Numerical and experimental study of mixed convection heat transfer and fluid flow characteristics of plate-fin heat sinks

Abstract

This study applies three-dimensional computational fluid dynamics (CFD) commercial software along with the inverse method and experimental data to determine the mixed convection heat transfer and fluid flow characteristics of a plate-fin heat sink in a wind tunnel. The inverse method of the finite difference method along with the experimental temperature data is applied to determine the unknown heat transfer coefficient on the fin. Commercial software combined with various flow models is used to obtain air temperature and velocity profiles, heat transfer coefficient on fins, fin surface temperature and pressure drop. More accurate heat transfer and fluid flow characteristics can be obtained by the appropriate flow model and the number of grid points, if the resulting heat transfer coefficient and the fin temperature at each measurement location are close to the inverse results of the heat transfer coefficient and the experimental temperature data, respectively. The interesting finding is that the results obtained by the RNG k-ε turbulence model are more accurate than those by the laminar flow model. FLUENT 4 has better accuracy than FLUENT 15 along with standard wall functions and enhanced wall treatment. In addition, the total number of grid points needs to be increased with increasing air velocity and fin spacing. The dimensionless wall distance can vary with air velocity. The pressure drop has a large variation in the specific range of the fin spacing, and the secondary vortices can be found at both corners of the wind tunnel. It is worth mentioning that the strength of the secondary flow decreases with decreasing fin spacing. The effect of the flow model, near-wall treatment, FLUENT version and grid points on the results obtained cannot be ignored. To our knowledge, few researchers have used similar methods to investigate this problem in the open literature. The two proposed correlations are closer to the obtained inverse and numerical results than the existing results.