Research on structural improvement of microchannel heat sink by using topology optimization method based on ε-constraint model

For the comprehensive performance of microchannel heat sink (MCHS) for electronic devices, in addition to optimizing the heat transfer rate and flow power dissipation, achieving temperature uniformity is also an essential performance target that requires urgent attention. This study proposes a multi-objective topology optimization model based on the ε-constraint algorithm to design the shape and layout of microchannel heat sink, aiming to simultaneously improve heat transfer rate and temperature uniformity, and reduce flow loss. The results indicate that the established optimization model using relevant improvement schemes accurately achieves the collaborative solution of three objectives. The iterative evolution of the objective function and microchannel structure reveals the trade-off game mechanism and dynamic response between the fin generation and physical field changing for seeking the optimal solutions. The optimization algorithm responds to the continuous variation of three-objective weight factors by systematically altering the shape and global distribution state of optimized fins. The three-objective Pareto frontier indicates that there exists an optimal trade-off interval among flow loss, heat transfer rate, and temperature uniformity to achieve the optimal comprehensive performance. Compared to conventional straight channel and circular fin, representative topology optimized structures respectively reduce the pressure drop loss by 38.2 % and 45.3 %, increase the convective heat transfer coefficient by 55.5 % and 29.6 %, reduce the overall temperature variance by 45.4 % and 51.5 %. Physical field analysis reveals that the structure and distribution characteristics of topology optimized fins suppress the formation of separation and wake vortices, hinder the development of boundary layer thickness and length, and enhance local mixing and convective transport effects of the fluid.