X-822 . < MA ST OAK RIDGE NATIONAL LABORATORY Operated By T . UNION CARBIDE NUCLEAR COMPANY 0 R N l_ | (I3 ~,, POST OFFICE BOX P CENTRAL FILES NUMBER OAK RIDGE, TENNESSEE "y 58-2-40 DATE: February 18, 1958 copy No. & 2 SUBJECT: Mclten Salt Heat Transfer External Transmittal Authorized TO: Distribution DISTRIBUTION l. A, L. Boch 2. A. P, Fraas 3-520 Ho W- HOffmaIl ‘ 53. W. H. Jordan " 54, H. G. MacPherson " 55. A. J., Miller \ 56. J. A. Swartout v 57T. A. M, Weinberg 58-59. Centrsl Research Librafy 60-61. laboratory Records 62. ORNL-RC 63. REED Library 64-78. TISE 79. M, J. Skinner 80, Doc. Ref. Section NOTICE This document contains information of a preliminary - nature ond was prepared primarily for internal use at the Oak Ridge National Laboratory. It is subject to revision or correction and therefore does not represent a final report. UNCLASSIFIED LEGAL NOTICE This report was prepared as an account of Government sponsored work., Neither the United States, nor the Commission, nor any person acting on beha!f of the Commission: A. Mokes any warranty or representation, express or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this repert, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B. Assumes any licbilities with respect to the use of, or for damages resulting from the use of any information, apparatus, method, or process disclosed in this report, As used in the above, '‘person acting on behalf of the Commission’ includes any employee or contractor of the Commission to the extent that such employee or contractor prepares, handles or distributes, or provides access to, any information pursuant to his employment or vontract with the Commission. MOLTEN SALT HFEAT TRANSFER H. W. Hoffman Oak Ridge National Laboratory Post Office Box Y Oak Ridge, Tennessee o ABSTRACT The experimental system used in the determination of turbu- lent forced-convection heat-transfer coefficients with molten salts is described, and the resfilts of the experimental program are detailed. The salts studied have been the mixtures Nafloa- NaN03-KNO3, LiF-NaF-KF and NaF-Zth-UFh, and sodium hydroxide. Low results with the LiF-NaF-KF mixture in Inconel tubes are ex- plained in terms of en interfacial film. A nonwetting phenomenon is postulated to describe the decreased heat transfer in the NaF- ZrFu'UFu system, The advantages of molten salts as coolants are briefly discussed. It can be concluded that molten salts behave, in general, as ordinary fluids (0.5 < Np,. < 100) as far as heat transfer is concerned., An analysis has shown that the molten salts compare Tavorably with the liquid metals as reactor coolants. }." INTRODUCTION The study of the heat-transfer characteristics of molten salts has received much attention in recent years as these fluids have been found to be potentially good high-temperature reactor coolants and fuels. These materiels are generslly characterized by low vapor pressures (allowing near atmospheric operation), high melting points, and good thermal properties. The thermal conductivities of these fluids lie somewhere between those of water and the poorer liquid metals, and the thermal capacities per unit volume are as good as that of water., Studies of the heat and momentum transfer characteristics of molten salt coolants indicate that the molten salts compete favorably with the liquid metals. Table 1 compares the working temperature range (melting temperature to boiling temperature) and the Prandtl moduli of a number of fused salts with those of scme of the more common heat transfer fluids. Thus, Dow- therm'A has a working range of only 400°F as compared to 1000°F - 2000°F for the molten salts. It is to be noted, of course, that this does not provide the salts with an advantage over the liquid metals. However, fused salts do possess an advantage over many of the liquid metals in that they are nonflammable. A comparison of the final column in Table 1 shows that the fused salts have Prandtl moduli which lie in the same range as those for ordinary fluids (0.5 < N, < 100). From this, it is to be Pr anticipated that the molten salts behave as ordinary fluids with respect to heat transfer. Since littlie information was available in the technical literature on