International Journal of Refrigeration 28 (2005) 851858 www.elsevier.

com/locate/ijrefrig

Heat transfer characteristics of at plate nned-tube heat exchangers with large n pitch
Yonghan Kim, Yongchan Kim*
Department of Mechanical Engineering, Korea University, Anam-dong, Sungbuk-ku, Seoul 136-701, South Korea Received 13 September 2004; received in revised form 26 January 2005; accepted 26 January 2005 Available online 14 April 2005

Abstract The objective of this study is to provide experimental data that can be used in the optimal design of at plate nned-tube heat exchangers with large n pitch. In this study, 22 heat exchangers were tested with a variation of n pitch, number of tube row, and tube alignment. The air-side heat transfer coefcient decreased with a reduction of the n pitch and an increase of the number of tube row. The reduction in the heat transfer coefcient of the four-row heat exchanger coil was approximately 10% as the n pitch decreased from 15.0 to 7.5 mm over the Reynolds number range of 500900 that was calculated based on the tube diameter. For all n pitches, the heat transfer coefcient decreased as the number of tube row increased from 1 to 4. The staggered tube alignment improved heat transfer performance more than 10% compared to the inline tube alignment. A heat transfer correlation was developed from the measured data for at plate nned-tubes with large n pitch. The correlation yielded good predictions of the measured data with mean deviations of 3.8 and 6.2% for the inline and staggered tube alignment, respectively. q 2005 Elsevier Ltd and IIR. All rights reserved.
Keywords: Experiment; Heat exchanger; Finned tube; Geometry; Heat transfer coefcient

` plaques a ` tubes ailete s a ` distance Echangeurs de chaleur a dalignement importante

changeur de chaleur ; Tube ailete rimentation ; E ; Ge ome trie ; Coefcient de transfert de chaleur s : Expe Mots cle

1. Introduction With increasing emphasis on energy savings, extensive efforts are being made to enhance the heat transfer performance of heat exchangers. The heat exchanger

* Corresponding author. Tel.: C82 2 3290 3366; fax: C82 2 921 5439. E-mail address: yongckim@korea.ac.kr (Y. Kim).

performance often is limited by the air-side heat transfer coefcients, because those coefcients are naturally lower than the refrigerant-side values. Therefore, many active and passive methods have been developed to improve the airside heat transfer performance to reduce heat exchanger volume and manufacturing costs. Air-conditioning system utilizes enhanced nned-tube heat exchangers with small n pitch, while refrigerators or freezers use at plate nnedtube heat exchangers with large n pitch to obtain high performance and reliability during the frosting and

0140-7007/$35.00 q 2005 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2005.01.013

852

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

Nomenclature A Ac CP Dc DF Dh D0 F Fc G j h k L LMTD _ m N Pr _ Q ReDc ReDh surface area (m2) minimum free ow area (m2) specic heat (kJ/kg K) collar diameter dimensionless n pitch (F/D0) hydraulic diameter (mm) (DhZ4AcL/A) tube diameter (mm) n pitch (mm) correction factor mass ux (kg/m2 s) Colburn j-factor heat transfer coefcient (W/m2 K) thermal conductivity (W/m K) core depth (m) log mean temperature difference (K) mass ow rate (kg/s) number of tube row Prandtl number capacity (W) Reynolds number based on collar diameter (GDc/m) Reynolds number based on hydraulic diameter (GDh/m) ReDo r T x U Reynolds number based on tube diameter (GD0/m) tube radius (mm) temperature (8C, K) at plate length (cm) overall heat transfer coefcient (W/m2 K)

Greek symbols d boundary layer thickness (cm) m dynamic viscosity coefcient (Ns/m2) h0 surface effectiveness Subscripts air air i inside in inlet loss loss mea measured ref refrigerant o outside out outlet pre predicted

defrosting processes. Since most existing heat transfer data and correlations were developed for small n-pitched heat exchangers, more comprehensive studies on heat exchangers with large n pitches are required to develop a design tool for heat exchangers in refrigerators and freezers. Experimental and numerical studies on the inuence of n pitches and tube alignment on the heat transfer performance have been conducted extensively in nnedtube heat exchangers with small n pitches (F!8.7 mm). Rich [1,2] studied the effects of both n pitches and the number of tube rows for at plate nned-tube geometry. He found that the Colburn j-factor decreased by 50% when the n pitch was reduced from 8.7 to 1.23 mm. In addition, the Colburn j-factors became smaller for each successive row in the low Reynolds number range. Wang et al. [3,4] examined the effects of the number of tube rows, n pitch, and tube diameter on the thermal hydraulic characteristics, and developed a correlation for nned-tube heat exchangers having plain n geometry. They found that the effect of n pitches on the Colburn j-factor was negligible for N R 4 and ReDcR2000. Mon and Gross [5] studied the effects of n pitches on four-row annular nned-tube bundles in staggered and inline arrangements. They found that the boundary layer and horseshoe vortices between the ns were substantially dependent on Reynolds number and the ratio of n pitch to height. Saboya and Sparrow [68] measured the local heat transfer coefcients for one-row,

