Parametric Generative Schemes with Geometric Constraints for Encoding and Synthesizing Airfoils

TL;DR

This paper introduces two deep learning-based generative schemes for airfoil design that effectively capture complex geometries while satisfying constraints, enhancing flexibility and accuracy over traditional methods.

Contribution

The paper presents novel CVAE and VAE-based generative models with geometric constraints for airfoil synthesis, improving diversity and adherence to design specifications.

Findings

Soft-constrained scheme produces accurate, smooth airfoils with slight constraint deviations.

Hard-constrained scheme generates diverse, constraint-adherent airfoils with broad objective space coverage.

Both methods outperform traditional techniques in design flexibility and optimization efficiency.

Abstract

The modern aerodynamic optimization has a strong demand for parametric methods with high levels of intuitiveness, flexibility, and representative accuracy, which cannot be fully achieved through traditional airfoil parametric techniques. In this paper, two deep learning-based generative schemes are proposed to effectively capture the complexity of the design space while satisfying specific constraints. 1. Soft-constrained scheme: a Conditional Variational Autoencoder (CVAE)-based model to train geometric constraints as part of the network directly. 2. Hard-constrained scheme: a VAE-based model to generate diverse airfoils and an FFD-based technique to project the generated airfoils onto the given constraints. According to the statistical results, the reconstructed airfoils are both accurate and smooth, without any need for additional filters. The soft-constrained scheme generates airfoils that exhibit slight deviations from the expected geometric constraints, yet still converge to the reference airfoil in both geometry space and objective space with some degree of distribution bias. In contrast, the hard-constrained scheme produces airfoils with a wider range of geometric diversity while strictly adhering to the geometric constraints. The corresponding distribution in the objective space is also more diverse, with isotropic uniformity around the reference point and no significant bias. These proposed airfoil parametric methods can break through the boundaries of training data in the objective space, providing higher quality samples for random sampling and improving the efficiency of optimization design.