1、4781 Study of Cross-Flow Cooling and Heating of Air via an Elliptical Tube Array Mesbah G. Khan Student Member ASHRAE Amir Fartaj, PhD Member ASHRAE David S-K Ting, PhD, PEng Member ASHRAE ABSTRACT Experiments were conducted on air cooling and heating viaanarrayofl8ellipticaltubeswith their31.7mm ma
2、joraxis parallel to the air cross-ow. These tubes had a minor-to- major axis ratio of 0.3 and were evenly spaced by a 6.1 mm air gap. Water entered the bottom tube and exited at the top tube of the array, which wasplaced in a 600 mm long, 300 mm by 300 mm wind tunnel section. The temperature differe
3、nce between the mean approach air and the inlet water was main- tained at 15.7k1.5C for both cooling and heating of the air. Reynolds number, based on the mean free stream air velocity and streamwise major axis length of the elliptical tube, was varied from 1 O, 000 to 36,000. In both the cooling an
4、d heating processes, the results showed that the Nusselt number varied in a power law manner with respect to the Reynolds number. It is also observed that the heat transfer coeflcient appears somewhat higher in cooling than in the heatingprocess. The results compared well with previously established
5、 correla- tions. INTRODUCTION Heat transfer between the surface of a body and the ambi- ent fluid is of great practical interest in most thermal and chemical engineering applications. The study of the process of heat transfer between fluids and bodies is vital in addressing the engineering problems
6、associated with heating and cooling systems. Traditionally, air-to-water cross-flow heat exchang- ers, consisting of tubes of various shapes, have been used in many applications. The airside generally accounts for about 90% or more of the total thermal resistance of typical air-to- water heat exchan
7、gers (Wang 2000). Thus, a proper choice of airside heat transfer correlation in the relevant design and application is essential. To design such heat exchangers, some basic parameters to be considered are the geometry and shape of the heat transfer surface, heat exchanger size, air and water flow ra
8、tes, and tube orientation. In addition, the economic and environmental issues in todays industry place the needs for system compactness and further performance improvement in the forefront. This necessarily entails the improvement of heat transfer, minimization of pressure drop, and promotion of eas
9、y fabrication. Thus, tube or banks of tubes in cross-flow have been the focus of a large number of investigations. As system components, different body shapes and orientations have been extensively studied and utilized in various applications. Although numerous studies have been made on a single cir
10、cu- lar cylinder or an array of circular tubes, little effort has been focused on an elliptical tube or array. It is anticipated that the elliptical tube has superior combined thermal-hydraulic features over the circular cylinder in terms of (a) enhanced heat transfer rate and (b) minimized pressure
11、 drop and vortex or flow-induced vibration. Moreover, elliptical tubes can increase the system compactness because of their larger heat transfer area per unit volume compared to circular cylinders. Enhancing Heat Transfer Using Elliptical Tubes Several investigators have reported that, in general, a
12、n elliptical tube enhances heat transfer relative to a circular one. For example, Matos et al. (2001) investigated the Reynolds number-based on the length of the tube row (i.e., from 75 to 200 when based on streamwise major axis length of the tube)-where they used eliptic axis ratio (AR) of 0.75, tu
13、be length to minor axis ratio of 6.2, and constant Prandtl number of 0.72. Compared to a circular tube, their results for the same flow obstruction area showed a 13% relative heat transfer gain Mesbah G. Khan is a graduate student, Amir Fartaj is an associate professor, and David S-K. Ting is an ass
14、ociate professor in the Department of Mechanical, Automotive and Materials Engineering, University of Windsor, Windsor, Ontario, Canada. 02005 ASHRAE. 423 due to elliptic Configuration. Hiso, Sohal and OBrien (2001) found that for the same cross-sectional area, a single elliptical tube with AR of 0.
