Effect of aspect ratio of air jet on heat transfer rate in the impingement cooling of electronic equipment--an experimental study.

Anwarullah, M. ; Rao, V. Vasudeva ; Sharma, K.V. 等

Introduction

Air jets emanating from nozzle are used where high rates of heat transfer are required. A jet impingement device can produce a flow field that can achieve relatively high local heat transfer rates over a surface area which is to be cooled or heated. One major application of jet impingement is in electronic cooling. Other industrial uses of impinging air jets include tempering of glass, annealing of metal and plastic sheets, drying of paper and textiles and cooling of turbine blades. Due to the many industrial applications, extensive research has been undertaken to estimate heat dissipation rates. The heat transfer rate to or from a jet impinging onto a surface is a complex function of Reynolds number (Re), Prandtl number(Pr), nozzle-to-resistor spacing (H/d), and non-dimensional displacement from the stagnation point (r/d).

In recent years, the design of electronic cooling system has gained significant importance due to miniaturization of components requiring dissipation of high heat flux. The problem of estimating the temperature of electronic components with time during startup, shutdown and malfunctioning of the cooling systems has gained the attention of the investigators. A majority of analyses reported the influence of nozzle diameter on the heat transfer rates in the wall jet region. The effect of the nozzle geometry, flow confinement, turbulence, and the variation of jet temperature have been shown to be significant on heat transfer coefficient by Jambunathan et al . [1].

A large amount of work relating to heat transfer from a single-jet was elaborated by Garimella and Rice [2], Fitzgerald and Garimella [3] and San et al. [4]. An experimental analysis has been performed from confined impinging jets by Guerra et al. [5]. San and Shiao [6] expressed the stagnation Nusselt number as a function of jet Reynolds number, ratio of jet height-to-diameter, jet plate length-to-diameter and jet plate width-to-diameter. Gardon and Akfirat [7] estimated from experiments the local and average heat transfer

[Nu.sub.0] = 1.2 [Re.sup.0.58] [(H/d).sup.-0.62] (1)

valid in the range 2000 < Re < 50,000 and 14 < H/d < 60, with an average deviation of 5 %. Yang et al. [8] presented heat transfer results of jet impingement cooling on a semi-circular concave surface bringing out the significant effects of the nozzle geometry and the curvature of the plate. A wide variety of nozzles were tested by Garimella and Nenaydykh [9], and McMurray et al. [10] have performed heat transfer measurements for both laminar and turbulent boundary layer and their correlation for laminar boundary layer is of the form

[Nu.sub.0] = 0.73 [Re.sup.1/2] [Pr.sup.1/3] (2)

For a laminar flow, Kendoush [11] studied theoretically the heat and mass transfer mechanics of an impinging slot jet by means of the boundary layer theory. The results were restricted just to the stagnation zone. The recommended correlation for the stagnation heat transfer rate is given below

[Nu.sub.0] = 0.75 [[Re Pr (1.02 - 0.024 (H/d)].sup.1/2] (3)

Zhou and Lee [12] experimentally investigated the fluid flow and heat transfer characteristics of a rectangular air jet impinging on a heated flat plate. The effect of jet Reynolds and nozzle-to-plate spacing on local and average Nusselt number were studied in the range of 2,715 < Re < 24,723. The following correlation was recommended for the estimation of stagnation heat transfer coefficient

[Nu.sub.0] = C [Re.sup.m] [Pr.sup.0.4] (4)

valid in the range 0.244 < C< 0.156 and 0.620 < m< 0.587 for 4 <H/d < 30. Colucci and Viskanta [13] determined the effect of nozzle geometry on local convective heat transfer rate for jet impingement cooling on a flat plate. Gauntner et al. [14] presented a review report of the flow characteristics of a turbulent jet impinging on a flat plate. Zumbrunnen et al. [15] carried out studies on convective heat transfer from a plate cooled by water jets. Lytle and Webb [16] investigated the flow structure and heat transfer characteristics of air jet impingement for nozzle-plate spacing of less than a nozzle diameter in the range of 3600 < Re < 27,600. Baydar [17] carried out an experimental investigation for low Reynolds number up to 10,000 at various nozzle-to-plate ratios. An expression for the stagnation Nusselt number was derived by Vader et al. [18] given by

[Nu.sub.0] = 0.505 [Re.sup.0.5] [Pr.sup.0.376] (5)

