Mark-é fuel-saltrpump represents a modification to the original design of the fuel-salt pump to provide additional volume needed to ac- commodate the thermal expansion of the systen fuel salt. For the original fuel-salt pump, an overflow tank had to be installed at the MSRE to pro- vide the additional volume. | Although the Mark-2 pump was fabricated and Operated with salt in the molten-salt pump test stand, 1t was never installed into the MSRE. It is still being operated in the test stand and in April 1970 had cir- ~culated the salt LiF-BeF, -ZrF, -ThF, -UF, (68.4-24.6-5.0-1.1-0.9 mole %) for more than 14,000 hr, mainly at 1200°F. Its performance has been satisfactory 1n every respect, except for partlal restrictions to the purge-gas flow that occur occa51onally in the off-gas line and in the pump shaft annulus. Description of Pump - The Mark-2 pump has the same hydraulic components (impeller and fvolute), bearing housing, and drive motor designs as those used in the i original fuel-salt pump. However, the height of the pump tank was in- creased by 9 3/4 in. to provide 5 3/h ft3 of additional volume for the a}thermally expanded fuel salt. The pump shaft is the same, except that the length of the salt-wetted portion was increased a corresponding o amount. The conflguratlon of the pump is shown in cross section 1n - Fig. 17 | The detailed designs of 1nternal equipment in the pump tank (i.e., salt splash baffles and salt spray ring) differ from those provided for 34 . ORNL-DWG 66-2048 ) DRIVE MOTOR - 0 t 2 3 e ] FEET AR THERMAL WELL AIR COOLER VENTURI SALT : PUMP v L T T T T == T . _____________ i - DISCHARGE SALT FLOW PRESSURE TENER ' STRAIGHTENER INLET PRESSURE PRESSURE ‘-B-in. P‘PE 6-in. PIPE wa—— SALT FLOW S FREEZEVAVE — | RS ‘ MSRE FREEZE fi FLANGE bump TANK ORIFICE FLOW RESTRICTER Fig. 16. Configuration of Salt Piping in the Molten-Salt Pump Test Stand as Modified After Failure of Weld Attachment of Parts in quW" Straightener. . " ) SHAFT COUPLING - 35 _ ORNL-DWG 69-10459 "LUBE OIL IN BALL BEARINGS {FACE TO FACE) BALL BEARINGS {BACK TO BACK) SHAFT SEAL ——— LUBE OIL OUT — LEAK DETECTOR GAS PURGE OUT GAS-FILLED EXPANSION SPACE NORMAL OPERATING LEVEL - me e == - ] J BEARING HOUSING GAS PURGE IN T ‘ . —SHIELD COOLANT PASSAGES (IN PARALLEL WITH LUBE OIL) SHIELD PLUG ‘BUBBLE TYPE LEVEL INDICATOR XENON STRIPPER (SPRAY. RING} . - o AN = e M- \, f - . . J S & — “ - J_ i’ . .A el Lo * ) - v . — . —_— e = = i Lo : D—\—"—'———/Eb I 1 I | _) J T o SHAFT SEAL | ’ ) 7 i) U “—_LUBE OiL BREATHER " Cross Section of Mark-2 Fuel-Salt’ Pump. - 36 the original tank (Fig. 1). The jets from the spray ring do not impinge directly into—the pool of salt in the pump tank but are aimed at the inner surface of the salt spray baffle. In addition the Mark-2 pump tank is equipped with a buoyancy level indicator (not shown in Fig. 17), vhich was not used in the original pump tank. - | Hydraulic Performance Hydraulic performance data were obtained at various pump speeds along & constant line of flow resistance and at a constant salt ten- perature of 1200°F. Theidata are‘superimposed in Fig. 18 on a graph of water test performance data provided by thé vendor of the impeller and volfité. The head values from the molten-salt data are within 2 ft of the vendor's values. Measurement of Undissolved Gas Content in Circulating Salt The content of undissolved gas circulating