15、33 can enhance heat transfer coefficient by 2535% compared to a single circular (AR = I) cylinder depending on various design parameters. Effect of Tube Axis Ratio (AR) on Heat Transfer. The effects of tube AR on cross-flow heat transfer have been stud- ied by many researchers. One such study was ma
16、de by Eckert and Livingood (1 953) on elliptical cylinders for ARS of 0.25 and 0.50, where the major axis was used to define the Reynolds and Nusselt numbers. The results, based on exact solutions of the laminar boundary layer equations, were compared with other analytical methods and available expe
17、r- imental results. Their findings show that the heat transfer coef- ficient is higher for elliptical cylinders than circular ones and that the elliptical cylinder with the smaller AR has relatively higher heat transfer coefficient. Rocha et al. (1997) numen- cally studied the heat transfer of circu
18、lar and elliptic cylinders for a constant Prandtl number of 0.70 and Reynolds numbers (based on hydraulic diameter) up to 1600. They observed that the elliptical configuration with an AR of 0.86 and a ratio of semi-minor axis to the length of the tube row of 0.23 is the most efficient one among the
19、ARS of0.75,0.86, and 1 (circular cylinder) studied. The numerical study by Ba hkauskas and Ulinskas 1988). However, Gnielinski (1 979) claimed that the same Nusselt number for a single tube in cross-flow is valid for a tube in an array in cross-flow if the Nusselt number for the single tube is deduc
20、ed with a Reynolds number in which the mean velocity in the gap between the tubes is used as the characteristic veloc- ity. Minimizing Pressure Drop and Vortex-Induced Vibration Using Elliptical Tubes Many authors reported that, in addition to enhancing heat transfer, an elliptical tube can reduce t
21、he pressure drop and, hence, the drag coefficient and vortex-induced vibration compared to a circular cylinder. For example, enhanced heat transfer as well as relative pressure drop reduction of up to 25% were observed when utilizing the elliptical arrangement (Brauer 1964). A heat exchanger built f
22、rom finned elliptical tubes requires less heat transfer surface area and consumes less power for driving fans than an exchanger built from finned circular tubes for a given heat transfer duty (Schulenberg 1966). Ota et al. (1 987) examined the flow around an elliptical cylinder with AR of 0.33 in th
23、e Reynolds number range (based on major axis length) from 35,000 to 125,000. Hasan and Sirn (2004) investigated heat exchangers made with circular (AR = I) and elliptical tubes (AR = 0.32) in the Reynolds number range from 500 to 4,000 (when based on streamwise major axis length and approach air vel
24、ocity). They found the average pressure drop and, hence, the friction factor to be smaller in elliptical configuration, which is 46% of the circular one. It is obvious from these and other investigations that at cross-flow the elliptical tube performs better than the circular configura- tion in term
25、s of reduced drag coefficient, pressure drop, and vortex induced vibration. Almost all of the earlier studies on cross-flow dealt with cylinder surface heated by electrical means either for constant surface temperature, such as Badr (1998), or for uniform heat flux, for instance, Laetitia and Kondjo
26、yan (2002) and Ota and Nishiyama (1 984). Studies on the airside heat transfer coeffi- cients, where tube surface is cooled or heated by water flowing inside the tube and the tube surface temperature and heat flux may not be considered as uniform, are scarce. Even fewer studies may be found that qua
27、ntifi and compare the effects of direction of heat flow on the airside heat transfer process, especially for the cases when the Prandtl number is roughly constant. While authors such as Dittus and Boelter (1930), McAdams (1 942), Winterton (1 998), Cengel (2003), etc., 424 ASHRAE Transactions: Resea
28、rch Figure 1 Closed-loop thermal wind tunnel. suggest a larger Prandtl exponent for heating the fluid than for cooling it, the role of heat flow direction has not been system- atically studied nor postulated based on sound reasoning. On this subject, Scholten andMurray (1995) studied the effects of
29、direction of heat flow on Nusselt numbers for a gas-particle cross-flow over circular tube arrays. Their study mainly deals with the heat transfer mechanism of gas-particle suspension, which does not represent the case of atmospheric airflow. Current Study The cooling and heating of air were experim
30、entally inves- tigated in this study. An elliptical tube array with an AR of 0.30 was oriented in a closed-loop wind tunnel with airflow parallel to the major axis. An AR of 0.30 was employed, as it falls into the optimal AR range of most studies reviewed above (Eckert and Livingood 1953; Harris and
31、 Goldschmidt 2002). The angle of attack was kept at zero order because it is the best orientation indicated by Badr (1998). Airside Reynolds number, based on mean free stream air velocity and the streamwise major axis length of the tube, was varied from 10,000 to 36,000. The effects of Reynolds numb
32、er on Nusselt number and the airflow pressure drop across the array were investigated. The airside Nusselt numbers for the cooling process were compared with those for heating process. EXPERIMENTAL DESIGNS AND TEST PROCEDURES The tests were performed in a closed-loop thermal wind tunnel with a contr
