Spatio-Temporal Flow Field Prediction of Turbulent Boundary Layer Based on PINN

Based on the PINN neural network for the study of turbulence, it can solve the nonlinear equations without grid, which is different from the CFD method. This paper mainly uses the PINN method to study the temporal and spatial prediction of the turbulent boundary layer flow field, and in the absence of label values, based on the known flow field information, the unknown information is reasonably predicted by solving the N-S equation. The comparison of PINN prediction results with numerical simulation and experimental results verifies the feasibility of applying PINN to the study of turbulent boundary layer.