Sandia researchers utilized machine learning techniques to address the limitations of Reynolds-averaged Navier-Stokes (RANS) turbulence models in predicting hypersonic turbulent flows, with a particular emphasis on inaccuracies in wall heating predictions for flows involving shock boundary layer interactions.
This research has led to the development of neural-network-based machine learned turbulence models that provide state-of-the-art predictions for a range of hypersonic flows, along with a deeper understanding of RANS modeling for hypersonic flows and an understanding of current deficiencies in Sandia’s modeling toolkit. This project, a collaboration with the Sandia National/Regional partner University of Michigan, resulted in published papers and presentations at conferences organized by the American Institute of Aeronautics and Astronautics (AIAA) and NASA. The findings are relevant to various mission areas, including nuclear deterrence. Elements of the research are ongoing under the Advanced Simulation and Computing initiative, which aims to extend the applicability of the developed models.
Sandia experts linked to work
- Eric Joshua Parish
Read more
July 14, 2025