Airfoil Optimization in Propeller Slipstreams Using Generative Adversarial Networks

This study explores airfoil design optimization in propeller slipstreams, leveraging the capabilities of Generative Adversarial Networks (GANs). With advancements in AI, the research integrates a GAN-based airfoil generation algorithm, emphasizing its benefits in input dimension reduction and curve quality. Using Information Maximizing Generative Adversarial Networks (info-GAN), essential airfoil features are extracted, showcasing the network's inferential prowess. The focus then shifts to propeller-wing coupled optimization, where a 6% drag reduction was achieved using rigorous validation techniques. The paper introduces a novel method, substituting expert feedback with GAN's discriminator in airfoil optimization. This approach not only meets design point criteria but also enhances robustness and applicability, aligning with real-world engineering scenarios. In summary, this work presents a novel approach to airfoil design in propeller slipstreams through GANs.