Multi-objective optimization of high-efficiency wavy microchannel heat sinks using response surface methodology and NSGA-II

Microchannel heat sinks (MCHS) are key components for ensuring the efficient operation of data center equipment. Wavy microchannel heat sinks (WMCHS) are widely regarded as an important measure of improvement to enhance the performance of traditional MCHS. However, current research still lacks a systematic analysis of the multi-parameter coupling mechanism of WMCHS. In this study, the influence of three parameters, namely the aspect ratio (AR), wave number (Lw), and coolant mass flow rate (q_m), on the thermal performance of the wave microchannel heat sink (WMCHS) was investigated through univariate analysis, and the optimal value range of each parameter was determined. The interaction effects and significance of these parameters were further clarified using analysis of variance (ANOVA). Numerical simulations were employed to obtain results under varying parameters, and a surrogate model was constructed using the response surface methodology (RSM). Finally, the average Nusselt number (Nu), maximum CPU temperature difference (ΔT_CPU,max), and inlet-outlet pressure drop (ΔP) were optimized using the NSGA-II algorithm, yielding the Pareto optimal frontier. With Nu as the primary optimization objective and ΔP as the secondary objective, an optimal solution was selected from the Pareto frontier. The results show that when AR=4.0013, Lw=10, and q_m=0.0085 kg/s, Nu increases by 20.5% and ΔT_CPU,max decreases by 27.5%. Although the pressure drop at the inlet and outlet has increased, the temperature distribution within the flow channel has become more uniform, and the coolant flow has become more vigorous, effectively enhancing the heat exchange effect. This study can provide engineering references for the subsequent structural optimization design of WMCHS.