Data-Driven Transition Models for Aeronautical Flows with a High-Order Numerical Method

Over the past years, there has been innovative ideas about data-driven turbulence modeling proposed by scholars all over world. This paper is a continuity of these significant efforts, with aim offering better representation for physics. Previous works mainly focus on viscosity or Reynolds stress, while are few transition. In our work, two mapping functions between average flow parameters and transition intermittency, virtual physical quantity describing amount at given position, refactored, respectively, neuron networks random forests. These then coupled Spalart–Allmaras (SA) model to reconstitute models prediction. To demonstrate that provide improved prediction accuracy compared previous SA models, we conduct test cases under high-order weighted compact nonlinear scheme (WCNS). The results both significantly capture natural transitions occurring in flows. Furthermore, interpolation generalisation extrapolation abilities also demonstrated this paper. emphasize potential machine learning as supplementary modeling.