On 4th February 2019, Romain DUPUIS has defended his thesis. Awarded by the University of Toulouse III – Paul Sabatier, his work was supervised by both CERFACS, MEGEP doctoral school, and IRT Saint Exupéry.
The high quality and the relevance of Romain thesis contributed to get significant results as a part of MDA-MDO project.
“Surrogate Models Coupled with Machine Learning to Approximate Complex Physical Phenomena Involving Aerodynamic and Aerothermal Simulations”
About his thesis
Numerical simulations provide a key element in aircraft design process, complementing physical tests and flight tests. They could take advantage of innovative methods, such as artificial intelligence technologies spreading in aviation. Simulating the flight mission for various disciplines pose important problems due to significant computational cost coupled to varying operating conditions. Moreover, complex physical phenomena can occur. For instance, the aerodynamic field on the wing takes different shapes and can encounter shocks, while aerothermal simulations around nacelle and pylon are sensitive to the interaction between engine flows and external flows. Surrogate models are used to substitute expensive high-fidelity simulations by mathematical approximations in order to reduce overall computation cost and to provide a data-driven approach.
In this thesis, we propose two developments: (i) machine learning-based surrogate models capable of approximating aerodynamic experiments and (ii) integrating more classical surrogate models into industrial aerothermal process. The first approach mitigates aerodynamic issues by separating solutions with very different shapes into several subsets using machine learning algorithms. Moreover, a resampling technique takes advantage of the subdomain decomposition by adding extra information in relevant regions. The second development focuses on pylon sizing by building surrogate models substituting aerothermal simulations. The two approaches are applied to aircraft configurations in order to bridge the gap between academic methods and real-world applications. Significant improvements are highlighted in terms of accuracy and cost gains.
MDA MDO Project
Methods and tools for multi-disciplinary analysis and optimization (aeronautical study cases)
|M.V. SALVETTI||Rapporter||Professor / University of Pisa|
|C. CORRE||Rapporter||Professor / Lyon Central School|
|A. IOLLO||Examiner||Professor / University of Bordeaux|
|P. BREITKOPF||Examiner||Engineer / CNRS Search Center, Technological University|
|P. SAGAUT||Thesis Director||Professor / University of Aix, Marseille|
|J.C. JOUHNAUD||Thesis co-Director||Senior researcher / CERFACS|
|A. GAZAIX||Invited / Thesis Advisor||Project Manager / IRT Saint Exupéry|
|C. LOURIOU||Invited||Engineer / Airbus|
- Surrogate modeling of aerodynamic simulations for multiple operating conditions using machine learning – R. Dupuis, J.C. Jouhaud, P. Sagaut – AIAA Journal, vol. 56, pp. 3622-3635 – 2018 – DOI
- Aerodynamic data predictions for transonic flows via a machine-learning-based surrogate model – R. Dupuis, J.C. Jouhaud, P. Sagaut – AIAA – ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, AIAA SciTech Forum – Kissimmee, Florida – 2018 – DOI