Study of transfer learning from 2D supercritical airfoils to 3D transonic swept wings

TL;DR

This study demonstrates that transfer learning from 2D supercritical airfoil models can effectively predict 3D swept wing flows, significantly reducing sample requirements and prediction errors in fluid mechanics applications.

Contribution

The paper introduces a transfer learning approach that reuses 2D models for 3D wing predictions, embedding swept theory to enhance accuracy and reduce data needs.

Findings

Transfer learning achieves good accuracy with ~500 samples.

Transferred models reduce prediction errors by 60% and 80%.

Embedding swept theory improves 3D flow predictions.

Abstract

Machine learning has been widely utilized in fluid mechanics studies and aerodynamic optimizations. However, most applications, especially flow field modeling and inverse design, involve two-dimensional flows and geometries. The dimensionality of three-dimensional problems is so high that it is too difficult and expensive to prepare sufficient samples. Therefore, transfer learning has become a promising approach to reuse well-trained two-dimensional models and greatly reduce the need for samples for three-dimensional problems. This paper proposes to reuse the baseline models trained on supercritical airfoils to predict finite-span swept supercritical wings, where the simple swept theory is embedded to improve the prediction accuracy. Two baseline models for transfer learning are investigated: one is commonly referred to as the forward problem of predicting the pressure coefficient distribution based on the geometry, and the other is the inverse problem that predicts the geometry based on the pressure coefficient distribution. Two transfer learning strategies are compared for both baseline models. The transferred models are then tested on the prediction of complete wings. The results show that transfer learning requires only approximately 500 wing samples to achieve good prediction accuracy on different wing planforms and different free stream conditions. Compared to the two baseline models, the transferred models reduce the prediction error by 60% and 80%, respectively.