# CFD analysis of heat transfer and fluid flow characteristics of swirl and non swirl impinging jet for electronic cooling application.

IntroductionDuring the last decades, the rapid development of the semiconductor industry made the integration degree of the IC (Integrated Circuit) increase continually. Simultaneously, the power density and heat flux on the electronic chips has also become higher and higher. As most of the electronic chips operate only in some range of temperature, it is very important to remove the heat from apparatus. It's obviously that the traditional natural and forced convection methods can't satisfy the cooling of such high heat flux density any more. In this situation, the impinging jet draws more and more attention as a very potential alternative because of its much better heat transfer effect than the previous methods. Liquid jet impingement has higher heat transfer coefficient than gas, but the complicated configuration and the high cost restrict its development. In all the cooling schemes of gas jets, air jet impingement has its unique preponderance because of its abundant source, low cost, high reliability and no pollution to environment. According to the former research [1], it can be known that the straight impinging jet strengthened the heat transfer on the simulant chip remarkably, and made the heat transfer coefficient present a bell-shaped distribution. That means, at the stagnation point, the heat transfer effect is the strongest; while in the region apart from the stagnation point, the heat transfer becomes weaker and weaker. This non uniformity of heat transfer on the chips can lead to the failure of the material. Therefore, the straight impinging jet array and the swirling impinging jet are presented to solve this problem Yang [2] presented the heat transfer results of jet impingement cooling on a semi-circular concave surface and clearly explained the significant effects of the nozzle geometry and the curvature of the plate Both the jet array and the swirling jet can strengthen the heat transfer as well as uniform the temperature distribution on the impinging surface. LI Yong kang [3] made a 3-D numerical computation about the jet array impingement with initial cross flow, they found that the impinging cooling with inline arrangement was better than that with staggered arrangement for a given jet-to-cross flow mass flux ratio. M. Emin Arzutug [4] compared the mass transfer between a plate and submerged conventional and multichannel impinging jets, it was found that the values of the mean mass transfer coefficient over the surface for conventional impinging jet and multichannel impinging jet were relatively close to each other, and the multichannel impinging jet could make a more uniform local mass transfer coefficient distribution. P. Brevet [5] researched the optimization scheme of heat transfer about a row of impinging jets, the influence of the spacing between every two adjacent holes both in the transverse and longitudinal direction is analysed Nozaki (6) combined whole field velocity and temperature measuring techniques (PIV and LIF) to investigate swirl impinging jets. The work was conducted for a small value of the jet Reynolds number, namely 4000. The results confirmed the presence of the recirculating zone at the stagnation region, and relate the heat transfer behavior to the dynamic character of this zone. Aldabbagh and I. Sezai [7] investigated the flow and heat transfer characteristics of impinging laminar multiple square jets numerically, they found that the magnitude of the local maximum Nusselt number at the stagnation point was not affected by jet-to-jet spacing. L. Huang [8] studied the flow channel and the swirl angle of swirling multichannel impinging jets. F. di Mare [9] simulated the large eddy in a model gas turbine combustor using a strong swirling flow. Mao-Yu Wen [10] carried out flow visualization experiments to study the heat transfer of a conventional impinging jet and two kinds of swirling jets. Chen Yu-Yang [11] used thermal liquid crystal technique to study the heat transfer rate of jet impingement with and without swirling, obtained the Nusselt number distribution under some different convective conditions.

In this paper, a variety of factors that affect the flow and heat transfer of straight and swirling jet are studied. In the numerical analysis, circular nozzle of 15 mm diameter with helical inserts are used; the influence of Reynolds number, the nozzleto-plate spacement and injection angle are investigated. These influencing factors are simulated by the commercial software, CFX for a comparative study. It reports the results of heat transfer and flow visualization which are performed to investigate and compare the heat transfer performance for the swirling jet on the surface with those of a straight impinging jet having the same diameter at the same conditions. The performance of these jets is evaluated in terms of the measured increases in the values of the local Nusselt numbers and the uniformity of their radial distributions on the impinged surface. In the heat transfer analysis the effects of (a) Pitch distance of helical insert, (b) jet-to-surface distances, H/D for 5, 7.5 and 10 (c) air Reynolds number, Re for 8000, 15000, 23000, 30000 are investigated.