Zhou and Ma [19] experimentally investigated the radial heat transfer behavior of impinging submerged circular jets. An expression for the local Nusselt number valid at the stagnation point in the radial direction is obtained as

[Nu.sub.0] = 1.32 [Re.sup.0.499] [Pr.sup.0.33] (6)

Lienhard et al. [20] experimentally investigated the splattering and heat transfer during impingement of a turbulent liquid jet. The recommended equation for the local Nusselt number at the stagnation location is given by

[Nu.sub.0] = 1.24 [Re.sup.0.5] [Pr.sup.0.33] (7)

Siba et al. [21] experimentally studied impingement cooling of a flat circular disk made of conducting material SS304. Recently, the flow characteristics of both confined and unconfined air jet impinging normally onto a flat plate have been experimentally investigated by Baydar and Ozmen [22]. Schwarz and Cosart [23] presented measurements and and theoretical analysis on fluid flow characteristics of impinging slot jet, but only for the turbulent wall jet zone. The present study is concerned with the experimental investigation of the confined impinging jet flow fields at various nozzle-to-resistor surface distances. The main objective of the present work is to study the effect of geometric parameters on the heat transfer characteristics of resistor surface normal to impinging air jet

Experimental setup

The experimental set up as shown in Fig. 1, consists of five cylindrical electrical resistors fixed to an insulating plate of diameter 100mm and 2mm thick located centrally on an aluminum heater plate. A chip assembly on PCB is simulated with the electrical resistors which are 25 mm long and 4 mm in diameter. The resistors each of 5 W rating are connected to supply through volt and ammeter. Five J-type thermocouples are attached to measure the surface temperature of each resistor. Thermocouples of Type J would normally have an error of approximately 0.75% of the target temperatures when used at a temperature lower or higher then 277[degrees]C. A heater plate of 240 mm diameter and 20 mm thick is connected to a heating coil of 500 W rating through a dimmerstat to enable the temperature of the insulating plate to be higher than ambient. Two thermocouples are connected to the heater plate and another one measures the ambient temperature. All these eight thermocouples are connected to a temperature indicator through a scanner to observe the readings and store the values in a personal computer. The airflow rate through a nozzle of different diameters located above the resistors is measured with a rotameter. Air at 20bar is made available to the nozzle from a reciprocating air compressor of 220 liter stororage capacity through the rotameter. Provision is made to vary the distance between the nozzle tip and the test surface. The axis of the nozzle is always aligned with the centre resistor and normal to the plane on which heat sources are mounted.

[FIGURE 1 OMITTED]

Experimental Procedure

The air jet emanating from the nozzle and impinging on the resistors is depicted as free jet and wall jet regions respectively and shown in Fig 2. Power is supplied to the resistors through a step down transformer and the aluminum plate through a dimmerstat.

[FIGURE 2 OMITTED]

The volumetric energy generation due to heating of the resistors using AC current is assumed to be uniform. The temperature of the resistors is allowed to rise up to 95[degrees] C and then cooled by forced convection mainly from the top surface by the air stream flowing in the wall jet region. The surface temperature of the resistors are recorded till they attain 40[degrees]C The procedure is repeated at different flow rates of air with temperature values recorded in the Reynolds number range of 5850 to 12200. The velocity of jet is measured using a Pitot tube. The heat loss from the bottom of the resistors is assumed to be negligibly small.

Results and discussion

Air jet from the nozzle is forced over the resistors when they have attained a maximum steady temperature of 98[degrees]C in the range of 5850 < [Re.sub.d] < 12200 and 2 < H/d < 10. It is observed that the surface temperature of the resistors drop down rapidly in 50 seconds from the time of starting of air flow. As expected the temperature gradient is higher at larger values of Reynolds and lower values of H/d as can be observed from Fig. 3. The rapid decrease in temperature is also due to large temperature potential between the surface and the ambient.

[FIGURE 3 OMITTED]

Fig.4 illustrates the effect of Reynolds number on heat removal rates for different H/d ratios. The heat removal rate increases with increase in Re and decreases with increase in H/d ratio. As H/d increases, the distance between the nozzle and the heated resistors increase, resulting in lower heat dissipation.

