1、C IC 1 i k 0 RESEARCH MEMORANDUM FORCED-CONVECTION HEAT-TRANSFER CHARACTERISTICS OF MOLTEN SODIUM HYDROXIDE By Milton D. Grele and Louis Gedeon Lewis Flight Propulsion Laboratory Cleveland, Ohio NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS WASHINGTON February 17,1953 “ . . . . . . . . . . “. Provided
2、 by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1R * 9 N NACA RM E52LO9 3 1176 . 01435 6563 . NATIONAL ADVISORY COMMFETE FOR AERONAUTICS FORCED-CONVECTION HEAT-WSEER CHAFACTERISTICS OF MOLTEN SODIUM EYDROXILcrE By Milton D. Grele and Louis Gedeon An investig
3、ation of the forced-convection heat-transfer character- istics of sodium hydroxide was made for a range of Reynolds .nugber Yrom 5300 to 30,000, corresponding to velocities from 3.8 to 15.4 feet per second, average fluid tenperatures cp to 938 F, and heat-flux densities up to 226,000 Btu per hour pe
4、r square foot for both heating and cooling. In addition, some heat-transfer tests were ,made with an aqueous solution of sodium hydroxide. Y Water heat-transfer tests were also made to check the instrumentation. * When the sodium hyd2oxide heating data are correlated by the famil- iar Nusselt relati
5、on, the data fall slightly above the McAdam correla- tion line. The sodium hydroxide cooling data are fairly well repre- sented by the McAdams correlation line. This report contains the heat-transfer data obtained at the NACA Lewis laboratory for sodim hydroxide flowing in an electrically heated Inc
6、onel tube. Heating data were obtained for a range of Reynolds number from 5300 to 29,000, corresponding to velocities from 3.8 to 15.4 feet per second, average surface tempemtures up to 967O F, average fluid temperatures up to 918O F, and heat-flux densities up to 226,000 Btu per hour per square foo
7、t. Cooling data were taken concurrently with the heating data, with the Sodium hydroxide flawing through the center passage of a single-tube, Inconel, counter-flow, sodium-hydroxide-to-air heat exchanger. Data were obtained for a range of Reynolds number from 6500 to 30,000, correspond- ing to veloc
8、ities from 5.9 to 15.4 feet per second, average surface tem- peratures up to 915O F, average fluid temperatures up to 938O F, and heat-flux densities up to 120,000 Btu per hour per square foot. . t Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-2 AP
9、PARATUS NACA RM E52L09 A photograph and a schematic drawing of the test setup are shown in figure 1. A centrifugal pw we6 usee the molten sodium hydroxide from the rmmp tank through the heating test section, cooling test section, heat exchanger, and into the volume tank. From the volume tank, the mo
10、lten sodium.hydmxide returns to the sump tank. The setup is so arranged that upon completion .of a run the sodium hydroxide drains back into the sump tank. Each section of the test setup- is described in the following paragraphs. cy Sump tank. - The sump tank housed the centrifugal pump and served a
11、s a storage tank for the 20 pounds of sodiwn hydroxide circulated through the system. The tank was made of Inconel with an inner diameter of Y inches and a depkh of 5 inches. During operation, an atmosphere of nitrogen was kept in the space above the liquid level. .- Circulating pump. - The circulat
12、ing pump was a centrifugal pump driven by a 1 hp air motor. The rotor, housing, and shaft were made of Inconel. The gasket material used at the split housing was two sheets of 0.004-inch nickel. The pumg was totally immersed in the molten hydroxide and supported from the cover plate of the sump tank
13、. The bearings and shaft-supports were located in a water-cooled housing above the sump tank and a slinger ring on the pump shaft prevented leakage. The flow rate wa8 controlled by varying the alr supply to the motor. Heating test section. - Preliminary tests showed that an elec- trically heated tes
14、t section could be used for the heat-transfer tests c because only a negligible amount of heat would be generated.in the sodium hydroxide. A schematic drawing of the heating test section is shown in figure 2. The heating test section was fatiricated from Inconel tubing having an outside diameter of
15、3/8 inch, R wall thickness of 1/16 inch, and an effective heat-transfer length of 24 inches. Three stainless steel electric flanges were welded.12 inches apart to the test section and were connected by flexible. straps and buss bars to a power source for resistance heating. Electrically, the test se
16、ction was con- nected in parallel, the current being divided at the center flange. Since the test sectian was not electrically insulated from the rest of the system, the outer flanges were grounded, thus maintaining the rest . of the system at the same electrical potential. For starting purposes, gu
17、ard ring heaters were used on each of the electrical flanges to eli- minate cold spots on the test section. Mixing tanks provided with baffles were located at each end of the test section. A thermocouple was located at the dawmtream end of each mixing t the outer tube was 1- inches in diameter with