the heat-transfer characteristics of fused salts, an experimental pro- gram to obtain heat-transfer coefficients for fused salts was initiated at the Oak Ridge National ILaboratory. The fluids studied have been sodium hydroxide, the eutectic mixture of sodium, potassium, and lithium fluoride (11.5-42.0-46.5 mole %), the mixture NaNO -Nal\T03-KNO3 (40-7-53 weight %) 2 commercially known as "HTS", and the mixture NeF-ZrF) -UF) (50-46-4 mole %). EXPERIMENTAL SYSTEM The experimental determination of the heat-transfer coefficients of molten salts presented a series of unususl problems. To insure against leskage of the molten salt, an all-welded experiments]l system was used. The high melting temperatures of the salts required that those parts of the system in contact with the fluid be held at elevated temperstures at 8ll times. This seriously affected the lifetime of some system components and increased the difficulty of servicing the equipment during operation. In sddition, the problems of temperature, pressure, and flow-rate measure- ment were aggraveted by the high system temperatures. The genersl tech- niques used were not new to experimental investigations but did require modification to satisfy the restrictions imposed by the fused salt systenms. To prevent contamination of the salt, helium was used as the blanket gas with residuasl oxygen and water vapor removed either by bubbling through sodium-potassium alloy or by passage over titanium sponge at 80000° The experimental system is shown in Figure 1. The flow of fused salts through the test section was effected by gas pressurization. A gas prgssurized flow system possesses the advantages over a pump system of simplicity of construction and of trouble-free operation but requires greater quantities of salt to insure temperature equilibrium at high flow rates. In this system the fluid was pushed through the test section, lo- cated between the two tanks, by pressurizing one of the tanks with the inert blanket gas while venting the other. At the end of a cycle - i.e., vhen a1l of the salt had been transferred into the second tank - the pres- sures were reversed and the fluid was pushed in the opposite direction. Generally, the device used to measure the fluid flow rate can be used also to control the fluid cycling. During the course of this experimentsl program, numerous modifications were made to the physical system to provide a more flexible apparatus. However, the only major change was & decrease in the size of the system. This was done to reduce the total salt inventory and to alleviate some of the effects of thermsl expension on the test section. Thus, the photograph of Figure 1 shows the experimental system in the later stages of its devel- opment. The two fluid reservoirs were Jjacketed by thermally insulated elec- trical heaters. The test section was heated by the passage of a high amper- age AC through the tube wall. The power to the test section was supplied through electrodes soldered to the tube at each end., Mixing chambers were located at the outlet and inlet ends of the test section. The fluid flow rate was obtained by both weight and volume measuring techniques. To obtain the weight rate directly, one of the two tanks was mechanically isolated from the rest of the system by a flexible bellows. This tank was then supported by a beam balance or a cantilever weigh beam, and the change in weight was observed as the fluid flowed into or out of the tank. With the cantilever beam, the movement of the beam as it de- flected under the varying load of the tank was sensed at the free end by a differential transformer, The cantilever weigh beam possessed high sensitivity with beam deflections as smell as 10 microinches being readily detected. Direct calibration of either system was obtained by placing weights on the tank or by flowing water at a constant rate into a contain- er resting on the tank. In another approach to flow measurement, the time required for the fluid to fill a known volume was obtained. The volume was defined by two bare metal probes inserted into one of the tanks at two different levels. Theée probes were arranged such that the rising fluid on contacting the first probe would complete