Numerical simulation of two-dimensional backward-facing step flow using OpenFOAM

To enhance the accuracy of turbulence modeling in incompressible flows and to develop efficient turbulence models, this study builds upon the previously proposed Partial Averaged Fluctuation Velocity （PAFV） turbulence model and introduces several improvements. By incorporating lateral fluctuation velocity to adjust the viscosity coefficient, a turbulence model was developed that is essentially free of empirical constants and primarily applicable to incompressible flow computations, with improved capability in predicting near-wall boundary layer separation. The model solver was implemented on the open-source platform OpenFOAM, and high-resolution numerical simulations were conducted for two-dimensional backward-facing step flows to obtain instantaneous velocity fields, fluctuation velocity distributions, and Reynolds stress data. The results demonstrate that, compared with the commonly used Spalart–Allmaras （S-A） and k–ω SST turbulence models, the PAFV model exhibits better agreement with theoretical values in certain regions. Due to the introduction of lateral fluctuation velocity and the reduction of viscosity, the new turbulence model shows enhanced sensitivity in capturing vortices in the near-wake region downstream of the step, thereby providing stronger capability in vortex prediction and more accurate characterization of turbulent boundary layer separation. However, in terms of capturing the upper-wall vortex region, this model performs worse than traditional turbulence models, and there is a lag in predicting the far-field vortex structure, which needs further optimization. This study provides new ideas for near-wall flow optimization, and future work will be dedicated to improving the universality of the model.