Forecasting Lift of Airfoil Randomly Maneuvering Under Dynamic Stall Conditions

Dynamic stall is an abrupt separation of flow from an airfoil under rapid motion, characterized by delayed stall and subsequent plummet of lift. This work forecasts the lift of a randomly maneuvering airfoil under dynamic stall conditions by using sparsely spaced surface pressure data, which the authors believe can maximize the effectiveness of various dynamic stall suppression schemes. Using data from large-eddy simulations, the authors demonstrate that a long short-term memory network delivers accurate predictions and that two pressure sensors positioned near the leading edge, on different sides of the airfoil, capture most of the information needed to forecast lift. Further analysis of the flowfield around a periodically pitching airfoil identifies the characteristic frequency range of pressure fluctuations amplified during dynamic stall and associates them with the chord-length-scale vortex dynamics. However, evidence hints that the forecast mostly relies on even lower-frequency components tied to the airfoil motion. This suggests that predicting the incoming flow or the wing’s structural response, rather than focusing on the precise vortex dynamics, may be more helpful in forecasting forces for practical engineering applications.