in the salt was measured with the pump operating at three different salt levels in the pump tank. At the normal level and the high level, 5 3/8 in. above normal, there was no gas detectable'in the circulating salt by use of the radiation densitometer. At the low level, 2.4 in. below normal, a gas content of 0.1 vol;% was measured. (The radiation densitometer is described in a previous section,) At the normal and high levels, the baffles are evidently effective in keeping gas bubbles from being carried into the circulating salt at the pump inlet. Restrictions to Purge-Gas Flow During operation of the pump, partial restrictions have been experi- enced in the purge-gas flow passages. Examination after initial operation revealed solid material in the off-gas line as far as 40 £t downstream from the pump tank. It had collected in valves and other restricted flow areas, and therefore it was difficult to maintain the pufgefgaé flow of 4 liters/min, the MSRE design rate. A commercial filter'yas therefore installed approximately 15 ft downstream from the pump fiéfik, 37 _ - , . o , - o ~ ORNL-DWG 70-6824 60 — = : o ——— 1150 rpm e I L ALLIS CHALMERS DATA o - \ " : - : 1153 rpm .| DATA POINTS FOR ORNL MOLTEN-SALT PUMP. -~ | . . -fh""“‘-7~ P - PERFORMANCE AT 1200 o - - “h\i::;fl» 1164 rpm N - — - ' 1151 rpm- \ ha ~ o0 J031 rpm 40 . < ;:._,‘ 860 rpm — e ——— z 130 ‘ ' i L T —— = ALLIS CHALMERS Mfg. CO. IMPELLER AND VOLUTE ;:;:::-~‘\ - 5 SIZE 8 x 63, TYPE E; IMPELLER P-482, 11.5 in. 0.D. - 861 rpm s - . . nt- s 20 -.-‘Q\ _..5..9‘9 rpm ol 60z rpm ~ 80z rpm 10 . o0 200 400 600 800 1000 1200 1400 1600 Fig. 18. Hydraulic Performanqe_qf_Marku_Fuelr$qlt_qup.n' | 38 and there was no further plugging. A restriction'now occurs on the average of about once per month in the pump tank off-gas hozzle or at a valve Jjust fipstrefim of the filter. It is removed by rapping on the line or by application of heat with a torch. | Examination of a sample 6f the plug material revealed that it was salt in nearly 8pherical droplet form 15 , in diameter or less. The salt material is carried as an aerosol in the purge gas from the pump tank gas space into the pump tank off-gas line. There have been eight occasions of partial plugging in the shaft annulus (inlet purge-gas_paséage). This plugging can be cleared by either stopping the pump momentarily and restarting it or by heating the system salt to 1325°F for a short period (about 1 hr) and then re- turning it to 1200°F. | Performance of Molten-Salt Pumps in MSRE Two molten-salt pumps, one for fuel salt and one for coolant salt, were instelled in the MSRE. They were developed with the aid of water tests® and the salt tests conducted in the molten-salt pump test stand, as described above. The two pumps are identical except for the hydraulic design (impeller and volute); the fuel pump tank has a spray ring to pro- vide for xenon removal, whereas the coolant pump does not; and the fuel pump is driven at 1175 rpm, while the coolant pump is driven at 1775 rpm. Their accumulated operating statistics, up to the shutdown of the MSRE on December 12, 1969, are presented in Table k4, which gives the accumu- lated pump operating hours for circulation of molten salt, helium, and the combined hours_for helium and salt when the system was at 900°F or above. The service of the MSRE salt pumps, which began in August 1964, was satisfactory, and