Improvement on the cyclic thermal shock resistance of the electronics heat sinks using two-objective optimization

Abstract

Phase change materials (PCMs) can isothermally absorb heat in electronic devices, and optimizing PCM heat sink designs to improve heat dissipation is therefore important to improve the performance of a system, including its ability to withstand cyclic thermal shocks. In this study, a low melting point alloy (LMPA)-based heat sink with crossed copper fins is developed and optimized to cope with cyclic ultra-high thermal shocks (100 W/cm²). Gallium is selected as an LMPA candidate due to its high thermal conductivity. The numerical model is validated against available experimental results where the transient temperature is predicted with a maximum discrepancy of 4.9%. The main goal of the study is to determine the design configurations that minimize the heat sink peak temperature after thermal shock and with minimal cooling time. An approach that couples the Response Surface Methodology (RSM) with numerical simulations is applied for this purpose. The height and thickness of fins and the PCM thickness, are treated as design variables in the two-objective optimization. Two correlations are presented to estimate the peak temperature and cooling period of the system in presence of thermal shock. The effect of the three parameters is thoroughly analysed following the optimization process. The optimum values for fin height, fin thickness and PCM thickness are found to be 8.77 mm, 0.55 mm, and 1.51 mm, respectively, where the PCM volumetric percentage is found to be 53.73%. Single-objective optimizations are also conducted to design a system satisfying a minimum peak temperature or minimum cooling period separately. Overall, the cyclic evaluation shows that the two-objective optimized heat sink delivers peak temperature up to 8.63% and 36.26% lower than that of single-objective optimized systems after five cycles of thermal shocks within periods of 30–60 s. This study shows how LMPA-based heat sink may be optimized and demonstrates a potential approach in the design of novel thermal management systems to prevent overheating from thermal shocks.