Unsteady Aerodynamic Wake Modes in Pitching Airfoils: Machine Learning Enabled Classification and Prediction

Accurate prediction of aerodynamic wake structures is critical for the analysis and design of bio-inspired propulsion systems. This study investigates the sinusoidal pitching oscillations of a teardrop airfoil using a computational approach. The study examines the wake structure for a range of pitching frequencies and amplitudes at a constant Reynolds number. The identification of the wake structure is based on vortex shedding from the airfoil and the arrangement of these vortices in the wake. The transformation of the wake behind the airfoil in a laminar flow is presented for various pitching parameters. The study discusses how the wake structure influences aerodynamic forces and propulsion efficiency. High-fidelity computational fluid dynamics (CFD) simulations are integrated with supervised machine learning to classify and predict these wake patterns. The dataset generated from high-fidelity simulations comprises approximately 2500 samples spanning ten distinct wake categories. The dataset captures the relationships between pitching frequency, pitching amplitude, and the resulting vortex wake structures. Different machine learning models are employed to classify and predict the wake patterns that directly impact thrust and drag. Each model's performance is evaluated using standard classification metrics to identify the most effective predictive framework. An optimum model with the highest prediction accuracy is proposed. The proposed model also demonstrated strong robustness to class imbalance within the dataset, highlighting its suitability for characterizing complex and unsteady flow phenomena in fluid dynamics. The proposed data-driven approach significantly reduces computational cost by enabling rapid prediction of wake structures, effectively bridging high-fidelity numerical simulations with intelligent analytics. These findings lay the groundwork for future advancements in propulsion systems, energy harvesting systems, and aerodynamic optimization.