1、4670 A Comparative Study of Shell-Side Condensation on Integral-Fin Tubes with R-II4 and R-236ea Wade W. Huebsch, Ph.D. ABSTRACT A test facility was constructed to perform shell-side condensation testing, and test results are presented for CFC- 114andHFC-236eausingapIain, a 1024 fpm, anda 1575fpm tu
2、be surface. Condensation coejcient data are presented for both saturated and superheated refrigerant vapor on a single test tube. The heatjlux range for the data was 15,000 to 40,000 W/m2 (4.8 x Id Btdhfi-12.7 x Id Btuh fi) at a saturation temperature of 40C (104F). The results of this study show th
3、at the two refrigerants produced similar performance char- acteristics in shell-side condensation on integral-Jin tubes. The data also show that the 1024fpm tubeperformedslightly better with R-236ea than with R-114. The 1575fpm tube demon- strated the same trend in the higher heatjlux range. The she
4、ll- side condensation heat transfer coeflcient results for super- heated vapor were similar to the saturated results. Specjcally, thesuperheated vapor data for thefinnedtubes were within 7% of the saturated vapor results. INTRODUCTION There is continued interest in every aspect of the refrig- erant
5、industry to discontinue the use of chlorofluorocarbons (CFCs) as the refrigerant of choice. The movement is to incor- porate refrigerants, which are termed HFCs or hydrofluoro- carbons, that do not contribute to ozone depletion and minimize global warming. The United States Navy has, in the past, us
6、ed CFC-114 (also designated as R-114) as the working refrigerant in shipboard and submarine chiller units. With the mandatory phase-out of CFCs as dictated by the Montreal Protocol and national policy, it was imperative for the Navy to find a replacement that is environmentally safe and performs in
7、a similar fashion to CFC-114 in the cooling systems. CFC- M.B. Pate, Ph.D. Member ASHRAE 114 has the characteristics of being a moderate-pressure refngerant, more stable than other refrigerants with tempera- ture and when exposed to water vapor, and with very low toxicity. The Navy required that the
8、 alternative refngerant should have attributes similar to CFC-114, less the ozone-harming characteristics. Following preliminary evaluations, there were two leading candidates that appeared to be potential replace- ments for R- 1 14: HFC-236ea and HFC-236fa (also designated R-236ea and R-236fa). Eve
9、n though experiments were performed on both of the alternative refrigerants, only the results for the HFC-236ea are presented and compared to CFC-114. The results for HFC-236fa will be presented in a later paper. The heat transfer results for HFC-236ea show that, for existing systems, there would be
10、 no performance loss with a refrigerant conversion and, for new systems, the data can be used to size and design the heat exchangers. For single-tube condensation testing, there are five major factors that can impact the heat transfer: (1) tube geometry, (2) fluidproperties, (3) superheated vapor, (
11、4) vapor shear, and (5) fluid contaminants (i.e., oil, noncondensible gas, etc.). These factors can either enhance or degrade the heat transfer perfor- mance of a given refrigerant. This research effort concentrates on the first three factors. The tube geometry variable was investigated by the use o
12、f three different tube surfaces, while the fluid properties variable was investigated by the two differ- ent refrigerants tested. Examining the effects of superheated vapor was accom- plished by comparing the condensation heat transfer coeffi- cients to those obtained in condensation of saturated va
13、por. It has been shown that calculating the average heat transfer coef- ficient using the temperature difference between the saturation temperature of the superheated vapor and the surface temper- W.W. Huebsch is an assistant professor in the Mechanical and Aerospace Engineering Department, West Vir
14、ginia University, Morgantown, W.V. M.B. Pate is a professor in the Mechanical Engineering Department, Iowa State University, Ames, Iowa. 40 02004 ASHRAE. Thernocouple Pressure Tap Thernistor Pressure Transducer I Fliter Drier Refrigerant Punp Figure 1 Schematic of condensation test facilis? ature re
15、sults in negligible error (McAdams 1954). There have also been published results from Goto et al. (1 980) for conden- sation of R- 1 13 superheated vapor. The results show that the heat transfer coefficient is only increased by about 5% for the superheated vapor as compared to saturated results. The
16、refore, it was concluded that the effect of superheated vapor on heat transfer is small. There has been a great deal of research performed in the area of condensation on integral-fin tube surfaces. Typically, these enhanced surfaces are only used to condense refrigerants that have a relatively low s
17、urface tension. High surface- tension fluids, such as water and ammonia, use plain tube surfaces for condensation. Karkhu and Borovkov (1971) conducted experiments on condensing CFC- 1 13 and steam on four different tubes with trapezoidal fins. The measured heat transfer coefficients were between 50