an electric circuit and activate a timer. When the second probe was reached, the timer would be stopped. From the volume-rate and the fluid density, the weight-rate was calculated. This technique possessed difficulties, particularly in the deterioration and electrical shorting of the probes and was not used to any great extent. Fluid cycling was accomplished by using the signals from the probes or the weigh-~beam to actuate gas solenoid velves. The test sections were fabricated from small diameter nickel, Inconel, and type 316 stainless steel tubing. The tube dimensions are shown in . Table 2. The tube outside surface temperatures were measured with 32-gage chromel-alumel thermocouples resistance-welded to the tube., The fluid- metal interface temperatures were obtained by correcting the outside sur- face temperatures for the calculated radial temperature drop in the tube wall. The thermocouple leads were wrapped around the tube for at least one-quarter turn to minimize thermocouple conduction errors. The fluid mixed-mean temperatures were obtained in baffled mixing chembers with the thermocouples contained in relatively low-mass thermowells. Several types of mixing pots, differing only in the method by which fluid mixing was achieved, were used. One design consisted of a short length of pipe closed at both ends. A whirling flow was indficed by a tangentisl inlet, and fur- ther mixing was obtained through a perforated disc located at the center of the unit. A thermcwell extending along the axis enabled temperature meas- urement at the inlet and outlet of the mixing chamber. Alternatively, a modified "disc-and-doughnut" mixing chamber was used. Thermocouples were calibrated at the freezing points of lead, zinc, and aluminum. By sectioning the tube after the completion of each test series, the calibration of the thermocouple attached to each of the small tube pieces could be determined. The voltage distribution along the test section was obtained using one of the wires of the tube surface thermocouple at each position as a voltage probe. The test section was Jjacketed by a container filled with a granular thermal insulation. The mixing chambers, guarded by small heaters, were adjusted to approximately the temperature of the test fluid before operation. Part of the total energy put into the test section is lost to the sys- tem environment. Since the magnitude of this loss must be known to obtain a heat balance for the system, a heat loss calibration was made. This de- termination was made for each test unit, as the system heat loss is a func- tion of the density and distribution of the insulation around the test sec- tion and of the end conditions. A typicel heat loss calibration for a 2k-inch test section is shown in Figure 2. EXPERIMENTAL RESULTS Figures 3 through 8 show the results of these investigations with the molten salts. Figure 3 presents the ORNL data on NaOH (closed circles) cor- related as the Colburn j-function. The generally accepted correlating equa- tion for ordinsry fluids is given by the line, j = 0.023 NR;O'E. The data of Grele and Gideon (1) obtained at NACA using a resistance-heated tube in 8 pump system are shown by the closed triangles. In addition to heating data, Grele and Gideon made measurements on the cooling of sodium hydroxide in an air-cooled double-tube heat exchanger. These results are given also (open triangles) in Figure 3. The reason for their high results and con- siderable scatter is not clear. Figure 4 shows the results of the study with "HTS". The first reported measurements for molten salt heat transfer were those of Kirst, Nagle, and Castner (2) using "HTS" in an electrically heated iron tube. ILacking data on the thermal conductivity of this salt mixture, these investigators -10- correlated their data in terms of the dimensional function, hd/uo'u. Over the Reynolds modulus range of 2,000 to 30,000, the average line through their experimentsl values was given by the equation hd dG P—:E = 0,000442 [";.T Jl.lh ) In Figure 4 the data of Kirst, Nagle, and Castner have been recalculated as the j-function using thermal conductivity values determined at ORNL. Tn this figure the ORNL results with "HTS" are given by the closed