Data-driven Turbulence Modelling with the Machine Learning Toolbox of the CFD Solver TRACE

Kurzfassung

The presentation shall give an overview over the machine learning interface of the flow solver TRACE, developed at the Institute of Propulsion of the German Aerospace Center (DLR) as well as recent work and results achieved using it. Firstly, focus is laid on the general philosophy of augmenting turbulence models using data-driven methods. Secondly, features and design of the TRACE machine learning toolbox are presented. Lastly, a selection of recent results is shown