18、% and 100% higher than for a smooth tube. Marto (1988) conducted a literature review for film condensation on integral-fin surfaces for both single tube work as well as tube-bundle experiments. The conclusions from the survey that are relevant to the present work are that fin spacing is a critical v
19、ariable and the condensate flooding significantly affects finned tube performance. The optimum fin spacing for a given working fluid increases as the ratio of the surface tension to density increases. Masuda and Rose (1985) performed tests on ethylene glycol in addition to CFC-113. These results sho
20、wed that the maximum enhancement was 4.7 at a spacing of 1 .O mm. The three fluids (CFC-113, ethylene glycol, and steam) used above were chosen because they offered a wide range in To DC Rectifier L the surface tension-to-density ratio (dp). The results indicate that the maximum heat transfer enhanc
21、ement increases and the optimum fin spacing decreases as the ratio dp decreases. Marto (1988) also showed that condensation testing of integral-fin tubes in a single-tube test facility accurately models the performance of a tube bundle (this work used a single-tube test facility). These results were
22、 confirmed by Webb and Murawski (1990), who showed that integral-fin tubes have negligible inundation effects during condensation within a tube bundle and, therefore, have similar performance characteristics to those found for a single-tube test. This paper focuses on comparing the shell-side conden
23、- sation performance of HFC-236ea and CFC-114 in a single- tube test environment. The tube surfaces that are used in this condensation performance analysis are the plain, the 1024 fpm, and the 1575 fpm tubes. Even though the plain tube is rarely used for industrial applications, it is useful to incl
24、ude this surface as a baseline reference. TEST FACILITY A test facility was constructed to study shell-side film condensation on a single, horizontal tube. This paper docu- ments the first condensation results obtained from this test facility. Therefore, a thorough test facility description is inclu
25、ded. The facility is capable of producing either saturated vapor or superheated vapor for condensation studies. A sche- matic of the test facility can be seen in Figure 1, and a detailed description is presented below. ASHRAE Transactions: Research 41 Table la. Tube Geometry Specifications in SI Uni
26、ts Do Nominal, Di Nominal, Tube mm mm plain 19.4 17.3 An An Ai or9 Fin Height, Nominal, Actual, Nominal, mm mm m2/m m2/m m2/m 19.4 0.0512 0.0512 0.0456 1024 fprn 18.8 14.3 15.9 1.45 0.0588 0.195 0.0448 1 Table 1 b. Tube Geometry Specifications in I-P Units 1575 fpm 18.8 15.6 17.1 0.86 0.0593 0.197 0
27、.0488 Tube plain 26 fpi Test Section The test section is constructed of stainless steel with a diameterof 101.6mm(4in.)andalengthof406.4mm(16in.). There are two glass-view ports (one per side) in the middle of the test section. At 45 degrees off top dead center, along oppo- site sides, are the two r
28、efrigerant vapor inlets. Two liquid outlets lie along bottom dead center of the test section. There is a stainless-steel tube sheet attached to each end of the test section; each has two threaded ports to allow for the insertion of the copper test tube. A drip-deflector was also installed in the tes
29、t section, which consists of a thin aluminum sheet placed directly below each vapor inlet port. The test section is equipped with two thermistors, one per side placed on oppo- site ends, that measure the refrigerant-vapor space tempera- ture. A pressure tap is also installed that is connected to a p
30、ressure transducer. This allows for measurement of the satu- ration pressure within the test section, which can be used to calculate a saturation temperature. Test Tubes Three different tube surfaces were tested: plain, 1024 fpm (26 fpi), and 1575 fpm (40 fpi) tube, which have a nominal outer diamet
31、er (OD) of 19.1 mm (0.75 in.). There is 101.6 mm (4 in.) of plain tube on each end that is located outside of the test section (see Table 1 for the exact dimensions on each test tube). These plain tube sections make up the water inlet, outlet, and turn-around. There were two test tubes (419.1 t r26i
32、piTube-Rwl i26 Tube - Rcm 2 -40fp1 Tute - Rcm 1 I x 40 fpi Tube - Run 2 4 0-7 - O 2000 4WO 6000 8000 10000 12WO 14000 Hwt Flux (Btuin n) Figure 6 presents the repeatability analysis for condensa- tion of CFC-114 on the 1024 fpm and 1575 fprn tubes. The repeat run for the plain tube was within 3% of
33、the original data but is not included in this plot. The deviation between runs for the 1024 fprn tube is less than 2%. The 1575 fpm tube shows a deviation of 4% at the lowest heat flux and less than 2% at 40 kW/m2 (12.7 x lo3 Btu/hft2). Therefore, all three tubes showed excellent repeatability withi
