What is it about?
To this day, there is no model for laminar to turbulent transition physics that can be implemented in low-cost simulations. The current models are based on empirical correlations, and sometimes their performance is not great. We show a methodology to calibrate the models to match experimental data using artificial intelligence.
Featured Image
Photo by Stockholm Paris Studio on Unsplash
Why is it important?
It is important because there is no consistent way of modeling transition, and it can play a huge factor in some problems. With these results and this methodology, industry and academia can tune their models so that it is easier to find a realization that produces values that match experimental references. Those experimental references could come from new designs they are testing, and if they had a computational setup that mimic them, they could extract very valuable insights that otherwise would be impossible to obtain.
Perspectives
I think this publication offers a methodology and results that could be very valuable for engineers that are looking to obtain better approximations to their problems. It also digs into the intricacies of transition modeling through a state-of-the-art model, and that discussion could be helpful for other people trying to enhance existing models or creating their own.
Javier Capel Jorquera
Universidad Politecnica de Madrid
Read the Original
This page is a summary of: Deep Learning for Inverse-Problem-Based Calibration of the γ Transition Model, AIAA Journal, July 2025, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/1.j065391.
You can read the full text:
Contributors
The following have contributed to this page