The physical properties of Epoxy plate with copper foil (common material for printed circuit board) have been considered for impinging surface.

Computational Modeling

The computational domain of impinging jet system is shown in Fig.1

[FIGURE 1 OMITTED]

The diameter of the jet nozzle is D. The ratio of the height of the channel to the diameter of the jet, H/D, is equal to 5. To study the effect of geometry on flow fields, larger H/D up to 10 is also considered. Selection of these parameters is based on the analysis that under these conditions the confined wall may strongly affect the jet flow. The dimensions for the extended domain in Fig.1 are: L/D=6.66 L'/D= 26.66

The following assumptions and boundary conditions have been specified: Boundary 1 is the jet inlet (x =0 and y = D). Jet flow is at a uniform velocity (u = 0) and constant temperature (T = [T.sub.0]). Boundary 5 is the target wall (x= L+H) and constant heat flux (q= [q.sub.x]) applied. The wall is assigned at constant temperature (T = [T.sub.w]). Boundaries 2 and 3 are the confined walls, where no-slip and adiabatic conditions are applied.

At the channel exit (Boundary 4, y =D/2+L'/2) in Fig. 1 a constant pressure is applied: p = constant.

[FIGURE 1.1 OMITTED]

The domain was discretized with an unstructured tetrahedral mesh with refined mesh at the impinging surface and at nozzle exit for taking the effect of exit flow structure. Optimization of meshing element is essential to reach a balance between the amount of computing time and the accuracy in the solution of flow variables. Initially the domain is discretized with 357537, 482067, 652451, 806049 and 968994 elements. It was learnt that velocity contours of 357537, 482067 and 652451 are different those of 806049 and 968994 elements. The latter two are good agreement with each other. Considering the lesser computing time the domain with 806049 elements was selected over the one with 968994 elements

Governing Equations

The governing equations of mass, momentum and energy for steady state compressible flow are defined using Navier Stokes equations.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

These equations are integrated over a control volume, and Gauss' Divergence Theorem is applied to convert volume integrals to surface integrals. The commercial software package CFX from ANSYS, is used in this study. The simulation uses the coupled solver, which solves the governing equations (for u, v, w, p) as a single system. CFX code uses a co-located (non-staggered) grid layout such that the control volumes are identical for all transport equations.

Turbulence Model

The Standard [kappa]-[epsilon] model is the most extensively used and validated turbulent model among all the [kappa]-[epsilon] models. It has achieved notable successes in calculating a wide variety of thin shear layer flows. Nevertheless, it is reported not to perform well in some important cases like flows with large extra strains (e.g. curved boundary layers, swirling flows) and rotating flows. The RNG [kappa]-[epsilon] model, on the other hand, is similar in form to the Standard [kappa]-[epsilon] model, except for some added refinements to improve the accuracy for rapidly strained flows. As for the Realizable k-s model, initial studies have shown that

It is the best [kappa]-[epsilon] model in solving fluid dynamic problems. However, the Realizable k-s model has yet to prove in exactly which instances it consistently outperforms the RNG [kappa]-[epsilon] model. It is more responsive to rapid strain and streamlines curvature relative to the other two [kappa]-[epsilon] models. In addition, it takes into account the effect of swirl on turbulence, thus enhancing the accuracy for swirling flows. Therefore, for a rapidly strained air flow with a confined injection in domain, the RNG [kappa]-[epsilon] model was preferred. The transport equations of RNG [kappa]-[epsilon] used in this work are as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

where [rho] is the fluid density, k is the kinetic energy, s is the dissipation rate, u is the mean velocity, [[mu].sub.eff] is the effective viscosity, [[alpha].sub.k] is the inverse effective Prandtl number for the k term, [[alpha].sub.[epsilon]] is the inverse effective Prandtl number for the s term, [G.sub.k] is the generation of turbulent kinetic energy due to mean velocity gradients, [G.sub.b] is the generation of turbulent kinetic energy due to buoyancy, [Y.sub.m] is the contributions of the fluctuating dilation in compressible turbulence to the overall dissipation rate, [S.sub.k] and [S.sub.[epsilon]] are the user defined source terms for k and [epsilon], respectively. [R.sub.[epsilon]] is an additional term in the RNG [kappa]-[epsilon] model which is modeled by

[R.sub.[epsilon]] = C[mu][rho][[eta].sup.3](1-[eta]/[[eta].sub.0])[[epsilon].sup.2]/1 + [beta][[eta].sup.3] [[epsilon].sup.2]/k (3)

Where [eta] = [S.sub.k]/[epsilon], [[eta].sub.0] = 4.38 and [beta] = 0.012.

