MMC-based heat sink topology optimization design for natural convection problems

Abstract

The optimal design of heat sink using Moving Morphable Components (MMC) is presented for natural convection problems. The topology optimization design of the heat sink has a heavy computational burden. For one thing, the multi-physics model is a fully coupled nonlinear system, and for another, a large number of degrees of freedom (DOF) of the design variables are involved in the optimization process. The governing equations of the multi-physics model are composed of the incompressible Navier-Stokes equations related to velocity and pressure and the heat transfer equation related to temperature. A reduced-order model that simplifies physical quantities is introduced to reduce the number of state variables, and the Gaussian Seidel iteration algorithm is used to further reduce computing scale. Compared with the density-based topology optimization, the MMC-based topology optimization shows similar heat sink design through fewer design variables and good heat dissipation performance. In addition, the heat sink structure with fully explicit boundary information is obtained. The numerical results show that the number of components and the lower bounds of the components have an effect on the heat dissipation performance. As the number of components increases, the heat dissipation performance improves. However, when the number of components reaches a certain level, the heat dissipation performance reaches saturation. Moreover, the higher the processing accuracy of the mechanical equipment, the smaller the lower bounds of the components that can be designed, and the better the heat dissipation performance.