1、4678 Convective Evaporation on Plain Tube and Low-Fin Tube Banks Using R-I23 and R-134a Liang-Han Chien, Ph.D. ABSTRACT This experimental study investigates the convective evap- oration heat transfer in a tube bundle. Heat transfer experi- ments wereperformed using R-l34a and R-123 on aplain tube an
2、d ajn tube having 15.9 mm (0.626 in.) OB. over a range of vapor qualities (0.03-0.34) with low mass velocities (8 to 40kg/m2s, or 1.6 to 8.2 lbmys) for a wide range of heatflux. Pool boiling data were also tested in the same apparatus. The jin tube havingjns O. 6 mm (O. 024 in.) high, with O. 6 mm (
3、O. 024 in.)jnpitch and 0.3 mm (0.012 in.)jin thickness, results in up to 170% boiling performance enhancement over the plain tube. The present experimental data were compared with corre- lations using superposition and asymptotic models. The super- position-type correlation provided better predictio
4、n than the asymptotic model. INTRODUCTION Convective boiling on tube banks is an important heat transfer mode in a flooded evaporator. The study of convective evaporation on tube banks is an important foundation in the design of flooded evaporators in chillers. Casciaro and Thome (2001) and Browne a
5、nd Bansal (1999) have surveyed the previous work on flooded evaporators, including experimental work and predictive models. Many researchers have tested the convective boiling heat transfer performance of tube banks in various conditions. For example, Comwell and Scoones (1988), Jensen and Hsu (1987
6、), and Webb and Chien (1994) have tested tube bank performance for R-113 and R-123. Some researchers have also provided correlations to predict the heat transfer performance in two-phase conditions. Gupte and Webb (1992) surveyed correlations for prediction of J.4. WU convective vaporization in tube
7、s and tube banks. The phenom- enological model results from combining the nucleate boiling and convective terms. In general, this may be written as = (hnb)n + (hcv)? l?n ? (1) which consists of nucleate boiling (hnb) and convective (hcv) contributions. Two main types of correlations have been used.
8、If n = 1, Equation1 is called the ?superposition model?; if n 1, it is called the ?asymptotic model.? Chen (1966) proposed the superposition model and argued that the flow velocity suppresses nucleate boiling. Hence, he proposed that the nucleate boiling heat transfer in two-phase flow be calculated
9、 by hnb = Shnbp, where hnbp is nucleate pool boiling, and Sis the suppression factor (O 1) in the asymptotic model inherently accounts for suppression by inhibiting the smaller of the contributing components in the region between the asymptotic limits of h, and h, The present work provides more expe
10、rimental data of tube bundles of plain and low-fin tubes using R-l34a and R- 123. R-123 is tested at 15C and 30C saturation temperature, and R- 134a is tested at 10C saturation temperature. The low- fin tube has 0.6 mm (0.024 in.) height, 0.3 mm (0.012 in.) thick fins, and 0.6 mm (0.024 in.) fin pit
11、ch. The asymptotic model and superposition model were compared with the present data. EXPERIMENTAL DESIGN Experimental Apparatus A diagram of the test facility is shown in Figure 1. Refrig- erant of known vapor quality enters from the bottom of the test section, where a bundle of tubes has a 15.87 m
12、m (0.625 in.) outer diameter and 9.2 mm (0.362 in.) inner diameter with 23.85 mm (0.94 in.) tube spacing. The evaporated two-phase mixture enters a condenser, and then condensed liquid flows into a receiver. The condenser is cooled by glycol water, which was circulated between the condenser and a co
13、nstant temper- ature bath. The glycol water tank is maintained at a constant temperature by an R-22 chiller. Following the receiver, a gear pump (Cole-Parmer variable flow drive P-75225-00 equipped with 07003-04 pump head) is used to circulate the refrigerant. A turbine flow meter (Cole-Parmer U3224
14、9-00: flow rate = 0.3 to 3.0 GPM, accuracy = 1% of reading) is connected after the pump to measure the flow rate. After passing through the flow meter, the liquid enters a pre-heater, where a given amount of heat is supplied. The maximum heating power of the preheater is 4.2 kW, and its heating powe
15、r is controlled by a variac. The fluid temperatures before and after entering the preheater were measured for the calculation of the vapor qual- ity at the inlet of the test section. Figure 2 shows the details of the test cell made from 20 mm (0.78 in.) thick stainless steel plates. The rectangular
16、internal space of the test cell is 150 mm (5.9 in.) high, 71 mm (2.8 in.) wide, and 120 mm (4.7 in.) long. A sight glass is made on one side of the test cell to observe the liquid level and boil- ing phenomena during the test. Tubes are soldered to the brass end flange to form a five-row staggered a
17、rray in an equilateral triangular pitch of 23.85 mm (0.94 in.). An O-ring is inserted between the end flange and the test cell for sealing. All tubes are 120 mm (4.7 in.) long, and have a 15.87 mm (0.625 in.) outside diameter and 9.2 mm (0.362 in.) inner diameter. Except for the tubes in the top and