circles. These data cover the Reynolds modulus range of 5,000 to 25,000 and show good agreement with the empirical equation describing forced-convection heat trans- fer in ducts containing ordinary fluids. Figure 5 summarizes the results of the heat-transfer experiments with the LiF-NaF-KF eutectic mixture. The first tests were conducted using a nickel tube for the test section. Data (open trisngles) were obtained in the transition flow region; and the results were in general agreement with predictions as to the heat transfer with this molten fluoride. Attempts to extend the NRe range of the data by increasing the temperature were unsuc- cessful due to fatigue failure of the nickel tube when the temperature ex- ceeded 1000°F, Following this the nickel tube was replaced by an Inconel section, since Inconél fetains its structural strength to very high temperatures. The re- sults of the tests with the Inconel test section (open circles of Figure 5) indicated that the heat transfer was a factor of two lower than that predicted from the earlier NaOH, "HTS", and LiF-NaF-KF nickel tube data. Three pos- sible explanstions for the observed difference were possible: (1) an error in the experimental measurements, (2) incorrect physical property dste, or (3) an unknown thermal resistence at the fluid-metal interface., Reanalysis of the experimental system indicated the possibility of a thermocouple con- duction error in the fluid mixed-mean temperature measurements. Therefore, the mixing chamber design was altered, and a third series of runs were made,. The results, shown by the closed circles, verified the previous test series in Inconel., Remeasurement of the physical properties eliminated this factor as the source of error., Visual examination of the inside surface of the tubes from both series of :uns with the fluoride salt showed a green deposit on the exposed tube surface, This was identified, chemically and petrograph- ically, as a mixture of complex fluorochromates of which the major constituent was K CrF6. 3 The true heat-~transfer coefflcient, h, is given by the equation 1 by = Ty 1) oy a/A’ where A' is the inside surface area of the film, y, the film thickness and ki’ the thermal conductivity of the film. Since the films are thin (in this case 0.4 mils as indicated by photomicrographic study), the ratio of the heat-transfer area for the film to that for the clean tube, A'/A, is approxi- mately anity. Thus, we can replace (qf/A') by (qf/A) and write h = 2 (3) 0.2 Substituting h as determined from the equation, j = 0.023 NR; , and the experimental coefficient, h', into equation (3) resulted in a velue of 0.0002 hr-rt°-°F/Btu for the thermal resistence, y/k,. As a cross-check, measurements were made of the film thickness and the thermal conductivity of pure KBCrF6' Using v = 0.4 mils and the preliminsry value ki = 0.133 Btu/hr-ft-oF, a thermal resistance of 0.00025 hr-fta-oF/Btu was obtained. This compared closely with the value (0.0002) deduced from the heat-transfer data. Two final experiments were made with the LiF-NaF-KF mixture. The first was with a second nickel tube to verify the earlier results. The results (closed triangles) were in essential agreement with the previous nickel tube data. The second test used a type 316 stainless steel tube. Preliminary tests had indicated that the fluoride eutectic mixture could be contained in this alloy for short times without serious corrosion or film formstion. This was verified by examination of the tube at the con- clusion of the test. The results (indicated by the open squares) served as additional evidence that the reduced heat transfer in Inconel could be attributed directly to the deposits on the heat-tranéfer surface, Heat-transfer studies were also made with the fluoride mixture NaF-ZrF) -UF), in Inconel. The results are shown in Figure 6 which compares the data obtained by two investigators at ORNL. At a Reynolds modulus of 10,000, both sets of dats fall approximatley 22% below the genersl correlation for ordinary fluids. Salmon (3) used a double-tube heat ex- changer with sodium-potassium alloy flowing through the annulus as coolant and s length-to-diameter ratio of 40. Center-line temperatures were