By substituting (3) in (2), the equation (2) can be written as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where [C.sub.le] = 1.42 and [C.sub.2e] = 1.68

Results and Discussion

Non Swirl

Fig 1 shows radial distribution of Nusselt number for H/D=5 for vertical and 20 position of nozzle. It is observed that for lower value of Reynolds number (Re = 8000) the radial uniformity of Nusselt number is more in the case vertical position.

[FIGURE 2 OMITTED]

It is observed from the result that for vertical position immediately after stagnation region Nusselt number increases in the wall jet region for a while after that it decreases gradually. When the Reynolds number increases, the radial uniformity decreases.

[FIGURE 3 OMITTED]

It is observed form the fig. 3 that when H/D distance increases (H/D=10) the value of Nusselt number also decreases. But the comparatively radial uniformity decreases. The radial distribution of Nusselt number is quiet uniform in the case of vertical Position.

[FIGURE 4 OMITTED]

Inclination

Fig 4 shows that for impinging angle 20 deg the stagnation point heat transfer is high in the case of lower value of H/D distance. At the higher value of H/D distance the more uniform heat transfer is obtained both in lower and higher value of Reynolds number.

[FIGURE 5 OMITTED]

It is observed from fig 5 that when H/D distance increases the value of Nusselt number decreases both in the lower and higher value of Reynolds number range but the radial uniformity is more in the case of high H/D distance.

Swirl and Non swirl

[FIGURE 6 OMITTED]

Fig 6 shows the Nusselt number distribution for swirl and non swirl flow. The results show that for H/D = 5 the non swirl flow exhibits increased Nusselt at the stagnation point and decreases along the wall jet region. It is clear from result that though swirl flow exhibits low value of Nusselt number the distribution of Nusselt number over the surface is more uniform.

[FIGURE 7 OMITTED]

When the Nusselt value increases with the increase of Reynolds number the radial uniformity of Nusselt is maintained at higher Reynolds number (Re = 23000). When the H/D distance increases (H/D = 10) at the lower value of Reynolds number, radial uniformity was maintained in the case if swirl flow. It is also observed.

[FIGURE 8 OMITTED]

Effect of Flow parameters

Details of exit flow structures in the case of an annular impinging jet with swirl and non swirl have been discussed in this section

[FIGURE 9 OMITTED]

Fig.9 shows velocity vector distribution for non swirl jet with varying H/D distances. It is observed from the result that for H/D = 5 recirculation occurs symmetry to nozzle axis. This is due to reverse flow after impingement. This will obviously enhance the heat transfer rate. In the case of H/D = 7.5 no such recirculation zones in the reverse flow. When the H/D is further increased the stream of jet is carried away by fluid particles.

[FIGURE 10 OMITTED]

Fig.10 shows exit flow pattern with separation distance of H/D= 5. The result shows the exit flow acquires radial velocity components due to the presence of swirl diverging radially as soon as it exits the annular nozzle. The flow field of swirl jet shows that increased uniformity in the radial velocity component with respect to nozzle axis. Strong recirculation zones have been observed in swirl jet. In the case of non swirl no such radial component of velocity has been observed. Recirculation in flow has been observed for this case about the flow axis due to reverse flow.

[FIGURE 11 OMITTED]

Fig.11 shows the turbulence kinetic energy variation. The turbulence kinetic energy is characterized by measured root mean square (RMS) velocity fluctuations. It reaches higher value in the recirculation zones. This will lead to increased heat transfer rate in that region. Uniform distribution of turbulence kinetic energy has been observed in the swirl jet. This will significantly increase the uniformity in heat transfer. In the case of non swirl jet concentrated turbulence level is observed in the stagnation point region and decreases over the wall jet region. This will comparatively reduce the heat transfer rate in the wall jet region.

