ABSTRACT
This chapter provides the results of experimental, numerical and geometric optimization of a matrix type heat sink, a potential alternative to the 72 pin fin heat sink. The heat sink is meshed with tetrahedral meshes on the aluminium material and with hexahedral mesh on the phase change material (PCM). The operation of a heat sink takes place in two cycles, namely charging and discharging cycles. To develop the numerical model, several thermocouple measurements which include the temperature reading inside the PCM were considered to estimate the overall heat transfer coefficient. Geometric optimization of the matrix heat sink is carried out. MATLAB 2013b is used to implement the non-dominated sorting genetic algorithm (NSGA) for multi-objective optimization. Using NSGA multi-objective optimization algorithm, a set of non-dominated solutions was obtained instead a single optimum solution. The algorithm aims to minimize the distance between the observation data point and the seed point.