two-row and three-row plate nned-tube heat exchangers using the naphthalene mass transfer method. Kim and Song [9] found that the total heat and mass transfer rate from the plate increased with F/D, and it reached saturation beyond F/DZ0.5. Due to its performance and reliability under frosting and defrosting conditions, the continuous at plate nned-tube with large n pitch has been widely used in refrigerators. However, studies focusing on heat transfer characteristics on the air-side of at plate nned-tube heat exchangers equipped with large n pitches are very limited. The major objective of this study was to experimentally investigate the effects of n pitches, the number of tube rows, and tube alignment on the heat transfer performance of at plate nned-tube heat exchangers with large n pitches. In addition, the present investigation aimed at developing a heat transfer correlation that can be used in the optimal design of evaporators for refrigerators and freezers.

2. Experimental setup and test conditions A schematic of the experimental setup used to determine the characteristics of the heat transfer is shown in Fig. 1. The equipments were installed in a psychrometric chamber to provide a pre-controlled ambient temperature. The psychrometric chamber was maintained at 2 8C using an airhandling unit including a cooling coil, a heating coil, and a

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

853

Fig. 1. Schematic of experimental setup.

humidier. The air temperature entering the test section was adjusted with an electric resistance heater installed at the entrance of the ow chamber. The humidity was maintained at a set point by adjusting the power input to an ultrasonic humidier. For measurement convenience, the ethylene glycolwater mixture was utilized as a refrigerant inside the coil. A constant temperature bath including a chiller and an electric heater controlled the temperature of the ethylene glycolwater mixture. Air temperatures before and after the test coil were measured by arithmetically averaging the values obtained with nine thermocouples installed in a uniform grid at each location. The thermocouples were calibrated with an accuracy of G0.1 8C. The absolute humidity was measured using a chilled mirror dew point sensor with an uncertainty of G0.2 8C. The airow rate was measured using a nozzle in an open wind tunnel with an estimated uncertainty of G 3.0%. The pressure drop across the test coil was measured using a differential pressure transducer with an uncertainty of G1.0%. The refrigerant ow rate was measured using a Coriolis effect ow meter with an uncertainty of G0.2% of reading. Twenty-two at plate nned-tube heat exchangers were tested by varying the airow rate, n pitch, and tube alignment. As shown in Fig. 2, the longitudinal and transverse tube pitches were 27 and 26 cm, respectively, and the outer tube diameter was 8 mm. The cross-sectional dimension of the test coil was 400 mm!65 mm. Table 1

shows the conditions under which this study was operated. Generally, the air-side heat transfer coefcient is a function of air mass velocity and the shape of the heat exchanger surface. Therefore, the air inlet temperature, the refrigerant inlet temperature, the refrigerant mass ux, and the humidity ratio were maintained constant in the experiments. The test section was made of transparent acrylic resin,

Fig. 2. Fin and coil congurations of two-row heat exchanger.

854 Table 1 Test conditions Parameter Inlet air temperature (8C) Inlet air relative humidity (%) Airow rate (m3/min) Refrigerant inlet temperature (8C) Refrigerant mass ow rate (kg/ h) Fin type Fin pitch (mm) Number of row Tube alignment

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

The air-side heat transfer coefcient, h0, was calculated from Eqs. (1), (3), and (4).
Value 3 60 0.8, 1.1, 1.4, 1.7 33 150 Continuous at plate 7.5, 10.0, 12.5, 15.0 1, 2, 3, 4 Inline, staggered