Conclusion

Thus the analysis of heat transfer and flow characteristics of swirl and non swirl flow jet impingement has been carried out and following conclusions have been drawn.

In the case of non swirl flow, for vertical position immediately after stagnation region Nusselt number increases in the wall jet region for a while after that it decreases gradually.

While increasing the separation distance the value of Nusselt number decreases and the radial uniformity increases. Strong recirculation zones have been observed in the lower separation distances.

In the swirl and non flow case, increased radial distribution of Nusselt number has been observed in the swirl flow but comparatively Nusselt number decreases slightly.

When the Reynolds number increases the radial uniformity of Nusselt number decreases in both cases.

Increased Radial velocity components have been observed in the swirl flow and it leads to comparatively stronger recirculation zones.

Intensity of turbulence is uniform both in the stagnation and wall jet region for swirl flow jet

References

[1] Mao-Yu Wen, Kuen-Jang Jang, An Impingement Cooling on A Flat Surface by Using Circular Jet with Longitudinal Swirling Strips, International Journal of Heat and Mass Transfer, (46) 4657-4667, 2003

[2] G. Yang, M. Chio, J.S. Lee, 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, 1999

[3] LI Yong-kang, ZHANG Jing-zhou, TAN Xiaoming, "3-D Numerical Computation of Jet Array Impingement with Initial Cross flow," Transaction of Nanjing University of Aeronautics & Astronautics, (21) 128-133, 2004

[4] M. Emin Arzutug, Sinan Yapici, M. Muhtar Kocakerim, "A Comparison of Mass Transfer between a Plate and Submerged Conventional and Multichannel Impinging Jet," International Communications in Heat and Mass Transfer, (32) 842-854, 2005

[5] P. Brevet, C. Dejeu, E. Dorignac, M. Jolly, J.J. Vullierme, eat Transfer to a Row of Impinging Jets in Consideration of Optimization, International Journal of Heat and Mass Transfer, (45) 4191- 4200, 2002

[6] Nozaki, A., Igarashi, Y. Hishida, K., Heat Transfer Mechanism of a Swirling Impinging Jet in a Stagnation Region, Heat Transfer--Asian Research, 32, No. 8. 2003

[7] L.B.Y. Aldabbagh, I. Sezai, Numerical Simulation of Three-dimensional Laminar Multiple Impinging Square Jets," International Journal of Heat and Fluid Flow, (23)509-518, 2002

[8] L. Huang, M.S. EL-GENK, "Heat Transfer and Flow Visualization experiments of swirling, Multi channel, and Conventional Impinging Jets," Int. J. Heat Mass Transfer, (41) 583-600, 1998

[9] F. di Mare, W.P. Jones, K.R. Menzies, Large Eddy Simulation of A Model Gas Turbine Combustor, Combustion and Flame, (137) 278-294, 2004

[10] Mao-Yu Wen, Kuen-Jang Jang, An Impingement Cooling on A Flat Surface by Using Circular Jet with Longitudinal Swirling Strips, International Journal of Heat and Mass Transfer, (46) 4657-4667, 2003

[11] Chen Yu-Yang, Yuan Zhong- Xian, Ma Chong-Fang, Experimental Study of Heat Transfer of Swirling Jet Impingement with Liquid Crystal Technique, Journal of Engineering Thermo physics, (24) 646-648 2003

S. Mohamed Illyas (1) *, Dr. B.R. Ramesh Bapu (2) and Dr. V. Venkata Subba Rao (3)

(1) Research Scholar, Department of Mechanical Engineering, University College of Engineering, JNTU Kakinada, India

E-mail Address: s_illyas@yahoo.com

(2) Professor, Department of Mechanical Engineering, RVS Padhmavathy College of Engineering and Technology, Chennai, India

(3) Associate Professor, Department of Mechanical Engineering, University College of Engineering, JNTU Kakinada, India

Printer friendly Cite/link Email Feedback | |

Author: | Illyas, S. Mohamed; Bapu, B.R. Ramesh; Rao, V. Venkata Subba |
---|---|

Publication: | International Journal of Dynamics of Fluids |

Date: | Dec 1, 2011 |

Words: | 2933 |

Previous Article: | MHD natural convection from a heated vertical wavy surface with variable viscosity and thermal conductivity. |

Topics: |