The rate of heat removal from the resistors with H/d ratios for 5850 < [Re.sub.d] < 12200 is shown in Fig. 5. As expected, the heat removal rate increases gradually with increase in Reynolds number due to higher mass flow rate. It is also observed that the heat flux decreases with increasing values of H/d ratio at a particular value of Re.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

The computed values is shown plotted in Fig. 6 between surface heat flux and temperature difference between the resistor surface and the ambient for 2 <H/d < 10 and 5850 < [Re.sub.d] < 12200. It can be observed that at a value of Reynolds number, the temperature difference is higher for lower values of H/d. This may be due to lesser residence time for air to extract heat from the surface.

Fig. 7 shows the distributions of local Nusselt number for H/d ranging from 2 to 10. From the stagnation point to the exit, local Nusselt number increases, then decreases quickly. Near the stagnation point the flow velocity increases and results in increases of the local Nusselt number

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

The effect of nozzle diameter on heat transfer coefficient for different values of Reynolds number is shown in Fig. 8. It can be observed that at a particular value of Reynolds number, the heat transfer coefficient decreases with increase of nozzle Diameter

The results presented in figure 9 illustrate the effect of varying nozzle diameter on the heat transfer distribution for a Reynolds number of 5850 and H/d = 2. The heat transfer coefficient increases with decreasing in diameter of the nozzle is clear. This is due to the high air flow velocities involved for small diameter nozzles. An increase of over 60% is observed for a decrease in nozzle diameter from 10mm to 5mm. It is worth noting that figure 9 displays the heat transfer coefficient distribution for the dimensionless radial distance, r/d. The velocity at the stagnation point is maximum and hence heat transfer coefficient is maximum

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

Data reduction and uncertainty analysis

Present experimental data equation:

[Nu.sub.Exp] = 0.93 [Re.sup.0.5] [Pr.sup.0.4] (8)

The local heat transfer coefficient is calculated using the following equation:

H = q/([T.sub.s] - [T.sub.a] (9)

where q(W/[m.sup.2]) is local heat flux from the top surface of the resistor to the air.

Here, [T.sub.wall] (K) is the surface temperature: [T.sub.a] (K) is the ambient temperature. The local Nusselt number on the resistor surface is defined by Eq. (10).

Nu = h.d/[k.sub.air] (10)

where q(W/[m.sup.2] K) is the local heat transfer coefficient ; d (m) is nozzle diameter; [k.sub.air] is the thermal conductivity of the air.

Nozzle Reynolds number is defined as follows:

Re = Vd/v (11)

The uncertainty associated with the experimental data is estimated using the standard single-sample uncertainty analysis recommended by Kline and McClintock [24] and Moffat [25]. In the present experiments, the temperature measurements were accurate to within [+ or -] 0.5[degrees]C, the uncertainty of the convective heat flux q is estimated to be 2.65% and those of d [Re.sub.d] and [Nu.sub.0] for the ranges of parameters studied under steady-state conditions is within [+ or -] 2% and [+ or -] 5%, respectively.

Experimental validation

Experimental data is used to evaluate the local heat transfer behaviors of impinging confined circular jet at a fixed radial location for jet Reynolds number in the range of 5850 < [Re.sub.d] < 12200. The results for H/d = 2 in the form of Nu/[Pr.sup.1/3] ~ Re are depicted in Fig. 10.Stagnation point heat transfer is experimentally determined using Eq. (8) and compared with each other as a validation exercise in this study. Fig. 10 presented the effect of jet Reynolds number on stagnation point (i.e., r/d = 0) Nusselt number. Stagnation point Nusselt number increases remarkably with jet Reynolds number. For comparison, Fig. 10 also presents the experimental data of Zhou and Ma [18] for a submerged jet and Lienhard et al. [19] for water free jet. As illustrated by triangles in the figure, stagnation point heat transfer in this study was enhanced slightly. Independent of jet Reynolds number and jet type, the present data agree well with the previous experimental results of Refs. [18, 19].

For jet Reynolds number in the range 5850 [less than or equal to] [Re.sub.d] [less than or equal to] 12200, the stagnation Nusselt number were examined at H/d = 4 and compared with the previous experimental results as a validation process. Fig. 11 exhibits the variation of the stagnation Nusselt number obtained at H/d = 4 with jet Reynolds number, in which comparison of several empirical correlations of stagnation heat transfer from the work of Vander et al. [17], McMurray et al. [10], Zumbrunnen et al. [15], and Kendoush [11], are also plotted. In this figure, the calculation was carried out at H/d = 4 and the Prandtl number of air (Pr = 0.71) was adopted for all the correlations. Good agreements between the present data and the previous experimental and theoretical results were observed. The data is in close agreement with the data of McMurray et al. [10] and Kendoush [11].

