Sobol sensitivity analysis and multi-objective optimization of manifold microchannel heat sink considering entropy generation minimization

Abstract

Manifold microchannel (MMC) heat sink has been widely employed as an efficient scheme for high heat flux cooling of electronic devices. The parametric optimization design of the MMC heat sink is the key factor that directly determines the thermal management performance. In this study, Sobol global sensitivity analysis is firstly applied to quantitatively evaluate the effects of four geometric design parameters and inlet volume flow rate on the MMC heat sink performance. A multi-objective optimization method combining space filling method, Back-Propagation Neural Network and NSGA-II is then established to minimum the thermal resistance, total pumping power and total entropy generation of MMC heat sink, in which TOPSIS method based on entropy weight is developed to select the best compromise solution from the Pareto-optimal solutions. Finally, the overall performance among different optimal solutions is compared to identify the effectiveness of the optimization results. The results show that the obtained optimum design parameters can provide more than 67% reduction in power consumption without scarifying thermal performance compared with the MMC heat sink published in the journal “Nature” by Erp et al. in 2020. In addition, the best compromise solution can be recommended as selectable for designer because it shows the best performance, which can boost the heat flux beyond 1000 W/cm2 with pumping power consumption 0.236 W. These results are of great significance to the design of advanced MMC heat sink for cooling electronic devices.