Machine Learning–Assisted Airfoil Optimization and Perspectives for Morphing Wing Applications

The design of high-performance airfoils is central to aerodynamic efficiency in aircraft and aerospace systems. Conventional optimization methods often depend on iterative computational fluid dynamics (CFD) simulations or wind-tunnel experiments, both of which are accurate but costly in time and resources. This paper introduces a machine learning (ML)–assisted framework for airfoil optimization that accelerates design while maintaining predictive reliability. A dataset of NACA airfoil profiles is used to train surrogate models capable of estimating aerodynamic coefficients. Bayesian optimization then explores the design space to identify profiles with enhanced lift-to-drag ratios. Preliminary results demonstrate that the framework generates improved geometries compared to standard baselines. Beyond static optimization, the paper outlines a conceptual extension of the methodology toward morphing wing applications, where adaptive aerostructures can dynamically transition between configurations to maximize performance. The study is presented as ongoing work, with detailed validation and implementation results reserved for forthcoming publications.