meas- ured at both the inlet and outlet of the fluid streams. An adjustable probe was used to measure the surface temperatures on the annulus side of the cen- ter tube. Both the measured surface temperatures and the Wilson (4) graphical analysis technique were used in evaluating the heat-transfer coefficient. Salmon's results based upon the measured surface temperatures exhibit con- siderable scatter and are not shown. Unlike the LiF-NaF-KF: Inconel sys- tem, visual examinstion of the exposed surface of the test section showed no deposits, However, the surface appeared to have been unwet by the salt. If the salt did not wet the tube, then perhaps the low heat transfer re- sults could be explained on the basis of an additional thermal resistance due to a gas film at the metal surface, As a check on the possibility of 8 systemic error, a duplicate apparatus was operated with water as the test fluid. The results agreed withh the general heat-transfer correlation. In an effort to gain some further insight into the resson for the de- creased heat transfer with this zirconium-base fluoride, pressure-drop meas- urements were made with this salt flowing isothermally through a smooth cir- cwlar tube. The test section consisted of a 50-inch length of Inconel tubing, l/kmin. OD x 0.035-in., wall thickness, with two short-radius right-angle bends. The effects of entrance contraction, exit expansion, and the two bends were determined by experiments with water. The equivalent tube length determined -1Y4- from the water measurements was used in calculating the fused salt friction factor. The pressure drop across the test section was obtained as the dif- ference in gas pressures in the fluid reservoirs at each end of the test section. The results of these measurements are compared in Figure 7 with the equation representing the friction factor in the Réynolds modulus range from 5,000 to 200,000, The experimentally measured friction factor fell approximately 16% below the velue which would be predicted. On the basis that J = £/2, the heat-transfer and momentum-transfer experiments with the zirconium-base fluoride salt show good agreement. Some information on the thermal entrance length was obtained from the sodium hydroxide and LiF-NaF-KF experimental data. The entrance system con- sisted of a thermal entry region preceded by a hydrodynamic entrance region of 8 to 13 tube diameters. The entrance length, (x/d.)e, was teken as the position at which the local heat-transfer coefficient had decreased to with- in 10% of ité fully established value. The observed variation of (:g/d)e with the Peclet modulus is shown in Figure 8. DISCUSSION The physicél properties of the salts studied were determined in an in- dependent experimental program. The literature values for viscosity and heat capacity for sodium hydroxide and "HTS" were also checked. The largest uncertainty exists in the thermal conductivity of these salts. However, the experimental effort aimed at establishing the temperature dependence of the thermal conductivity of these fluids is continuing strongly. - =15- Additional thermal resistance at the heat-transfer surface, whether caused by a surface film-forming reaction or by "nonwetting", will result in drastically reduced heat transfer in small tube systems. Since the ex- perimental data indicate that such phenomens cen occur with LiF-NaF-KF and NaF-Zth-UFu, the use of these salt mixtures as coolants may be limited. This deficiency may be ameliorated by the development of suitable container materials or "wetting" agents. In general, it can be concluded from the ex- perimental results reported that the molten salts behave as ordinary fluids (0.5 < Ny, < 100) as far as heat transfer is concerned. An anelysis on the basis of a "cooling work modulus” (flow work per unit heat removel) shows that as reactor coolants the molten salts compare favorably with the liquid metals, This modulus, which is a function of the system geometry, the flow regime, the thermal properties of the fluid, and certain coolant temperature differences, is defined in the first eguation of Table 3.