Characterization of Model-Based Uncertainties in Incompressible Turbulent Flows by Machine Learning

Abstract: This work determines the inaccuracy of using Reynolds averaged Navier Stokes (RANS) turbulence models in transition to turbulent flow regimes by predicting the model-based discrepancies between RANS and large eddy simulation (LES) models and then incorporates the capabilities of machine learning algorithms to characterize the discrepancies which are defined as a function of mean flow properties of RANS simulations. First, three-dimensional CFD simulations using k-omega Shear Stress Transport (SST) and dynamic one-equation subgrid-scale models are conducted in a wall-bounded channel containing a cylinder for RANS and LES, respectively, to identify the turbulent kinetic energy discrepancy. Second, several flow features such as viscosity ratio, wall-distance based Reynolds number, and vortex stretching are calculated from the mean flow properties of RANS. Then these flow features are regressed onto the discrepancy using a Random Forests regression algorithm. Finally, the discrepancy of the test flow is predicted using the trained algorithm. The results reveal that a significant discrepancy exists between RANS and LES simulations and ML algorithm successfully predicts the increased model uncertainties caused by the employment of k-omega SST turbulence model for transitional fluid flows.