18、 bottom rows, 9.1 mm (0.358 in.) diameter 250 W cartridge heaters were inserted in these tubes. The total length of the cartridge heater is 100 mm (3.94 in.), but the actual heated region is 60 mm (2.36 in.) long, located in the middle ofthe heater. The heat flux ofthe test tube is calculated based
19、on the actual heated length ofthe heater and outer diameter of the tube. The heat input of the cartridge 6 sight glass atemperature w valve Figure 1 Diagram of tube bundle test facility. Sight glass Insuiiented tube Figure 2 Cross section of tube bundle test cell. ASHRAE Transactions: Research 102 T
20、hermocouple I I I I -ring Cartridge Heater End View U Side View Figure 3 Test tube instrumentation and heater assembly. heater in the test tube was controlled by a variable power trans- former, and the current and voltage were measured to deter- mine the heating power. Saturation pressure and temper
21、ature are measured at the top (Psi, Ts,) and bottom (Ps2, Ts2) of the tube bundle, and the saturation temperature is calculated at the test tube location by interpolation between the saturation temperature at the top and bottom (Tsl and Ts2). During the tests, the difference in the saturation temper
22、atures at the top and bottom is less than 05C (0.9“F). The Cole-Parmer 68001-24 pressure transducers (pressure range O to 100 psig; accuracy 0.028% of 100 psig) were used for the R-134a tests, and the Cole Parmer 07356-50 pressure transducers (pressure range O to 30 in.-Hg; accuracy 0.04% of 30 in.-
23、Hg) were used for R-123 tests. The mass velocity (G) is calculated based on the minimum flow area in the tube bundle. Heat is supplied through 250 W cartridge heaters inserted in all tubes in the test cell except the top and bottom rows. The function of unheated tubes is to mix the inlet flow from t
24、he preheater. The instru- mented tube is located at the center of the third row from the bottom. Figure 3 shows the end view and the side view of the instrumented tube. The instrumented tube is fixed on the end flange with two 3.0 mm (O. 1 18 in.) thick screws. An O-ring is placed between the tube a
25、nd the end flange for sealing. The tube is designed with the precise length to fit in the distance between two flanges, and the location of the O-ring is carefully designed to ensure proper compression of the O-ring. Two diametrically opposite axial grooves of 0.6 mm are made along two opposite side
26、s of the test tube. The center of the groove is 0.6 mm (0.024 in.) away from the tube wall of the plain tube or 0.6 mm (0.024 in.) from the root of the fins on the fin-tube. T-type thermocouples of 0.5 mm (0.02 in.) sheath diameter are inserted in the grooves for tube wall temperature measurement. T
27、he one-dimensional steady-state heat conduc- tion equation in cylindrical coordinates is used to correct for the conduction temperature drop between the thermocouple and the boiling surface. The grooves are located at the top and bottom positions in the tube wall. A heat sink compound is used to ens
28、ure good thermal contact of heaters and thermo- couples with the tube wall. Data Reduction and Experimental Uncertainty The temperature and pressure of the refrigerant entering the preheater were measured to determine the thermodynamic state of the subcooled refrigerant. Knowing the heat input to th
29、e preheater, the mass flow rate, and the thermodynamic properties, the thermodynamic state of refrigerant leaving the preheater is calculated assuming an isobaric process. By knowing the pressure at the bottom of the tube bundle, the vapor quality can be calculated. The vapor quality in the test is
30、calculated at the center position of the test tube from the heat input up to that point. All instruments are connected to an Agilent 34970A data acquisition system and transmit data through RS-232 interface to a personal computer. The thermocouples were frequently calibrated in a constant temperatur
31、e bath and are repeatable within fO.l“C (0.1 8F). Based on the instrumentation accu- racy of the heat flux, saturation pressures, and wall tempera- ture, an error analysis shows that heat transfer coefficients are calculated within 2.5% for the maximum heat flux condition and within 7.1 % for the mi
32、nimum heat flux condition for R- 134a. Similarly, the uncertainty is within 1.1 % for the maxi- mum heat flux condition and within 2.5% for the minimum heat flux condition for R-123. EXPERIMENTAL PROCEDURE Convective vaporization of a tube bundle and pool boiling data of a heated tube in a tube bund
33、le were tested in R-134a and R-123 boiling on a plain tube and a fin tube having 0.6 mm fin height, 0.3 mm fin thickness, and 0.6 mm fin pitch. Data were taken for three saturation temperatures: 10C (50“F), 15C (59“F), and 30C (86F) for various mass velocities, heat fluxes, and vapor qualities at th
