1、I n t e r n at ion a 1 J o u i- n a 1 of Heat in g ,Ve n til a t in g, Air-conditioning and Refrigerating Research HVAC mr may any pan of this book be reproduced, stored in a retrieval system, or transmitted in any form or by any mean-lectronic, photocopying, recording. or other-without permission i
2、n writing from ASHRAE. AbstrPcteAbsuacted and indexed by Engineering Information. Inc. Available electronically on Compendex Plus and in print in Engineer- ing Index. Information on the contents arc also presented in the follow- ing IS1 products: SciSearch, Research Ale* and Current Contents/ Engine
3、ering, Computing, and Technology. Diselahe-ASHRAE has compiled this publication with care. but ASHRAE has not investigated, and ASHRAE expressly disclaims any duty to investigate. any product the boiling curve shifted left with decreas-ing thickness. They also found that use of alternating current i
4、nstead of direct current resistive heating caused the boiling curve to move to the left. Despite the interesting findings, the authors did not provide any rationale for the trends or causes of the results.Results of different studies for boiling heat transfer of R-134a over Turbo-B tube are depicted
5、 in Figure 1. Palm (1995) investigated boiling performance of Turbo-B tube with R-134a using the electrical heating method. The tube diameter was 19 mm and tube length was 250 mm. Operating saturation temperatures were 0.7C and 20.4C. His data for the operating tempera-ture of 0.7C are shown in Figu
6、re 1. Oh and Kwak (1996) employed the water heating method to experimentally investigate the effect of a direct current electric field on nucleate boiling heat transfer for refrigerants R-11 and R-113 in a single-tube shell/tube heat exchanger. Even though they were successful in showing the enhance
7、ment in heat transfer due to the application of high voltage electric field, in the absence of the electric field the boiling heat transfer coefficient was almost 50% lower than the resistive heating values found by both Marto and Lepere (1982) for refrigerant R-113, and Papar (1993) for refrigerant
8、 R-11. The authors acknowledged that their data is lower than others are, but did not explain the cause. Figure 1. Comparison of data on Turbo-B tube with R-134a0 10 20 30 40 50 60 70 q“ (kW/m)0 1 2 3 4 5 Superheat (C)R-134aPool BoilingTurbo-B TubeThors (1994)Water heatingTsat=14.6 CWebb (1991)Elect
9、ric heatingTsat=26.7 CPalm (1995)Electric heatingTsat=0.7 CVOLUME 5, NUMBER 4, OCTOBER 1999 285Webb and Pais (1991, 1992) experimentally investigated the boiling performance of Turbo-B tube using R-134a with the electric heating method. They performed the tests at 4.4C and 26.7C. The tube diameter w
10、as 19 mm and the tube length was 165 mm. Their results for the operating temperature of 26.7C are shown in Figure 1. Turbo-B tube was tested using the water heating method for a tube length of 2.4 mm and outside diameter of 19 mm (Thors 1994). The working fluid was R-134a and the operating temperatu
11、re was 14.6C. The tube-side water velocity was 1.6 m/s. For the data obtained with fluid heating, the value of the wall superheat was averaged along the tube using average heat flux and heat transfer coefficient. From the data in Figure 1, it appears reasonable that the data of Webb falls to the lef
12、t the data from Palm, as the saturation temperature is higher in Webbs case compared to Palms case. However, the data from Thors, taken at a lower saturation temperature (14.6C) than that of Webb case (26.7C), should fall to the right of the data of Webb. This may suggest that the dif-ferences obser
13、ved are attributed to the heating method and boundary conditions that are dis-cussed later in this paper.Another comparison of heating methods is reflected in the results of electric heating for boil-ing performance of R-114 over a smooth tube by Memory et al. (1995) compared with the results of the
14、 water heating data of McManus et al. (1986) in Figure 2. Memory et al. (1995) conducted the experiments at 2.2C saturation temperature with a tube length of 450 mm and diameter of 16 mm. McManus et al. (1986) ran the experiments at 13.8C saturation temperature for a tube length of 304 mm and diamet
15、er of 16 mm. As can be seen from Figure 2, the results differ by as much as 50%. Also, the order does not match that of Figure 1. Although the satura-tion temperature is different for the two cases, it is very likely that part of the difference is due to the method of heating employed. Kedzierski (1
16、995) experimentally investigated the pool boiling performance of R-123 on four enhanced surfaces. The tubes were Turbo-BII, High Flux, GEWA-k, and GEWA-T. The sur-faces were either machined or soldered onto a flat, thick, high conductivity copper plate. He investigated the boiling performance of the
