PreprintArticleVersion 1Preserved in Portico This version is not peer-reviewed
Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack
Version 1
: Received: 28 April 2023 / Approved: 30 April 2023 / Online: 30 April 2023 (03:02:34 CEST)
How to cite:
Harmening, J.H.; Pioch, F.; Fuhrig, L.; Peitzmann, F.; Schramm, D.; el Moctar, O. Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack. Preprints2023, 2023041244. https://doi.org/10.20944/preprints202304.1244.v1
Harmening, J.H.; Pioch, F.; Fuhrig, L.; Peitzmann, F.; Schramm, D.; el Moctar, O. Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack. Preprints 2023, 2023041244. https://doi.org/10.20944/preprints202304.1244.v1
Harmening, J.H.; Pioch, F.; Fuhrig, L.; Peitzmann, F.; Schramm, D.; el Moctar, O. Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack. Preprints2023, 2023041244. https://doi.org/10.20944/preprints202304.1244.v1
APA Style
Harmening, J.H., Pioch, F., Fuhrig, L., Peitzmann, F., Schramm, D., & el Moctar, O. (2023). Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack. Preprints. https://doi.org/10.20944/preprints202304.1244.v1
Chicago/Turabian Style
Harmening, J.H., Dieter Schramm and Ould el Moctar. 2023 "Data-Assisted Training of a Physics-Informed Neural Network to Predict the Reynolds-Averaged Turbulent Flow Field around a Stalled Airfoil under Variable Angles of Attack" Preprints. https://doi.org/10.20944/preprints202304.1244.v1
Abstract
Physics-informed neural networks are a promising method to yield surrogate models of flow fields. We present a metamodeling technique for variable geometries based on physics-informed neural networks. The method was applied to the DU99W350 airfoil at a Reynolds number of 1×105. The model predicted the Reynolds-averaged velocity and pressure field around the airfoil for arbitrary angles of attack between 10.0° and 17.5°. The model was trained with data from CFD simulations for a limited set of angles of attack. Additionally, satisfaction of the a priori known boundary conditions as well as the Reynolds-averaged Navier-Stokes equations were trained. A sensitivity analysis concerning the Reynolds number, the amount and distribution of training data, and the turbulence model was conducted showing the superiority of the pseudo-Reynolds stress method and the demand of labeled training data in the domain. The trained network was capable of predicting the developing flow separation on the suction surface and exhibited excellent agreement with CFD results even in the proximity to the wall for interpolations as well as extrapolations from the labeled data set.
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.