34、e test section. The heat flux was varied decreasingly to avoid boiling hysterysis. The instru- mented tubes were cleaned with acetone and pure water before each test. Then they were immersed in de-ionized water for ten minutes and then dried by blowing air. The pool boiling test was performed accord
35、ing to the following procedures: A leak-tight test is performed before charging. The sys- tem was considered leak-tight if the system was able to hold vacuum (less than 5 kPa) for at least 24 hours. Evacuate the system for at least twenty minutes by a vacuum pump. Charge the working fluid into the s
36、ystem. About 22 kg of refrigerant is needed. Turn the heater to the maximum output (about 250 W). Keep the pool temperature at the desired saturation tem- perature by adjusting the cooling water chiller system and preheater and maintain for at least an hour. Then, evacuate the system for some three
37、seconds to remove the noncondensable gases. Keep the system boiling at the maximum heat flux for at least thirty minutes, and adjust the glycol water system and preheater heat input to keep the saturation tempera- ture at the desired value. ASHRAE Transactions: Research 1 03 Make sure that the liqui
38、d level is above the instrumented tube. Record the data after the system temperature is steady for at least five minutes. Stepwise reduce the heat input and repeat step 5 and 6. The test procedure of convective boiling is similar to the above procedures for pool boiling except that the pump was turn
39、ed on, and varied after every set of heat flux and vapor quality at a fixed flow rate was taken. EXPERIMENTAL RESULTS The pool boiling data of R-123 at 15C and 30C are shown in Figures 4 and 5, respectively. Predictions of Cooper (1 984) and Gorenflo (1 993) correlations are compared with *,A R-123
40、Ts=l 5C . m-. Plain Tube Data -O- Cooper Correlation R,=1.0 A.,n. _,.- -A- Gorenflo Correlation ,_,. ,e. I I I I I . 10 20 30 40 50 q“(kW/m2) Figure4 Pool boiling data of the plain tube at 15C saturation temperature in R-123. R-123 TI=3O0C O Cooper Correlation R,=l .O Gorenflo Correlation Plain Tube
41、 Data 4000 50001 - I I l I I 10 20 30 40 50 q“(kW/m2) Figure5 Pool boiling data of the plain tube at 30C saturation temperature in R-123. the present experimental data for plain tube (square symbols in Figures 4 and 5). For the curves of the Cooper correlation in Figures 4 and 5, the authors set the
42、 surface roughness parameter Rp = 1 .O. As shown in Figures 4 and 5, the present data are in good agreement with the Cooper correlation at both 15C and 30C. The Gorenflo correlation overpredicts the present R- 123 plain tube data at both 15C and 30C for heat flux greater than 20 kW/m2. Figure 6 show
43、s the pool boiling data of R- 134a at 10C. The Gorenflo (1 993) correlation and the Cooper (1984) correlation with Rp = 1 .O and 1.4 are shown in Figure 6 for comparison. The present plain tube data are in better agreement with the Gorenflo correlation at 10,20, and 30 kW/m2 but are in better agreem
44、ent with the Cooper corre- lation for heat flux q“ = 40 and 50 kW/m2. The slope of the R- 134a data does not agree with the predictions of either the Cooper or the Gorenflo correlations. Figure 7 shows the test results of convective evaporation for a plain tube in R-l34a, in the form of h versus hea
45、t flux (4“). The data shown in groups for 10, 20, 30, 40, and 50 kW/m2 heat flux were taken for vapor qualities (x) between 0-0.34. For a fixed vapor quality, the heat transfer coefficient increases with increasing heat flux, but the change of mass velocity (G) from 10 kg/m2s to 40 kg/m2s does not m
46、ake significant difference. However, one expects that the mass velocity will influence the convective evaporation perfor- mance at a higher mass velocity. The data in Figure 7 also show that the heat transfer coefficient increases as the vapor quality increases for a fixed mass velocity (9. For the
47、same mass velocity, G = 10 kg/m2s, the heat transfer coefficient increases as the vapor quality increases from 7% to 34%. Similarly, for mass velocity G = 25 kg/m2s, the heat transfer coefficient increases as the vapor quality increases from 7% Figure 8 shows the convective evaporation test results
48、of R- 123 in the same format. In addition to the plain tube data, the , to 14%. - R-I 34a TS=l 0C 9000 - 8000 - . .-O,-. Cooper Correlation R,=1.4 . -A- Cooper Correlation R,=1.0 7000- - Gorenflo Correlation I, ,. Plain Tube Data /- 2000- 10C lo00 ! 8 I I 8 1 4 10 20 30 40 50 q“(kW/rn2) Figure 7 Tes
49、t results of convective boiling in R-134a at 10C saturation temperature. Figure 8 Test results of convective boiling in R-123 at 15C and 30C saturation temperatures. results of the low-fin tube are also shown in Figure 8. The plain tube data were taken at 15C or 30“C, and the fin tube data were taken at 15C only. For the same test conditions (e.g., G = 1 O kg/m2s, x = 6%), the fin tube yielded 90% to 170% greater heat transfer coefficient than the plain tube. Similar to the plain tube, the fin tube yielded a greater heat transfer coe