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

Variation of the stagnation heat transfer rate with nozzle-to-resistor spacing is examined in greater detail in Fig. 12. It is seen from the figure that the variation of NuO with H/d, which may not be monotonic, exhibits a complex nature depending on jet Reynolds number. At a lower jet Reynolds number of Re = 3100, the stagnation heat transfer rate decreases initially with the axial distance from the nozzle exit till H/d = 4 and then increases. Thereafter, it decreases monotonously beyond the point at H/d = 8.

[FIGURE 12 OMITTED]

The present experimental data obtained with nozzle diameter of 8mm lies in between that of Gardon and Afirat [7] and Zhou and Lee [12] who have conducted experiments with 3.18 mm and 11 mm respectively.

The present experimental data is subjected to regression and given in a simplified form as

[Nu.sub.Reg] = 1.065 [Re.sup.0.58] [Pr.sup.0.54] [(H/d).sup.-0.0174] (12)

valid in range 2< H/d< 10, and 5850< Re< 12200 with average deviation 9% and standard deviation 10%.

The present experimental data is in good agreement with the values of Nusselt obtained with Eq.12 as shown in Fig.13

[FIGURE 13 OMITTED]

Conclusions

Based on the present experimental conditions, the jet Re, the nozzle tip- toresistor spacing and cooling time have an important influence on the heat transfer of impinging circular jet nozzle, especially on the wall jet and impingement region. For confined air jet nozzles, local heat transfer rate at given radial location were correlated and compared. The effects of the jet Re and nozzle tip-to-resistor spacing on the Nu of impinging jet were determined to develop an optimum parameter of heat transfer enhancement. The heat transfer rate increases as the jet spacing decreases owing to the reduction in the impingement surface area. It is observed that the heat transfer coefficients increase with H/d up to 8 for any Reynolds number and increases with increase in Reynolds. The present study will provide a better understanding on the fluid flow and heat transfer characteristics of impinging air jet

Acknowledgement

The first author is working as faculty in the Department of Mechanical Engineering and grateful to the management of Muffakham Jah College of Engineering and Technology, Hyderabad for the financial support in the fabrication of the experimental setup

Nomenclature

A                    surface area of the resistor, [m.sup.2]

[C.sub.p]            specific heat at constant pressure, J/ (kg K)

d                    diameter of nozzle, m

H                    distance between nozzle tip to resistor, m

Nu                   local Nusselt number, Eq. (10)

Re                   jet Reynolds number, Vd/v

q                    heat flux, W/[m.sup.2]

t                    cooling time, seconds

[T.sub.s]            surface temperature of the resistor before
                     cooling, [degrees]C

[T.sub.[infinity]]   ambient temperature, [degrees]C

V                    velocity of air, m/sec

H/d                  nozzle-to-resistor spacing to nozzle diameter

[Nu.sub.O]           Nusselt number at stagnation point

K                    thermal conductivity of air, W/(m K)

Pr                   Prandtl number

r                    radial distance measured from the stagnation
                     point, m

Greek symbols

[rho]                density of air, kg/[m.sup.3]

v                    kinematic viscosity of air, [m.sup.2]/s

Subscripts

Reg   regression

Exp   experimental

References

[1] K. Jambunathan, E. Lai, M.A. Moss, B.L. Button., 1992, "A review of heat transfer data for single circular jet impingement," Int. J. Heat Fluid Flow 13: 106-115

[2] S.V. Garimella, R.A. Rice., 1995, "Confined and submerged liquid jet impingement heat transfer," J. Heat Transfer 117, 871-877.

[3] J.A. Fitzgerald, S.V. Garimella., 1998, "study of the flow field of a confined and submerged impinging jet,"Int. J. Heat Mass Transfer 41 (8-9),1025-1034.

[4] J.Y. San, C.H. Huang, M.H. Shu. 1997, "Impingement cooling of a confined circular air jet," Int. J. Heat Mass Transfer 40 (6), 1355-1364.

[5] D.R.S. Guerra, J. Su, P. Atila, A.P. Silva Freire.,2005, "The near wall behavior of an impinging jet,"nt. J. Heat Mass Transfer 48 (14), 2829-2840.