34、n the tested heat flux range. It should be noted that the deviations found in this repeatability analysis are less than the uncertainties in the calculation of the shell-side heat transfer coefficient (Table 4). Repeatability testing was routinely performed during this study for both refrigerants an
35、d all tube surfaces tested. The 1024 fpm and 1575 fprn tube data were repeatable within 3% and 2%, respectively, for HFC-236ea. ASH RAE Transactions: Research 47 The condensation coefficient as a function of the heat flux for HFC-236ea is shown in Figure 7. As was the case with CFC-114, the two finn
36、ed tubes outperformed the plain tube. But, for HFC-236ea, the 1575 fpm tube outperforms the 1024 fprn tube only at the low end of the heat flux range. At 40 kWl m(12.7 x 103Btu/h-ft2), ho forthe 1024fpmtubeis5%higher than for the 1575 fprn tube. Since the value of 5% is within the uncertainty of the
37、 tubes at this heat flux, it is evident that they have similar performance in condensation of HFC-236ea. Figure 8 summarizes the condensation results for the inte- gral finned tubes. Several findings can be drawn from this figure: (I) the 1024 fprn tube produces higher condensation coefficients for
38、HFC-236ea than for CFC-114, (2) the heat transfer performance for the 1575 fprn tube is similar for both 2 1 8 ., R-11411024fpm Tute eR-114/ 1575fpmTute mR-PSeallO24 fpm Tube x R-Pea I 15/5 (pm Tube B NO R-114 /40 Tute x R-23Bea I $ -p Tu-te . O 5 10 15 20 25 30 35 40 45 Heat FIux(kWm) refrigerants
39、in the lower heat flux range and slightly higher for HFC-236ea in the higher heat flux range, and (3) the HFC- 236ea heat transfer coefficient data show a slightly greater increase with increasing heat flux than data for CFC-114. But, overall, any combination of the integral-fin tubes and the two re
40、frigerants produce similar heat transfer coefficients. Superheated Vapor Condensation data were also taken for superheated vapor of CFC- 1 14 on all three tube surfaces. Certain industrial appli- cations require condensation of superheated vapor and, there- fore, it was of interest to examine the ef
41、fect on the heat transfer performance. McAdams (1954) showed that the heat transfer coefficient for condensation of superheated vapor was within I 1400, I 1 NO o D m* . Figure 7 Condensation heat transfer coeficient for R-236ea at TSat = 40C. O WOO 4Mo 6wx) 8wo IWO 12M 14000 Heat Flta Buh fi Irno 14
42、00 I -1 MOI I est FIUX (i 1024 and o/p 1 x io4. In the present study, the surface tension-to-density ratio for both refrigerants falls within the requirements for both correlations. The 1024 fprn tube falls on the border of both, while the 1575 fprn tube matches the RW requirements. In the present s
43、tudy, it was found that the BK model outper- formed the RW model. The authors believe there are several possible reasons for this difference. The combined refnger- ants and tubes used in this study generally fall in the gray area of both correlations. For example, in Rudy and Webb (1983), for the 10
44、24 fpm tube with R- 1 1, the RW correlation predicted the condensation coefficients within 25% of the experimental data. For the 1024 fpm tube with R-114 in the present study, the BK model predicted the experimental data within 12.5%. A second possible reason would be the trend of increasing condens
45、ation coefficient with increases in heat flux for the experimental data. Both correlations (BK and RW) predict that the condensation coefficient will decrease as the heat flux increases. All of the experimental data in this work shows a slight increase in heat transfer with increasing heat flux. The
46、refore, neither correlation is going to be able to predict this data to a high degree of accuracy. In this case, the BK model happened to produce better results for these test cases and has been used for the comparison. For comparison between the predicted and measured results, the predicted heat tr
47、ansfer coefficient from the corre- lation is plotted as a function of the measured heat transfer coefficient. Figure 10 shows the plain tube results and the Nusselt correlation results for CFC-114 and HFC-236ea. For CFC-114, the correlation predicts the data within 53% for the heat flux range that w
48、as tested. The discrepancy between correlation and measured condensation coefficients was greater for HFC-236ea, but the deviation is still within 13%. Overall, the Nusselt correlation does a reasonable job in predicting the heat transfer coefficients for condensation of both refrigerants on a plain
49、 tube surface. Figure 11 shows the correlation comparison for the 1024 fprn tube surface with CFC-114 and HFC-236ea. The corre- lation predicts that the condensation coefficient would decrease with an increase in heat flux, while the experimental values show a slight increase with increased heat flux. This produces the mild negative slope when plotting predicted versus measured heat transfer coefficients. For CFC-114, the correlation still predicts the condensation coefficients to within +12.5%. The predicted value and the experimental value for ho are approximately equal
