Sound-source-oriented Bayesian optimization for efficient airfoil tonal noise control

Abstract

This study optimizes tonal noise suppression for a National Advisory Committee for Aeronautics 0012 airfoil. The airfoil operates at a Reynolds number of 5000 and an angle of attack of 5°. A sound-source-oriented Bayesian optimization framework integrated with multi-process acoustic theory is adopted. This framework optimizes the position and angle of steady wall blowing/suction under different amplitude conditions. It effectively overcomes the challenges of high-dimensional parameter space and time-delay effect in tonal noise control. Mode decompositions and continuous wavelet analysis are employed. The spatiotemporal features, energy sources, and instability origins of the acoustic modes are characterized in detail under varying control amplitudes. These analyses elucidate the underlying control mechanisms. The results indicate that the optimal suction locations are concentrated around the flow separation point of the uncontrolled baseline airfoil. The optimal suction angle is maintained near the wall-normal direction. As the control amplitude increases, the optimal suction position gradually shifts downstream. Concurrently, the dominant instability undergoes two distinct transition stages: from double-vortex shedding to single-vortex shedding and ultimately to suppression of the wake instability. Notably, the transition to single-vortex shedding yields the highest noise reduction efficiency. This shift in dominant mechanisms is accompanied by a migration of the instability mode's peak amplitude away from the wall. Energy analysis reveals that the instability modes are sustained by both the mean flow and nonlinear interactions. Both of these contributions are significantly attenuated under active control.