Turbulence Modeling for Physics-Informed Neural Networks: Comparison of Different RANS Models for the Backward-Facing Step Flow

Physics-informed neural networks (PINN) can be used to predict flow fields with a minimum of simulated or measured training data. As most technical flows are turbulent, PINNs based on the Reynolds-averaged Navier–Stokes (RANS) equations incorporating turbulence model needed. Several studies demonstrated capability solve Naver–Stokes for laminar flows. However, little work has been published concerning application RANS turbulent This study applied RANS-based PINN approach backward-facing step at Reynolds number 5100. The standard k-ω model, mixing length an equation-free νt and pseudo-Reynolds stress were applied. results compared favorably DNS data when provided three vertical lines labeled For five data, all models predicted separated shear layer associated vortex more accurately.