_ air Z UAFc LMTD Q 1 1 lnr0 =ri 1 C Z C UA hi Ai 2pLk h0 A0 h0

(3) (4)

which was not insulated from the ambient. To measure the heat loss from the test section to the surroundings, an electric resistance heater instead of the test coil was installed in that section. The heat loss was measured by comparing the power input of the electric resistance heater with the airside heat capacity. As shown in Fig. 3, the heat loss was expressed as a function of the temperature difference between the air inside the test section and the surroundings. With this relationship, the air-side heating capacity and the outlet air temperature were modied, and were used for the determination of the air-side heat transfer coefcient. Two independent methods were used to determine the capacity of the test coil: the refrigerant enthalpy method and the air enthalpy method. These capacities were calculated from Eqs. (1) and (2): _ air Z m _ loss _ air Cp;air Tair;in K Tair;out C Q Q _ ref Z m _ ref Cp;ref Tref ;in K Tref ;out Q (1) (2)

The refrigerant-side heat transfer coefcient, hi, was evaluated from the DittusBoelter correlation. The correction factor, Fc, and the surface effectiveness, h0, were determined based on existing correlations [10]. In this study, the estimated average uncertainties of the heat transfer coefcient, the j-factor, and the Reynolds number were analyzed with the method suggested by ASHRAE Guideline 2 [11], and were determined to be 4.0, 6.0, and 3.9%, respectively.

3. Results and discussion 3.1. Effects of n pitches The boundary layer thickness is very important in the determination of the heat and mass transfer coefcient over a at plate, and it was determined by solving the continuity, momentum, and energy equation. Governing equations and boundary conditions can be found in Ref. [10]. Eqs. (5) and (6) are the exact solution of the velocity boundary thickness, d, and the approximated solution regarding the thermal boundary layer thickness, dT, respectively. d 5:0 Z p x Rex dT 0:976 z p 3 d Pr (5)

Based on the experimental results of this study, the equations were consistent with each other within a relative deviation of 5%.

(6)

Fig. 4 shows the laminar velocity boundary layer thickness, thermal boundary layer thickness, and average heat transfer

Fig. 3. Relationship between heat loss and temperature difference.

Fig. 4. Boundary layer thickness and average heat transfer coefcient along at plate.

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

855

coefcient along the at plate. As the boundary layer thickness increases with a rise of the distance from the leading edge, the average heat transfer coefcient decreases. Consequently, slit or louver ns have been widely used to obtain relatively higher heat transfer coefcients by periodically cutting the boundary layer. Since the n height of the one-row coil used in this study was 2.7 cm, the boundary layer thickness should have been approximately 3.5 mm at the tailing edge with the air velocity of 1 m/s. Based on the theoretical solution, the n pitch should be greater than 7 mm to avoid boundary layer interruption between the ns. Fig. 5 shows the effects of n pitches on the j-factor of one-row evaporator coils. Generally, the j-factor increases with a rise of n pitches. However, the dependency of the jfactor on n pitches was reduced with an increase in the Reynolds number (ReDo), which was based on the tube diameter. The boundary layer interruption could not have occurred at large n pitches with faster inlet air velocities (higher Reynolds number) because the boundary layer thickness was smaller than half of the n pitches. Therefore, the heat transfer coefcients for faster inlet air velocities were independent of the pitches. However, for smaller n pitch ranges with lower inlet air velocities, the heat transfer coefcient decreased with a reduction of the n pitches. This is because the interruption of the boundary layers between the ns resulted from an increase in the boundary layer thickness with a reduction of inlet air velocity. Rich [1] reported that the j-factor for a at plate nned-tube heat exchanger decreased with a reduction of the n pitch in the ndez et al. small n pitch range (F!8.7 mm). Romero-Me [12] also showed that the Nusselt number almost proportionally increased with the n pitch in the small n pitch range, while they were almost independent in the large n pitch range. Fig. 6 shows the averaged j-factor for four-row evaporator coil with respect to ReDo. This result is

Fig. 6. Effects of n pitch on the j-factor for four-row evaporator coil.

substantially different from that of the one-row evaporator coil. With the one-row evaporator coil, the j-factor for each n pitches approaches the same value when the Reynolds number increases. However, in the case of the four-row evaporator coil, the j-factor increases with a rise of n pitches at all Reynolds numbers. The enhancement of the jfactor with n pitches may be the result of the delay of the boundary layer interruption to the next row in the inline tube alignment. 3.2. Effects of the number of tube row Fig. 7 shows the effects of the number of tube row on the j-factor of the continuous at plate nned-tube with a n pitch of 7.5 mm. The j-factor decreases with an increase of the number of tube row from 1 to 4. However, the decreasing slope of the j-factor gradually became less important at all Reynolds numbers. This may indicate that the ow elds became uniform at the rear portion of the heat

Fig. 5. Effects of n pitch on the j-factor for one-row evaporator coil.

Fig. 7. Effects of the number of tube row on heat transfer coefcient at a n pitch of 7.5 mm.

