On aeroelastic response of an airfoil under dynamic stall using time delay neural network

Abstract

To clarify the mechanism of the complex aeroelastic responses of flexible blades of helicopter rotors under dynamic stall, an aeroelastic system of a 2D airfoil forced to oscillate is investigated. A reduced-order model based on time delay neural network is introduced to cut down the cost in parametric study of the aeroelastic system. The model is validated against full-order model results both with and without elasticity, which is proved to be an accurate and computational efficient method. The model is then applied to study the influence of system’s natural frequency on the aeroelastic responses, and a phenomenon of secondary resonance is found in the results. Detailed investigation and full-order simulation are performed on the cases of secondary resonances. The phenomenon is believed to be associated with the strong aerodynamic nonlinearity under dynamic stall where the instability of the vortex shedding or interaction leads to subharmonics of the aerodynamic loads. When the structure’s natural frequency is coincident with the multiples of the subharmonic, secondary resonances happen.