the total operating times for the pumps exceeded 58,000 hr. During operation of the pumps only one problem was encountered — partial restriction of the off-gas flow. This problem did not inter- fere with the operation of the MSRE but was a nuisance in that it re- quired considerable attention from time to time. We believe that the restrictions resulted from two sources: the freezing of salt aerosol ® 39 and the radiolytic polymerlzation of 011 in the off-gas lines Measure? ments of the hydrocarbon content in the pump off—gas showed 1 to 2 g/day of hydrocarbons. The source of salt aerosol is dlscussed in a prev1ous sectlon, as is the source of the 011 Partial restrictions occurred at the off-gas outlet nozzles of the ‘pump tanks and at other locations, such as valve seats or other points of reduced flow ares. In all ‘other respects, the oPeration of the salt pumps was deemed satisfactory by the MSRE Operations Group. Table 4. Molten-Salt Pump Operation in MSRE Pump Process Head Flow Speed Temperature Total Fluid (ft) (gpm) (rpm). (°F) Hours Fuel Helium and 5900 30,848 . molten salt ' Molten salt 50 1200 1175 = 10001225 21,788 ' Helium - e - . 100-1225 7,385 Coclant Helium and | . 2900 7,438 ’ molten salt -~ . D T AMoiten salt 78 800 1175 1000-1275 26,076 Helium . 100-1275 4,707 *An incident occurred with the fuel-salt pump during one salt-filling ioperation in preparation for startup of the reactor. _The salt was forced to an- excessively high level in’ the pump tank, which resulted in salt '_entering the ‘shaft annulus ‘and freezing. Under these circumstances ‘the motor output torque vas not sufficient to. turn the shaft After appli- cation of extra heat to the: pump tank and sufflcient ‘time for- heat to . . transfer into the frozen region, the shaft became free to rotate and the pump—Was again operative. The filling procedures were modified and - this 1nc1dent did not recur. To) Cbnclusions The development and test programflproduced fuel-=and coolant-salt pumps that, on the whole, operated satisfactorily and dependably during the operating 1ife of the MSRE from August 196l to December 1969. The molten-salt hydraulic performance of the pumps compared favdrdblyfwith the room-temperature water test performance. The pump head with molten salt was observed to be within 2 ft, or 5%, of the water test head, and’ thié difference is within the accuracy of the instrumentation. | The pump shaft purge was effective in protecting the:lower shaft seal region df the fuel=salt pump from radiocactive fission products, and the path for removal of seal o0il leakage remained opened during all MSRE operation. The oil leakage rate for the lower shaft seals was ac- ceptable (maximum of 36 ce/day) throughout MSRE operation. A seal weld scheme, which Joiné'the shield plug and the bearing housing, was developed to prevent oil leaskage in the lower seal catch basin from entering the pump.tank. It was applied to the spare rotary elements, which were not needed in the MSRE and therefore were never installed. Cdnsideration should be given to the handling of salt aerosol in the design of salt pumps for advanced molten-salt reactor systems and to minimizing the production of aerosol. A device should be provided to remove the aerosol content in the purge gas and return it to the salt system. Acknowledgments The satisfactory design, development, fabrication, and