856

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

exchanger with the increasing number of tube row. These trends were consistent with Richs data [2], which showed signicant impacts on the heat transfer coefcient due to the number of tube row, when the Reynolds number was low. However, he reported that the effects of the number of tube row became negligible at Reynolds number above 15,000. As shown in Fig. 8, the evolutions of the j-factor in function of the number of tube row at a n pitch of 15 mm were consistent with those at a n pitch of 7.5 mm. For the n pitches of 7.5 and 15.0 mm, the j-factor decreased by approximately 20 and 14%, respectively, with an increase of the number of tube row from 1 to 4. In the case of n pitch of 7.5 mm, the boundary layer interruption between the ns occurred faster compared with that in the case of n pitch of 15.0 mm. Therefore, as the n pitch decreased, the heat transfer coefcient dropped more severely with increasing number of tube row in the multi-row heat exchangers. 3.3. Effects of tube alignment The air-side pressure drop across the evaporator coil under no-frosting conditions was very small. This was due to the combined action of the very large n pitch of continuous at plate n with a small number of tube row and the low air velocity. Therefore, the effects of tube alignment on the air-side pressure drop were not included in this study. Fig. 9 shows the effects of tube alignment on the j-factor with respect to the number of tube row. The staggered tube alignment had higher j-factor than the inline tube alignment in the continuous at plate nned-tube. By making the tube alignment staggered, the j-factor of the four-row evaporator coil increased by 11 and 14% over the inline tube alignment, at the Reynolds numbers of 450 and 900, respectively. Generally, the staggered tube alignment yields higher enhancement of the heat transfer coefcient with an increase in air mass velocity because of the increase intensity in turbulence.

Fig. 9. Effects of tube alignment on heat transfer coefcient with respect to the number of tube row.

As the number of tube row increases from 4 to 8, the jfactor for the staggered tube alignment decreases by only 2.2%, while that for the inline tube alignment regresses by 9.3%. Jang et al. [13] reported that the number of tube row had a small effect on the average heat transfer coefcient when the row number was greater than 4. The heat exchanger used in their study had a staggered tube arrangement and n pitch of 3 mm. Therefore, it can be concluded that the effects of the number of tube row were very limited for staggered tube alignment with NR4 in the continuous at plate nned-tube tested in this study. However, there was reduction in the heat transfer coefcient for the inline tube alignment with NR4, because the smallest n pitch tested in this study was relatively large and the inlet air velocity was small. 3.4. Development of heat transfer correlation Fig. 10 shows the comparison of the present data with the

Fig. 8. Effects of the number of tube row on heat transfer coefcient at a n pitch of 15.0 mm.

Fig. 10. Comparison of the present data with the existing data and correlation.

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858

857

existing data [1] and the correlation obtained from the literature [4]. Wang et al. [4] developed a heat transfer correlation for plain nned-tubes based on data concerning 74 samples. Since the evaporator coil tested in this study had a different n pitch and air velocity compared with Richs tests [1] and the Wang et al. correlation [4], their data and correlation were extrapolated or interpolated to match the test range of this study. For correct data comparisons, the experimental data shown in Fig. 10 were chosen within the valid range of data source in the references. The correlation developed from Richs data [1] largely under predicted the present data, and the average and mean deviations were K23.6 and 23.6%, respectively. The Wang et al. correlation [4] yielded relatively good predictions of the present data for the staggered tube alignment. With that correlation, the average and mean deviations were 3.2 and 12.8%, respectively. However, the Wang et al. correlation [4] largely under-predicted the present data regarding the inline tube alignment with average and mean deviations of K27.3 and 27.3%, respectively. This was due to the lack of their database for the inline tube alignment in their model development. In this study, an empirical correlation for the air-side heat transfer coefcients was developed for at plate nnedtubes with large n pitch based on the present data. The present database included 48 data points for the inline tube alignment and 25 data points for the staggered tube alignment. The test ranges and the geometries for each data set are given in Table 1. The basic correlation for the jfactor was represented as a function of the Reynolds number based on the hydraulic diameter, ReDh. Then, non-dimensional n pitches and the number of tube rows were added into the basic equation to account for their effects. The derived correlation covered both inline and staggered tube alignment since it included the hydraulic diameter into Reynolds number. The resulting correlation of the air-side jfactor for at plate nned-tubes with inline and staggered tube alignment is given in Eq. (7). The coefcients in Eq. (7) were determined by a multiple regression analysis using the above-mentioned database. j Z 0:170NK0:141 DF0:384 ReDh (7)

Fig. 11. Comparison of the measured with the predicted data using the present correlation.

correlation are as follows: number of tube rowZ14, n pitchZ7.515 mm, and Reynolds numberZ6002000.

