A semi-empirical mesh strategy for CFD simulation of indoor airflow

Abstract

Proper meshing is critical for accurate computational fluid dynamics (CFD) simulation. In the three stages of CFD simulation, pre-processing accounts for about 40–60% of the total time. In this study, the performance of the grid for a large number of numerical simulations was analyzed, and base mesh sizes were recommended for different indoor geometric models. The simulation results for temperature, velocity, mesh resolution and computation time were compared, and mesh types were recommended for different models. Considering factors such as wall function, wall y+, face quality of the mesh, grid volume change rate and prism layer setting, a semi-empirical mesh strategy was presented. This strategy was verified in three cases with different geometric complexity. The verification results showed that the mesh strategy can yield reasonable simulation results. The root-mean-square error of the velocity simulation results of the optimized grid can be reduced by nearly 20%. This strategy can save nearly 50% of the mesh adjustment time. Through analysis and fitting of the grid results, the evaluation criteria for this mesh strategy were obtained. The criteria can be used to evaluate the cases using this mesh strategy.