operation - of reactor-grade molten-salt pumps for the MSRE required the efforts, enthuSiasm, loyalty, and cooperation of more people than can be named in this report. From this large group, recognition musfi be given to L. V. Wilson for his efforts in the design of the pumps‘and to R. B. Briggs, A. G. Grindell, and R. E. MacPherson for the substantial guidance and encourééement they provided throughout the MSRE'salt pump development program. [ 13 10. hl; '_References R. B. Briggs, MSRP Semiann. Progr. Rept. July 31, 1964, USAEC Report ORNL-3708, p. 375; Osk Ridge National Laboratory. ‘P. G. Smith, Water Tesfi Development of the Fuel Pump for the MSRE, USAEC Report ORNL-TM-T9, p. U7, Oak Ridge National Laboratory, Mar. 27, 1962. C. H. Gabbard, Thermal-Stfess and Strain-Fatigue Analyses of the MSRE Fuel and Coolant Pump Tanks, USAEC Report ORNL-TM-T78, p. T1, Oak Ridge National Leboratory, Oct. 3, 1962. A. G. Grindell, W. F. Boudreau, and H. W. Savage, Development of Centrifugal Pumps for Operation with Liquid Metal and Molten Salts at 1100—1500°F, Nucl. Seci. Eng., T(1): 8391 (1960). R. W. Kelly, G. M. Wood, and H. V. Marman, Development of & High- Temperature Liquid Metal Turbopump, Trans. ASME, Series A, J. of Eng. for Power, 85(2): 99-107 (1963). C. Ferguson et al., Design Summary Report of LCRE Reflector Coolant Pumps and Sump, USAEC Report PWAC-384, p. 41, Pratt & Whitney Air- craft, Middletown, Connecticut, December 1963 0. S. Seim and R. A. Jaross, 5000'gpm Electromaegnetic and Mechanical Pumps for EBR-II Sodium System, Second Nuclear Engineering and Science Conference sponsored by ASME, Philadelphia, Pa., March 1114, 1957. H. W. Savage, G. D. Whitman, W. G. Cobb, and W. B, McDonald, Compo- nents of the Fused-Salt and Sodium Circuits of the Aircraft Reactor Experiment, USAEC Report ORNL-2348, pp. 21?33 Oak Ridge National Leboratory, Feb. 15, 1958 0. P. Steele III, ngh-Temperature Mechanlcal "Canned Motor" Liquid Metal Pumps, Nuclear Engineering Science Conference sponsored by Engineers Joint Council, Chicago, March 17-21, 1958. R. W. Atz, Performance of HNPF Prototype Free-Surface Sodium Pumps , ~ USAEC Report NAA-SR-h336,-Atomics International, June 30, 1960. 11, 12, J. G. Carroll et al., Field Tests on a Radlatlon-Re51stance Grease, ‘Lubricating Eng., 18(2): 6470 (1962). P. G. Smith, Development and Operational: Experlence w1th the. Lubrlca- tion Systems for the Molten-Salt Reactor Experiment Salt Pumps, Oak . lr'Rldge Natlonal Laborato;y, unpublished internal memorandum, June 1970. 13. R. B, Briggs, MSRP Semi ann., Progr. Rept. November 1964, USAEC Report ORNL-3708, pp. 233235, Oak Ridge National Laboratory. 1k, 15. Lo R. B. Briggs, MSRP Semiann. Progr. Rept. December 31, 1962, USAEC Report ORNL.-3369 » PP. 123-124, Oak Ridge National Laboratory. 4 R. B. Briggs, MSRP Semlann. Progr. Rept. January 31, 1963, USAEC . Report ORNL-3419, pp. 37-40, Oak Ridge National Leboratory. " 83 Fuel-Salt Pump Drawings BM-8930 F-9830 E-10965 D~1096L F-9846 D-103k) B-972T7 - D-9711 D~10068 F-98h1 D-10561 D-9849 ~ E-9710 D-9839 D-9837 D-98LT F-98LL E-9843 D-10967 D-9834 D-9832 F-9845 D-9838 F-9836 E-9835 D-9033. D-9831 D-9848 D-9850 D-9720 43 Appendix MSRE DRAWINGS MSRE Fuel Pump - Assembly Pump Tank Test Assembly Flange Wéldmentr' - Upper Shell Weldment Details Dished Head | Volute Bubbler Header Weldment Lower Shell Weldment Sprinkler Head Weldment