Multi-objective optimization of a composite phase change material-based heat sink under non-uniform discrete heating

This paper presents a novel heat sink that integrates paraffin/copper foam composite phase change material. The heat sink base is subject to non-uniform heating with multiple discrete heat sources on a circuit board. Numerical simulations are performed using FLUENT software to investigate its thermal transfer characteristics and performance. Three types of paraffin, including RT35HC, RT44HC, and RT54HC, are compared. An optimization scheme that combines Artificial Neural Network and Non-Dominated Sorting Genetic Algorithm-II is proposed to (a) maximize the operation time and (b) minimize the heat sink mass. Operation time is defined as the time for simulated chips to reach set point temperature, 80 °C. The result shows that local overheating caused by non-uniform heating limits operation time. The optimization method shows great robustness and effectiveness. Three sets of non-dominated solutions are obtained, and optimal configurations are determined with further analysis. A figure of merit is defined as the ratio of charging time to heat sink mass. Optimal configuration's figure of merit increases by 1.2–44.0% comparing to original designs.