17、 tubes by electric heating as well as water heating. He observed differences between the results. Figure 3 compares water to electric heating for the three tubes tested for the heat flux range of 10 to 70 kW/m2. Kedzierski found that water heating Figure 2. Comparison of data on smooth tube with R-1
18、140 10 20 30 q“ (kW/m)8 9 10 11 12 13 Superheat (C)Memory (1995)Electric heatingTsat= 2.2 CMcManus (1986)Water heatingTsat= 13.8 CR-114Pool BoilingSmooth Tube286 HVAC&R RESEARCHin most cases resulted in higher heat transfer coefficients compared to electric heating. The high-est case was for the GEW
19、A-K tube, in which water heating resulted in as much as a 32% greater heat flux compared to electric heating at a heat flux of 35 kW/m2.In an effort to explain the difference between water heating and electric heating, Kedzierski postulated that for the same time-averaged heat flux, a larger fractio
20、n of heat was used to super-heat liquid in the electric heating method than in the water heating method. He approximated the transient surface temperature of the heating plate as a square wave, which was low for boiling and high for liquid superheating modes. A thin penetration depth in the wall nea
21、r the boiling surface was defined to explain the transitional behavior of the plate temperature, defining an inner wall temperature Twiat the lower edge of and an outer wall temperature Twoat the upper edge of , as shown in Figure 4. The explanation was that during boiling the outer wall tempera-tur
22、e Twodropped, since it was a more efficient means of transferring heat than natural convec-tion (therefore having a square wave profile). The inner wall temperature was constant for the water heating method but varied for the electric heating method in phase with the same ampli-Figure 3. Comparison
23、of water to electric heating for three tubes (Kedzierski 1995)Figure 4. Speculative representation of temperature variation of the tube surface (Kedzierski 1995)VOLUME 5, NUMBER 4, OCTOBER 1999 287tude as in the outer surface temperature Twodue to the constant heat flux constraint. The analysis was
24、shown in the following equational form:(1)Where A is the amplitude of the assumed square wave temperature profile along a period of a timed average heat flux, and k is the thermal conductivity. The above equation shows that water heating superheats the liquid less than electric heating by the amount
25、 of Ak/2. In conclusion, for the same time-averaged heat flux, the boiling curve for the electric heating method would fall to the right, indicating a higher heat transfer coefficient for the water heating method.The above analysis explains some of the discrepancies found in the data. However, the m
26、odel of Kedzierski cannot explain all of the factors contributing to the differences found between boil-ing heat transfer for electric and fluid heating. For example, it did not explain the causes respon-sible for larger heat transfer coefficients being obtained for electrically heated data on a sho
27、rt test section as compared to averaged fluid heated data over a long test section. The model accounts for only boiling heat transfer differences between electric and fluid heating conditions on identical surfaces having the same length and operating conditions. In the present study a more encompass
28、-ing approach to explain the differences between the heating methods is provided.EXPERIMENTAL METHODSIn the present investigation, experiments were performed on the Turbo-BII tube using both electric and water heating methods. Two experimental setups were employed. The first setup was a small-scale
29、unit, which was used for pool boiling heat transfer of non-CFC refrigerants, with a resistive heating method. The schematic diagram, of the setup shown in Figure 5a, con-sists of a high-pressure boiling chamber, an air-cooled refrigeration unit, the test section, and a pressure control system. The n
30、ominal outside diameter of the test tube was 19 mm (3/4 in.) and the length was 63.5 mm (2.5 in.). An electrical cartridge heater was imbedded in the test tube to provide heat. The heater outside diameter was 6.35 mm (1/4 in.) with a heating length of 50.8 mm (2 in.). The detailed description of the
31、 setup is given in Ohadi et al. (1994).The second setup was an industrially scaled liquid-to-refrigerant loop, that was designed to address the applicability of the EHD technique for refrigerant-side heat transfer enhancement in industrial chillers. The heating was applied by a water heating method.