18、-a 1/16-inch wan. A stainless steel bellows in the outer tube was used to take up the differential expansion between the two tubes. The effective 1 4 N m 0 heat-transferlength was 24 inches. Sodium hydroxide Wet and outlet N temperatures were measured in mixing cans. Nine buried chromel-alumel therm
19、ocouples on the inside tube were , used to obtain wall temperatures. As shown in figure 3, .a c*cumferen- tial and.a longitudinal groove 0.031-inch deep and 1/32-fnch wide were cut at each thermocouple location. The thermocouple junction was placed in the longitudinal groove and each lead was wrappe
20、d around the tube 1/2 turn to reduce conduction losses from the junction. A ceramic was used to insulate the thermocouple leads from the tube. Inconel spray imide a 1/8-inch Inconel tube brought the leads to either end of the annulus. The location of each thermocouple junction was measured as surf a
21、ce. 9 weld covered the leads and located them permanently. Alundum tubing + accurately as possible and was found to be 0.045 inch from the inside Service air at 100 pounds per square inch was available for coolfng and was controlled by a regulator. Heat exchanger. - The heat exchanger was used to ma
22、intain a constant inlet temperature to the heating test section, since at a high rate of heat input to the heating test section, the cooling test section was inadequate in removing this heat. The heat exchanger was fabricated from two Inconel tubes forming an azII1u1us, with the. sodium hydroxide in
23、 the inner tube ahd the counter-flaw air in the annulus. The dimensions of the inner tube were 3/4-fnch outside diameter with a 1/16-inch wall, and of the outer tube, 1 inches with a 1/16-inch wall. 1 Volume measuring tank. - The volume rneamzrik tank was made of Inconel formed into a cylinder %inch
24、es inside diameter and ll inches deep. An air-actuated piston was used to open or close a plunger valve at the bottom of the tank. Electric contact points at measured depths in coqjunction with an electric stop clock determined the rate-of volume flow through the test sections. During oseration, the
25、 pluhger valve was closed, and as the liquid level rose, the time require$ to fill the tank level reached the upper contact point, the plunger valve was opened. Thermocouples located in the tarnk recorded the fluid temperature. From 3 .I between the electric contact points was determined. When the l
26、iwid I. the volume-flow rate and the density of sodium hydroxide at the measured - temperature, the weight-flow rate was calculated. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 NACA RM E52L09 . . Electric system. - Electric power. from a 208-vo
27、lt, 60-cycle line was supplied to the .heating test section, as shown in figure l(b), through two saturable reactors and a power transformer. A voltage regulator controlling the direct-current supply to the reactor main- tained close voltage. regulation at the primary of. the pow- transformer for an
28、y setting (heating-test-section input) of the variable trans- ,former control. The capacity of the electric equipment was -25 kilovolt- amperes at -a-maxirmun input of 10 volts across the heating test section. The power input was .read from and the specific heat, from reference 3. The thermal conduc
29、tivity k was assumed to be 0.6 Btu/(hr)(sq ft)(?F/ft) for the range of temperatures encountered in the investigation as reported in - reference 3. (33 co 0 N i The physical properties in this report are evaluated at the average bulk temperature of the fluid. FESULTS AWD DISCUSSION The basic data obt
30、ained in this investigation for heating and cool- ing are listed in tables I and 11, respectively. Water heat-transfer tests. - Water heat-transfer tests were run before and after the sodium hydroxide data were obtained. The same heat- ing test section was used for all the runs. The results are show
31、n in figure 6, where the Nusselt number divided by the Prandtl nuniber raised to the 0.4 power Nu/Pr0.4 is plotted against the Reynolds nmber Re. Water data obtained both before and after the sod.ium hydroxide tests are shown. There are three sets of water data shown: one for water runs before any s
32、odium hydroxide was used, a second for water runs after the first set of sodium hydroxide runs was made, and a third for water runs after the second set of sodium hydroxide runs was made. Fairly good agreement between the water data and the McAdams correlation line (ref. 4) was obtained for all thre
33、e sets of data, with the water data having a slightly greater slope. Also included for comparison is the correlation line obtained in reference 5. The water data show very good agreement with the correlation line of reference 5. These data indicate that the instrumentation of the heating test sectio