33、action ratio of 6.25 (Figure 1). The 600 mm long, 300 mm x 300 mm test section was made of plexiglass having a thermal conductivity of O. 19 W/m OC. The elliptical tubes, outside major and minor axes of 3 1.7 mm and 9.7 mm, were drawn from 0.825 mm thick, 20.60 mm inner diameter circular copper tube
34、s (type-M, ASTM B-88) with thermal conductivity of 339 W/m OC. The array, as shown in Figures 2 and 3, consisting of 18 elliptical tubes (each of 300 mm .- (Each) Figure 2 Schematic of the test section with elliptical tube array (dimensions and measurement setup). L I inlet measuring oort (M) I 1- I
35、 I ;:I- -! I ,- fl“ U“, Figure 3 Experimental domain with pressure measurement taps. length) equally spaced by 6.1 mm gaps and having outside minor-to-major AR of 0.30, was oriented in the test section at zero angle of attack, that is, with major axis parallel (stream- wise) and minor axis perpendic
36、ular to the direction of airflow. As shown in Figures 2 and 3, the tubes were oriented to form an in-line array of heat exchanger, where water entered at the bottom of the array into the 1st tube and exited at the top from the 18th tube. The tubes were connected at their ends (outside of the test du
37、ct) using small square pockets made of the same plexiglass material. Two half dummy tubes were placed at the top and bottom of the array to reduce any extraneous effects. For the cooling test, hot air at 29.0fl.5“C approached and flowed over the tube array, while cold water at 13.510.6C ASHRAE Trans
38、actions: Research 425 entered the tube to cooi the hot air. During the heating test, coid air at 7.5Il.O“C entered the test section and flowed across the tube array, while the hot water at 24.0f0.6“C entered the tube to heat up the cold air. As marked in Figure 2, two cross sections (Planes 1-1 and
39、2-2), approximately 14 a, upstream and downstream of the array, were chosen to measure the approach air velocity (V,) and temperature (T,J and that of the downstream air temperature (T,J respectively. At the beginning of each experiment, the approach air velocity (, = + Ts,0uter)/2, for the air and
40、at bulk temperature Tw,b = (T, + T,)/2, for the water. The waterside and airside heat transfer rates were estimated as and where positive q, and q, signify air cooling process and nega- tive q, and q, portray air heating process. As mentioned before, the approach air temperature (Ta,), which was mea
41、sured at nine locations around the upstream cross section at Plane 1-1 (Figure 2), was found to be uniform. The average temperature values of these nine locations were taken as the mean air inlet temperature (T,J, which could be represented by a single-point measurement at Position M (Figures 2 and
42、3) at the inlet. The downstream air temperature around cross section 2-2 (Figure 2), however, was found to be less uniform than the upstream section. So, the average value from the nine measuring locations was taken as the mean air outlet temperature (T,J. The mean values of the maximum deviations o
43、f Ta, from its nine grid points were found to be 1.2% for cooling and 5% for heating tests. The temperature difference between the air and water inlets, AT,-, = IT, - T,l, was maintained at 15.7f1.5“C during both the cooling and heating processes for which T, varied between 14C and 23C from one test
44、 to another depending on the total heat transfer rate. The heat transfer rate at waterside was compared with that of the airside. For all the experimental runs, the heat rate values obtained from Equations la and lb generally agreed within +6%. Thus, the q was set to the arithmatic aver- age of airs
45、ide and waterside heat rates (Rugh et al. 1992) in the following form: 426 ASHRAE Transactions: Research The radiant heat transfer between the room and the test section wall was neglected due to negligible temperature difference between them. The temperatures of the inner walls (Twu,inner) of the te
46、st section were assumed to be equal to the temperature of the flowing air inside the duct (Ta,J, that is, Twall,inner N Tu,? Thus, the radiant heat transfer coefficient (hrud) between the test section walls and the tube outer surface was estimated as (3) For commercial copper tubing with an emissivi
47、ty of E N 0.15, a conservative estimate of maximum hrad was found to be 0.95 and 0.8 W/m2“C for the cooling and heatingprocesses, respectively. In both processes, the h, was no more than 1 .O% of the respective airside convective heat transfer coeffi- cient (hJ. Thus, the influence of radiant heat t
48、ransfer between the tube surface and the surrounding environment inside the test section was also ignored. The effect of natural convection on overall heat transfer at airside was neglected. The maximum Grashof number, Gr, attained in the experiment was approximately 15,400 for cool- ing and 18,300
49、for heating tests, respectively. The Gr, to Re: ratios (buoyancy effect indicator) were about 0.000143 (i.e., Ta,): forRe,=4,000 Heating the sir (T,e, To,): for Re, = 7,100 - - - o 800 / - - . 8000 12000 16000 20000 24000 28000 32000 36000 40000 Rea Experiments Air cooling Air heating Figure 7 Change of q with Re, for different water mass flow rates (pical error bars are shown). Re, Cl W) c2 RZ 3,300 9.16 0.415 0.99 6,100 6.82 0.453 0.99 4,000 10.22 0.405 0.99 7,100 5.06 0.483 0.96 g - Re, Relationship To observe the effects of airside Reynolds number