4. Conclusions The air-side heat transfer characteristics of at plate nned-tube heat exchangers with large n pitches were experimentally investigated by varying n pitches, the number of tube rows, and the tube alignment. For the one-row heat exchanger coil, n pitches had negligible inuence on the heat transfer coefcient when the n pitches were larger. However, as the number of tube row increased, the heat transfer coefcient was enhanced with an increase in n pitches. Basically, the heat transfer coefcient decreased with a rise of the number of tube row. However, for the staggered tube alignment with NR4, the heat transfer coefcient was nearly independent of the number of tube row. The heat transfer coefcient of the staggered tube alignment was 10% higher than that of the inline tube alignment. These trends became more evident with an increase of the air mass velocity and the number of tube row. An empirical heat transfer correlation was developed for at plate nned-tubes with large n pitch based on the present data. The present correlation yielded good predictions of the measured data with mean deviations of 3.8 and 6.2% for the inline and staggered tube alignment, respectively.

Fig. 11 compares the measured and the predicted j-factor using the present correlation. Approximately 90% of the measured data tted well with the predictions within a relative deviation of G10%. The present correlation predicted the data with mean deviations of 3.8 and 6.2% for the inline and staggered tube alignment, respectively. Basically, the generalized form of the correlation can be applied to predict the air-side j-factor for at plate nnedtubes with extensive ranges of operating conditions and geometries. However, the accuracy of the correlation cannot be guaranteed when the correlation is applied beyond the range of the database. The database ranges of the present

Acknowledgements This work was jointly supported by the Samsung Electronics Co. Ltd and the Carbon Dioxide Reduction

858

Y. Kim, Y. Kim / International Journal of Refrigeration 28 (2005) 851858 [6] F.E.M. Saboya, E.M. Sparrow, Local and average transfer coefcient or one-row plate heat exchanger congurations, ASME J Heat Transfer, Ser C 96 (3) (1974) 265272. [7] F.E.M. Saboya, E.M. Sparrow, Transfer characteristics of tworow plate n and tube heat exchanger congurations, Int J Heat Mass Transfer 19 (1976) 4149. [8] F.E.M. Saboya, E.M. Sparrow, Experiments on a three-row n and tube heat exchanger, ASME J Heat Transfer 98 (1976) 26 34. [9] J.Y. Kim, T.H. Song, Effect of tube alignment on the heat/mass transfer from a plate n and two-tube assembly: naphthalene sublimation results, Int J Heat Mass Transfer 46 (2003) 30513059. [10] F.P. Incropera, D.P. Dewitt, Introduction to heat transfer, 3rd ed, Wiley, New York, 1998. [11] ASHRAE Guideline 2, Engineering Analysis of Experimental Data, ASHRAE, Atlanta, USA; 1986. ndez, M. Sen, K.T. Yang, R. McClain, Effect of [12] R. Romero-Me n spacing on convection in a plate n and tube heat exchanger, Int J Heat Mass Transfer 43 (2000) 3951. [13] J.Y. Jang, M.C. Wu, W.J. Chang, Numerical and experimental studies of three-dimensional plate-n and tube heat exchangers, Int J Heat Mass Transfer 39 (1996) 30573066.

and Sequestration Center, one of the 21st Century Frontier R&D Programs in the Ministry of Science and Technology of Korea.

References
[1] D.G. Rich, The effect of n spacing on the heat transfer and friction performance of multi-row smooth plate n-and-tube heat exchangers, ASHRAE Trans 79 (2) (1973) 135145. [2] D.G. Rich, The effect of the number of tubes rows on heat transfer performance of smooth plate n-and-tube heat exchangers, ASHRAE Trans 81 (1) (1975) 307317. [3] C.C. Wang, K.Y. Chi, Heat transfer and friction characteristics of plain n-and-tube heat exchangers. Part I. New experimental data, Int J Heat Mass Transfer 43 (2000) 26812691. [4] C.C. Wang, K.Y. Chi, C.J. Chang, Heat transfer and friction characteristics of plain n-and-tube heat exchangers. Part II. Correlation, Int J Heat Mass Transfer 43 (2000) 26932700. [5] M.S. Mon, U. Gross, Numerical study of n-spacing effects in annular-nned tube heat exchangers, Int J Heat Mass Transfer 47 (2004) 19531964.