Thimble Weldment Impeller Details Details | - Labyrinth Flange Weldment Shield Plug Assembly Details - Motor Gulde Weldment - Detaills | Piller Bar - Shaft | , _ ' ;shaft3Coblant Plug WEIdmehtfi | ,LBearing Housing,Assembly | _:f fBéarifig"HOusing Weldment Upper Seal meeafing Clamp Ring Modified Flexible Coupling - Modified Oval Ring, ASA B16-20 Details ' | 4y D-10067 Shield Plate D-98L40 Details l D-108890 Assembly: Shaft Leak Tester D-10888 Assembly: Leakage Test Fixture D-10887 Weldment: Leakage Test Fixture D-10886 Details BM-9T718 Seal, MSRE Pump D-9718 Seal Assembly D-9721 Seal Body Subassembly C-9722 Seal Body D-9720 Seal Spring Plate B-9719 Seal Nose BEM-9730 Universal Joint, MSRE Fuel Pump D-9730 Universal Joint Assembly D-9905 Details BM-10066 Bolt Extension, MSRE Fuel Pump D-10066 Bolt Extension Assembly Coolant-Salt Pump Drawings BM-10062 MSRE Coolant Pump F-10062 Assembly D-9837 Details E-10966 Pump Tank Test Assenbly D-10964 Flange Weldment F-10063 Upper Shell Weldment B-9727 Dished Head D-10068 Bubbler Header Weldment D-97T7L Volute D-103h4k Details D-103h44 Details F-1006L4 Lower Shell D-10065 Thinmble Weldment D-9T7T0 Impeiler D-9834 Details oy D-9847 D-9849 F-98LL E-9843 D-10967 D-9833 D-9837 D-9832 F-98L45 D-9838 F-9836 E-9835 D-9831 D-9848 D-9850 D-9839 D-9720 D-10092 BM-9718 D-9718 D-9721 C-9722 D-9720 D-9T719 653 J 015 653 J.009 828 D 1hh 472 B 183 319157 ED 18399 Lho A 811 Lo A 800 ks Lebyrinth Flange Weldment Details | | Shield Plug Assembly Details Motor Guide Weldment Details Details Filler Bar Shaft ‘Shaft Coolant Plug Weldment Bearing Housing Assembly Béaring Housing Weldment Bearing Clamp Ring | | MOdlfled Flexible Coupling Modified Oval Ring Tachometer Block Details ‘Impeller Wrench Assembly Séal, MSRE Pump Seal Assembly ' Seal Body Subassembly Seal Body - Details Seal Nose 5Fuel and Coolant Pgmp Drive-Motor Drawingé | General Assembly 'ereneral Assefibly General Assembly | Stator Lamination Details' ' Wiring Dlagram for 6 Pole H . E_Wiring Dlagram for 4 Pole ‘Terminal Ball Bearing Detail k72 B 233 Seal Lubrication Stand Drawings - - EM-41801 E-141801 E-41807 E-141803 E-41802 E-41806 D-41810 D-41808 D-41812 D-41813 D-141809 D-41811 D-4180k D-41805 D-4181kL E-L41815 BE-41816 D~55$73 D-555T2 Mark-2 Fuel-Pump Drawings F-56300 F-56301 E-9T10 D-9839 D-9837 D-98L7 D-98L49 F-98L4L D-10967 Lube 011 Packsge, MSRE Piping Assembly Support Frame Tank Piping Subassembly Tank Assembly Flange, Double O-Ring Seal Breather Line | Pump Discharge Manifold Details ' Filter Modification and Assembly Supply Manifold Return Manifold Style "SSD" Valve Modification General Notes Process Piping Connections Instrumentation Assembly and Power Connections - Instrument Support Tray Flow Element (Coolant) Flow Element (ILube) Assembly Test Assembly of Pump Tank Impeller Impeller Nut Slinger Labyrinth Flange Details Shield Plug Assembly ‘Motor Guide ) - D-48225 -Dg983h D-9832 F-5630h D-9838 F-9836 D-9833 D-9831 D-9848 D-9850 D-9720 D-10067 D-56307 D-55508 E-~k8hal E-48L23 ~ F-56302 F-56303 D-48Lop D-48936 C-48935 B-9727 D-10964 D-56305 ' D-56306 D-1034k BM-9718 D-9718 D-9721 C-9722 ~ D-9720 B-9719 BM-9730 D-9730 T o ~ Clamp Seal Filler Bar Shaft -Shaft Coolant Plug Bearing Housing Assenmbly Seal | | | Bearing Clamp Ring Modified Flexible Coupling Modified QOval Ring Details Shield Plate Spool Piece XFMR Mtg. Plate Details ' Splash Baffle Weldment Upper Shell Weldment Lover_Shell‘WEldment Volute | ‘Level Indicator — Float Assembly and Details | Lube Oil Jet Pump Assembly and Details 36~in. Flanged and Dished Head - Weldment Flange _ ~ Capsule Guide and Latch Stop .'