ANN-GA optimisation of rectangular fin array with closed top plate under forced convection

Abstract

Heat fins play an important role in removing undesired heat generated in many electronic applications. Heat fins are extended surfaces which are provided to enhance heat transfer. Quite a number of studies have been done on conventional open top rectangular fins. However, there is a necessity to increase heat transfer rate for particular fin geometry with geometric constraints. In the present study, experiments are done in a standard heat sink with open top and closed top rectangular fins. The numerical model of rectangular fins with closed top plate under forced convection conditions is validated with the experimental results. Studies are done using commercial CFD code ANSYS FLUENT©. The length, width and height of fins are kept constant. Numerical study is done by varying velocity, number of fins, top plate thickness, fin thickness and heat duty. The results of numerical analysis show that fins with closed top plate have enhanced heat transfer rate than that of fins without closed top plate. The base plate temperature is predicted by training the artificial neural network (ANN) based on the data obtained from numerical simulations. Dimensions to achieve minimum base plate temperature are predicted by applying genetic algorithm (GA) on the designed neural network.