Numerical study of convective heat transfer behaviors using the MPSCV method

This paper introduces the MPSCV method, a novel particle-grid hybrid numerical approach for simulating convective heat transfer. By combining the strengths of the finite volume method (FVM) and the moving-particle semi-implicit (MPS) method, this novel particle-grid hybrid method addresses the limitations of the MPS method in accurately calculating thermal conduction and boundary temperatures. The method employs the MPS particle interaction model to compute fluid flow and generates grids based on particle positions. It then utilizes the FVM to solve the energy equation, ensuring precise partitioning of particle control volumes. Extensive testing, covering 1D heat penetration under different conditions and square cavity natural convection at various Rayleigh numbers, verifies the method’s high accuracy and effectiveness. Compared to the MPS method, the particle-grid hybrid approach demonstrates a marked enhancement in simulation precision, overcoming issues of overestimating thermal conduction and inaccurately calculating boundary temperatures. In addition, MPSCV outperforms other particle methods in natural convection simulations and has been successfully applied to 3D calculations with satisfactory results. In terms of computational efficiency, MPSCV exhibits performance comparable to the MPS method in 2D scenarios, while requiring slightly more time for 3D calculations.