34、n is fundamentally correct, and furthermore, that the inner-tube wall surface was not affected, so far as heat transfer is concerned, by exposure to the flowing sodium hydroxide in that the water heat-transfer data were not altered. No water heat-transfer tests were made with the cooling test sec- t
35、ion because of the relatively small temperature difference between the heated water and the cooling air. This would result in an extremely- small temperature drop of the water, and hence make the accuracy of any data obtained very doubtful. Heat balance. - The heat balance for the water heat-transfe
36、r data . checked within 5 percent. The heat balance with sodium hydroxide for the heating test section is shown in figure 7, where the heat transferred to the sodium Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-8 NACA RM E52L09 hydroxide Q as dete
37、rmined from the flow rate, specific heat, and increase in total temperature is plotted against-the total electric heat input Qp minus the external heat loss %. A 450 line is drawn to represent the perfect heat-balance condition. The external heat loss was determined by mpply5ng various mounts of par
38、er to the test section with no sodium hydroxide flowing. The power input-for a given average tube-wall temperature was considered to be the external heat loss for the same average wall temperature with sodium hydroxide flowing through . the, tube. All the data, with the exception of ,two points, hav
39、e a maxi- mumdeviation of 10 percent from the line. No attempt was made to obtain a heat balance on the cooling test section, since the cooling-air temperature andweight flow were not measured. Tube-wall temperature distribution. - Representative axial tempera- ture distributions of outside-tube wal
40、l temperatures are shown in fig- ure 8(a) for the heating test section. The-slight irregularity at the center and the drop off at the ends are caused by the comparatively large mass of metalin_the electric- flanges. Representative axial temperature distributions of the buried-tube walltemperatures a
41、re shown in fig- ure 8(b) for the cooling test section. Some of the thermocouples were inoperative for the runs shown, but a sufficient number remained to give a smooth temperature distribution curve . Sodium hydroxide heat-transfer data for heating test section. - The sodium hydroxide heat-transfer
42、 data are correlated by the familiar Nusselt relation, where the Nusselt number divided by the Prandtl number raised to the 0.4 power Nu/ProD4 is plotted against the Reynolds number Re. A plot of Nu/- 4 against Re is shown in figure 9 (a) for the heating data. 1ncluded.for comparison are the McAdams
43、 correla- tion line and tho. s.adium hydroxide correlation line obtained from refer- ence 3. It is seen that the data fall approximately 20 percent above the McAdams correlation line and about 33 percent above the line from reference 3. The sodium hydroxide data were run in two sets, and the - plot
44、indicates that there.is no apparent change in the heat-transfer coefficient with time. Since a constant value of thermal conductivity was used in the calculations, the correlation may be altered as new physical property data are. made available. Sodium hydroxide heat-transfer data for coaling testse
45、ctioa. - A plot of Xu/Pro-4 against Re ia shown in figure.g(b) for the cooling data. The McAdams correlation line is included for comparison. The bulk of the data agree fairly weli with the McAdams line; however, several scattered points are high. A comparison of figures 9(a) and 9(b) shows that the
46、 heating data are higher than the cooling data but have the same slope. r - c- . “ .“ c Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA RM E52L09 9 * It should be noted that when the buried thermocouple was positioned in the cooling-test-section
47、 wall, the exact Location could not be deter- was removed from the test setup, the thermocouple locations were sec- tioned, and. the junction location was measured as accurately as possible. .) mined. Upon completion of the tests, however, the cooling test section M where the product of the heat-tra
48、nsfer coefficient and diameter hD is plotted against the product of the mass flow per unit cross-sectional area and the diameter GD. Included for comparison are the water data and the sodium hydroxide data, as well as the lines representing the water and sodium hydroxide data as determined fromthe M
49、CAdams equation. The sodium hydroxide data fall below the water data, and the data for a 50 percent by weight solution of sodium hydroxide and water fall con- siderably below both the water and sodium hydroxide data. A plot of this type was used because the thermal conductivity of the aqueous solu- tion was not known. The experimental setup was not designed to test this solution. A considerable amount of scatter is present in the data, an