* Substituting for Pp (the flow work based on the pressure drop), using the equation describing the friction factor in the turbulent flow regime, results in the expression of the cooling work modulus as a function of the parameters outlined above. The second equation in this table defines the over-all tempersture difference in the reactor-heat-exchanger system postulated. The third equation expresses the Nusselt modulus for the sys- tem in terms of a number of dimensionless parameters involving the geomet- rical proportions, heat generation rates, and temperature differences for the system. This equation was obtained by substituting the equations de- scribing the heat transfer in the reactor and the heat exchanger into the * " - The derivation of this equation has been given by Rosenthal, Poppendiek, and Burnett.5 ’ -16- temperature difference equation. The system on which this analysis was based is shown schematically in Figure 9. The over-all temperature difference defined is seen to be the difference between the hottest point in the system (the reactor wall at the exit) and the heat rejection temperature {the coolant-side wall temperature in the heat exchanger). It was assumed that heat generated uniformly within the reactor. fuel elements was removed by & non-heat-generating coolant flow- ing turbulently through circular, parallel tubes uniformly distributed through the core. The coolant passages were presumed of sufficient length to allow thermal and hydrodynsmic entrance effects to be neglected, The physical properties of the fluids were assumed temperature-independent, and the thermal sifik in the heat exchanger was chosen to be a liquid boiling on the outside surface of the tubes. The Nusselt modulus equation for this system was solved simultaneously with the Nusselt modulus equations for fully developed turbulent pipe flow for both ordinary fluids and liquid metals to obtain the Reynolds modulus. The resulting value of the Reynolds modulus was substituted into the first equation of Table 3 to obtain the cooling work modulus. The specific set of reactor dimensions and operating conditions are given in Table 4. The cooling work modulus as a function of tube diesmeter for the coolants con- gidered - lithium, sodium, bismuth, sodium hydroxide, and the LiF-NaF-KF eutectic - is shown in Figure 10. It can be seen that lithium is the best of the five reactor coolants while bismuth is the worst. Sodium, sodium hydroxide, and the Lif-NeF-KF mixture are comperable in heat transfer effec- tiveness with the two salts being better over the lower part of the tube i =17~ diemeter range. For very smell tube dismeters (approximately one-tenth inch), the alkali-metal fluoride salt is as good a coolant as lithium. CONCLUSIONS The molten salts, in general, behave as ordinary fluids as far as héat transfer is concerned. Since phenomens such as film formation and non- wetting (demonstrated in some molten salt systems) are difficult to predict, the heat-transfer characteristics of the molten salts for critical appli- cations must be experimentally established. An analysis has shown that the fused salts compare favorsbly with the liquid metals as reactor coolants. REFERENCES 1. Grele, M. D., Gideon, L., Nat. Adv. Comm. Areo, RM E52109 (1953). 2. Kirst, W. E., Nagel, W. M., Castner, J. B., Trans. AIChE, 36, 371 (1940). 3. Salmon, D. ¥., Oak Ridge National Laboratory, unreleased data. 4. Wilson, E. E., Trans. ASME, 37, 47 (1915). 5. Rosenthal, M. W., Poppendiek, H. F., Burnett, R. M., ORNL CF 54-11-63 (1954). & - Table 1 Table 2 Table 3 Table 4 Figure Figure Figure Figure Figure Figure Figure Figure Figure W = W Figure 10 -18- LIST OF TABLES AND FIGURES Working Temperature Range and Prandtl Modulil of Various Fluids Dimensions of Test Sections Equations for Cooling Work Modulus Analysis Reactor Dimensions and Operating Conditions for Cooling Work Modulus Analysis Experimental System for Molten Sslt Heat-Transfer Studies Heat Loss Calibration Curves Sodium Hydroxide Heat Transfer "HTS" Heat Transfer LiF-NaF-KF Heat Transfer N&F-ZrFH-UFh Heat Transfer Isothermal Friction Pactor for NaF—Zth-UFh Flowing in a Smooth Circular Tube Thermal Entrance Length for Sodium Hyfiroxide and the LiF-NeF-KF Eutectic Temperature Distribution in Resctor-Heast-Exchanger System Comparison of Coolants -19- TABLE 1 Working Temperature Range and Prandtl Moduli of Various Fluids Melting Boiling Temp. Prandtl Moculus Fluid Temp. (1 atm,) Temp. Range Pr °F °F °F Water 32 212 50 - 200 9.6 -1.9 1 Do~ ~vm A 53 496 100 - 450 | 3k.2 - 4.5 Mercury -38 675 50 - 600 0.027 - 0.008% ,' Sodium 208 1621 200 - 1300 0.011 - 0.0038 . Bismuth 520 2691 600 - 1400 0.014 - 0.0084 _’ Air --=(to 1600)---~ 0 - 1600 0.69 - 0.72 "9 NedNOL -NalNO,-KNO, | 288 d 1000 570 - 750 | 4.9 - 2.7 | Sodium hydroxide | 60k 1652 750 - 930 6.1 -~ 3.7 LiC1-KC1 679 950 1.1 LiF-NeF-KF 854 2858 1020 - 1300 2.6 - 1.7 ' 4 » & -20- TABLE 2 Dimensions of Test Sections S 1t Test Section & Wall Length x Material OD (in.) | Thickness (in.) Diameter ’ d NaOH Nickel 0.1875 0.035 204 "HTS" Inconel 0.250 0.035 47, 133 316 s. s. 0.250 0.035 W7 LiF-NaF-KF Nickel 0.1875 0.035 20k Inconel 0.225 0.025 137 316 s. s. 0.250 0.035 133 NaF-Zth-UFH Inconel 0.250 0,035 47 v -21- ¢ i<.} TABLE 3 Equations for Cooling Work Modulus Analysis P 1 3 VR/V P 0.092 R t 2.8 Fp=Jq_t=gcJ 2 ’+(' V) NRe,R PR 4" (9 /7, , A$t = Amr,R + Ama,R + Axr,HX o . - 1 _ P - XN 14 . 1 "Nu,R Re,R "Pr,R XB Nu, HX . *2+,R Ypr,R e -1 < ¢ 2P~ TABLE k4 Reactor Dimensions and Operating Conditions for Cooling Work Modulus Analysis Reactor length, LR 3.00 ft Resctor diameter, dR 3.00 ft Retio of volume of coolant in reactor ’ to total reactor volume, VR/Vf 0.30 Heat removal rate, dp 200 Mw Ratio of heat removal rate to resctor volume, qo/V, 35.1 x 10° Btu/hr-£t3 Over-sll temperature difference, At 500°F t Fig. 1. Experimental System for Molten Salt Heat-Transfer Studies. UNCL ASSIFIED PHOTO 26700 .-..E:Z_ OUTSIDE TUBE SURFACE TEMPERATURE (°F) 1800 1600 1400 1200 {1000 800 600 400 200 DISTANCE FROM ENTRANCE OF TEST SECTION (in.) Fig. 2. Heat Loss Calibration Curves. t » < ‘ . ':‘ . - - 'i . UNCLASSIFIED ORNL-LR-DWG 3449 o ] rd N A A A e A Are— A’ e e Lr ] o] "} 4——9—————0—_'—']—'— \‘ A A [/ e * . o—T 0 ° ST e N / A — F A A e M " ELECTRICAL HEAT INPUT (Btu/hr) QO 444 A 337 o 300 ® 250 A 204 2 4 8 10 42 14 16 18 20 22 24 ...VZ-. UNCLASSIFIED ORNL-LR-DWG, 15061 .00 008 006 .004 COLBURN FUNCTION, j=Ngy Np 2’3 .002 @ HOFFMAN® A GRELE,GEDEON, HEATING® A GRELE,GEDEON - COOLINGZ® .00 1000 10,000 REYNOLDS MODULUS, NRe 100,000 Fig. 3. Sodium Hydroxide Heat Transfer. —gz- 2 COLBURN FUNCTION, /= Ag, - My, '3 0.010 0.005 0.002 0.004 - ' L e @ o b S 3\‘.‘_ * r ” ,J UNCLASSIFIED ORNL~-LR-DWG 27704 O O O O o ) o © ° o? -0.2 0.023 Ny, O KIRST, NAGLE, CASTNER ® HOFFMAN, COHEN, GREGORY 103 2 5 104 2 5 10° REYNOLDS MODULUS, Ay, Fig. 4. '"'HTS' Heat Transfer. -.9z_ % “ o N = Ny T COLBURN FUNCTION - ~ %* - > » ” "1 ;fi —— - ‘ - UNCLASSIFIED ORNL«L R-DWG, 3453 0.0410 A NICKEL TUBE NO.4 v NICKEL TUBE NO.1 O INCONEL TUBE NO.4 ® INCONEL TUBE NO.2 0.005 A NICKEL TUBE NO.2 : A O TYPE 316 STAINLESS STEEL TUBE A T T o /'/ A \ . % / a3 -0.2 ——— , A A Ny N, = 0023 N 2 / 7V / I V e 0.002 2 o1 o ® g ©lo O 0.001 1000 2000 5000 10,000 20,000 50,000 Nge REYNOLDS MODULUS Fig. 5. LiF-NaF-KF Heat Transfer. _Lz_ 2/3 0410 UNCLASSIFIED ORNL-LR-DWG, 15022 .008 .006 .004 .002 O N .______,:/ : Re ' O ® HOFFMAN O SALMON 001 1000 10,000 REYNOLDS MODULUS, Ng, Fig. 6. Nc:F-ZrF4---UF4 Heat Transfer. 100,000 -Bz.. FRICTION FACTOR, f/2 004 .002 010 008 006 .00 {000 UNMCLASSIFIED ORNL-LR-DWG, 15023 -0.2 f/2=0.023 Ng, 10,000 100,000 REYNOLDS MODULUS, Nge Fig. 7. Isothermal Friction Factor for Na F-ZrF,-UF, Flowing in a Smooth Circular Tube. —6C- -30- E 1 UNCLASSIFIED ORNL—LR—~DWG 3455 50 A FLINAK ® SODIUM HYDROXIDE / / ® ® / ® ,® ¢ y 20 ® o~ ./ ® ® {x d)e +. 4y 10 / 3 v // // N 5 N A/ A 3 . 3 . { 6,000 10,000 50,000 100,000 ; Npe . PECLET MODULUS Fig. 8. Thermal Entrance Length for Sodium Hydroxide and the LiF-NaF-KF Eutectic. TEMPERATURE, REACTOR REACTOR TUBE WALL TEMPERATURE % COOLANT TEMPERATURE HEAT EXCHANGER UNCLASSIFIED ORNL-LR-DWG, 15024 ——me—— =g Atr’ R COOLANT-SIDE HEAT EXCHANGER TUBE WALL TEMPERATURE DISTANCE ALONG TUBES Fig. 9. Temperature Distribution in Reactor-Heat Exchanger System. - le_ COOLING WORK MODULUS, 7, ~32- UNCLASSIFIED ORNL.LR-DWG, 15025 5 /N(:IOH A / LiF-NaF-KF Na H // \\ —~— S ® Li o0 2N\lVd / 6 8 504 2 4 6 8 g4 2 4 REACTOR TUBE DIAMETER, dg (ft) 10° \ p Fig. 10. Comparison of Coolants.