[6] J.Y. San, W.Z. Shiao., 2006, "Effects of jet plate size and plate spacing on the stagnation Nusselt number for a confined circular air jet impinging on a flat plate,"Int. J. Heat Mass Transfer 49 (19-20) ,477-3486.

[7] R. Gardon, J.C. Akfirat.,1965, "The role of turbulence in determining the heat transfer characteristics of impinging jets," Int. J. Heat Mass Transfer 8, 1261-1272.

[8] G. Yang, M. Chio, J.S. Lee, 1999. "An experimental study of slot jet impingement cooling on concave surface: effect of nozzle configuration and curvature, Int. J. Heat Mass Transfer 42, 2199-2209.

[9] B. Garimella, B. Nenaydykh.,1996, "Nozzle-geometry effects in liquid jet impingement heat transfer," Int. J. Heat Mass Transfer 39, 2915-2923.

[10] D.C. McMurray, P.S. Myers, O.A. Uyehara, 1966, "Influence of impinging jet variables on local flat surface with constant heat flux, in: J.P. Hartnett (Ed.), Proceedings of the Third International Heat Transfer Conference, 1, Hemisphere Publishing Corporation," pp. 292-299.

[11] A.A. Kendoush, 1998. "Theory of stagnation region heat and mass transfer to fluid jets impinging normally on solid surface," Chem. Eng. Process.37,223-228.

[12] D.W.Zhou and Sang-Joon Lee., 2007, Forced convective heat transfer with impinging rectangular jets," Int. J. Heat and Mass Transfer 50, 1916-1926.

[13] D.W. Colucci, R. Viskanta., 1996, "Effect of nozzle geometry on local convective heat transfer to a confined impinging air jet," Exp. Thermal Fluid Sci. 13, 71-80.

[14] J.W. Gauntner, J.N.B. Livingwood, P. Hrycak, 1970. "Survey of literature on flow characteristics of a single turbulent jet impinging on a flat plate," NASA TN D-5652,

[15] D.A. Zumbrunnen, F.P. Incropera, R. Viskanta, 1989. "Convective heat transfer distributions on a plate cooled by planar water jets," J. Heat Transfer 111 (3), 889- 896.

[16] D. Lytle, B.W. Webb., 1994, "Air jet impingement heat transfer at low nozzle-plate spacings," Int. J. Heat Mass Transfer 37, 1687-1697.

[17] E. Baydar, 1999,Confined impinging air jet at low Reynolds numbers, "Exp. Thermal Fluid Sci. 19, 27-33

[18] D.T. Vader, F.P. Incropera, R. Viskanta., 1991, "Local convective heat transfer from a heated surface to an impinging, planar jet of water," Int. J. Heat Mass Transfer 34 b (3), 611-623.

[19] D.W. Zhou, C.F. Ma., 2006, "Radial heat transfer behavior of impinging submerged circular jets," Int. J. Heat Mass Transfer 49 (9-10),1719-1722.

[20] J.H. Lienhard, X. Liu, L.A. Gabour., 1992, "Splattering and heat transfer during impingement of a turbulent liquid jet," J. Heat Transf. 114 (2), 362-372.

[21] Erick A. Siba et. Al.,2003, "Heat Transfer in a High Turbulence Air Jet Impinging Over a Flat Circular Disk," Journal of Heat Transfer, vol.125:pp257- 265

[22] Baydar and Y. Ozmen., 2006,An Experimental Investigation on Flow structures of Confined and Unconfined Impinging air jets," Heat Mass Transfer,42:pp. 338-346

[23] W.H. Schwarz, W.P. Cosart., 1961, "The two-dimensional turbulent wall jet," J. Fluid Mech 10,481-495.

[24] S. J. Kline and F. A. McClintock., 1953, "Describing uncertainties in single-sample experiments, Mech. Engng 3-8.

[25] R. J. Moffat., 1988, "Describing the uncertainties in experimental results, Expl.Therm.Fluid Scii. 1, 3-17

M. Anwarullah (1), V. Vasudeva Rao (2) and K.V. Sharma

(1) Research Scholar, (3) Professor

(1,3) Centre for Energy Studies, JNTU College of Engineering, Hyderabad-500034, India

(1) E-mail address: manwar_sana@yahoo.com.

(2) Professor, Department of Mechanical Engineering, SNIST, Hyderabad. India