32、 The overall schematic of the setup is shown in Figure 5b. A horizontal shell-and-tube heat exchanger was built as the EHD-enhanced pool boiling test facility to withstand pressures up to 2 MPa (300 psi). To sim-ulate the operating conditions of practical industrial chillers, hot water passed throug
33、h the inside of the heat exchanger tubes and boiling took place on the outside of the tubes. There were two sub-loops: a hot water loop, provided the heating energy required for boiling of the refrigerant, and a cold water loop provided cooling water for the condenser loop to condense the refrigeran
34、t. Other components for the system included the instrumentation for measurement of temperature, pressure, and flow rates in the loop. A 1.87 m long stainless steel shell of 200 mm inside diameter with 8.1 mm thickness was designed to accommodate high pressure refrigerants. The two sight glasses are
35、located at the middle of the shell along its length. The test section tube was a Turbo-BII copper tube with 19 mm outside diameter and 2.87 m length. Approximately 1872 mm of the tube was located inside the shell. The part of the tube that was in the shell was enhanced surface and the developing par
36、t was plain tube.Data were taken in the electric heating setup at the saturation temperatures of 4C, 15C, and 19C for a tube length of 63.5 mm and diameter of 19 mm. The data for the water heating sys-qwqeATWk2-=288 HVAC&R RESEARCHtem were taken at the saturation temperature of 15C for the same tube
37、 diameter and a tube length of 1872 mm. Figure 6 shows the results of the tests for the common saturation tempera-ture of 15C. It is seen that the water heating data are to the left of electric heating data. In a sep-arate set of experiments, a 19 fins per inch (fpi) tube (fin spacing 1.34 mm) was u
38、sed for the boiling experiments with R-123 as the working fluid at an operating temperature of 26.7C using the water heating method. As shown in Figure 7, the data are to the right of the resistive heating results of Kumar (1993) for the same tube and saturation temperature. This trend does not matc
39、h the order in Figure 6. Figure 5a. Schematic of resistive heating setupCooling LoopStepper MotorComputerAccumulatorCompressorAir-cooledrefrigeration unitExpansionValveVertical Condensing TubesPressure TransducerReliefValveVoltmeterVariableTransformerPr essur econtr ol syst emTest ChamberTest TubeFi
40、gure 5b. Schematic of the water heating setupHeat ExchangerCold Water Loop Water inWater outChillerPumpFlowmeterR-12 Loop Control ValveHot Water Loop Refrigerant Sight glass HeaterPumpFlowmeterVOLUME 5, NUMBER 4, OCTOBER 1999 289ANALYSIS OF WATER HEATING AND ELECTRIC HEATING PROCESSESAs discussed by
41、 Dhir (1991), in a boiling heat transfer process the heat transfer coefficient strongly depends on (1) the magnitude of heat reaching the boiling surface, (2) the conduction through the thickness of the surface, and finally (3) the outer surface structure over which the boiling occurs. In the presen
42、t analysis, the comparisons between the fluid and electric heating conditions are made for identical surface structures and heat flux. Consequently, the focus of the analysis is on the conduction from the heated side of the heat transfer surface to the boiling side. In general, there are two ways th
43、at conduction and heating method can interact and cause an apparent difference in boiling heat transfer. The first conduction effect is observed for a compar-ison of local electric data to local fluid data. The local-local heat effect was observed by Figure 6. Comparison of data on Turbo-BII tube wi
44、th R-134a0 10 20 30 40 50 60 70 q“(kW/m)0 1 2 3 4 5 Superheat (C)Electric heating setup Water heating setupR-134aPool BoilingTurbo BII TubeTsat= 15 CFigure 7. Comparison of data on 19 fpi tube with R-1230 5 10 15 20 q“(kW/m)0 1 2 3 4 5 6 7 Superheat (C)R-123Pool Boiling19 fpi TubeTsat= 26.7 CKumar 1
45、993)Electric heatingPresent workWater heating290 HVAC&R RESEARCHKedzierski (1995). The second conduction effect is observed when comparing local electric data to global fluid data. Here the averaging process contributes to the differences attributing larger heat transfer coefficients to the electric
46、ally heated boundary condition. As will be seen in the following discussion, both the first and second conduction effects are linked to the temperature profile along the heated surface.Differences Due to AveragingIn order to address the difference between the two heating methods contributed by the a
47、verag-ing process, the results of electric heating (constant heat flux) were applied to water heating in a process of segmentation. This was done with the data obtained for the boiling performance of a 19 fpi tube with R-123. The length of the water heating tube (1872 mm long) was divided into 6 seg
48、ments, each approximately 312 mm. Then the results of the heat transfer coefficient obtained from electric heating were applied to each segment of the tube, assuming that the heat flux on each segment remained constant. Although this approach is not strictly correct, it is an appropri-ate approximat
49、ion.In order to obtain a heat flux for each segment of the tube, different temperature profiles were assumed along the tube for the mean water temperature. The two ends of the temperature pro-files were inlet and outlet water temperatures. A linear profile was assumed along with the two exponential profiles with two different arbitrarily selected radii of curvature. Then for each seg-ment of the tube and for each temperature profile, heat flux and heat transfer coefficient calcula-tions were performed and the arithmetic average of