_Bubblér Header Weldment ~ Details " Seal, MSRE MK-2 Fuel Pump | _;Seal Assembly ' Séa1 Body Subassembly_ Seal'Body Seal Spring Plate "'Seal Nose - Universal Joint, MSRE Fuel Pump Universal Joint Assembly L8 D-9905 © Detalls | BM-10066 Bolt Extension, MSRE Fuel Pump D-10066 Bolt Extension oy " 21, 22, 23, oL, 25, 26. o7, 28. 29, 30. 31. 32. 33. 3. 35. 36. - 37. © 38, 39. ko, L. - ke, - 43, Lk, ks, L6. L. R. T' W. F. Appler . J. Ball F. Bauman E. Beall Bender . E. Bettis . S. Bettis F. Blankenship . Blumberg . G. Bohlmann J. Borkowski B. Briggs W. Cardwell J. Claffey W. Collins L. Compere H. Cook . W. Cooke B. Cottrell L. Crovley L. Culler . H. DeVan . R. Distefano . Jd. Ditto A, Doss P. Eatherly R. Engel P. Epler . P. Fraas H. Frye C. Fuller H. Gabbard B. Gallaher - R. Grimes . S G. Grindell H. Guymon 0. Harms N. Haubenreich - E. Helms G. Herndon C. Hise _ W. Hoffman . P. Holz Houtzeel I;. Hudson R. Huntley ko Internal Distribution LS. k9. 50.' 51. 52. ~ 53. 5k, 55« 56. 5T 58, 59. 60. 61.. 62. 63. 6k. 65. 66. 67. 68. 69. T70. T1. 2. - T13. . T6. 77 T8. T9. 8o. 81, 82, . 83. 8L, 85. - 87. 89. ~90. 91, 92. o%! +o - 86. . ORNL-TM-2987 H. inouyé = . H. Jordan R. Kasten . Kerlin Keyes . Koger Korsmeyer . Krakoviak . Kress . Krewson . Lindauer . Lundin . Lyon MacPherson . McCoy . McCurdy . McGlothlan McNeese . McWherter Metz « Meyer . J. Miller L,. Moore . M. Perry Richardson C. Robertson . W. Rosenthal " C. Savage . W. Savolainen . H. Shaffer e 24 . ?lfllEi?%C)EiE!fi!FiUfiE:UJP4H1 GrEufoXrdrrHqQPORdHdYIY R DSGHEHEDY n ‘M. J. Skinner G. M. Slaughter A. N. Smith P. G. Smith I. Spiewak R. D. Stulting D. A. Sundberg E. H. Taylor R. E. Thoma D. B. Trauger A. M. Weinberg J. R. Weir J. C. White G. D. Whitman L. V. Wilson Gale Young H. C. Young 95-96. 97-98. 99-102. 103. 10L. 105. 106. 107. 108. 109. 110. 111. 112. -113. 11k, 115. 116. 117. 118. 119. 120. 12]. 122. 123-137. 50 ‘Internal Distribution (continued) Central Research Library Y-12 Document Reference Section ‘Laboratory Records Department Laboratory Records Department (RC) External Distribution- - C. E. Anthony, Westinghouse Electro Mechanical Division, . Cheswick, Pa. 15024 R. N. Bowman, Bingham-Willamette Company, Portland, Oregon 97200 Arthur Bunke, Byron Jackson Pumps, Inc., Los Angeles, Calif. 90000 E. J. Cattabiani, Westinghouse Electro Mechanical Division, Cheswick, Pa. 1502h , D. F. Cope, RDT, SSR, AEC, ORNL o Charles R. Domeck, USAEC, Washington, D.C. 20000 David Elias, USAEC, Washington, D.C. 20000 ' A. Giambusso, USAEC, Washington, D.C. 20000 : Kermit Laughon, USAEC, OSR, ORNL 1 Liquid Metal Engineering Center, c/o Atomics International, P.0. Box 309, Canoga Park, Calif. 91303 (Attention: - R. W. Dickinson) | C. L. Matthews, USAEC, OSR, ORNL T. W. McIntosh, USAEC, Washington, D.C. 20000 C. E. Miller, Jr., DRDT, USAEC, Washington, D.C. 20000 M. A. Rosen, DRDT, USAEC, Washington, D.C. 20000 H. M. Roth, USAEC, Oak Ridge Operations ' J. J. Schreiber, DRDT, USAEC, Washington, D.C. 20000 M. Shaw, USAEC, Washington, D.C. 20000 W. L. Smalley, USAEC, Oak Ridge Operations Laboratory and University Division, ORO Division of Technical Information Extension (DTIE)