Active control of flow-induced acoustic resonance through surface perturbation

Abstract

Vortex shedding after a bluff body in the cross flow creates alternating lift and drag forces on the rear surface of the body, which can cause serious structural vibrations, and considerable acoustic noise particularly due to the acoustic resonance with a downstream cavity. This phenomenon can be classified as a type of flow-structure-sound (FSS) interaction. FSS is a very complex and challenging research area, which is relevant to a large variety of applications in mechanical, civil and environmental engineering. One aspect awaiting technological breakthrough is the development of innovative technologies for stabilizing flow and suppressing flow-induced vibration vortex-induced noise at the same time. This issue motivated the present research, in which the active control of flow-induced acoustic resonance has been experimentally investigated by using a novel surface perturbation technique. Seven interrelated subtopics are addressed in the present thesis. Major conclusions are: 1) The vortex shedding from a semi-circular leading edge test model and its associated acoustic resonance was controlled by utilizing a novel surface perturbation technique in the open-loop control scheme. It was observed that the control performances were repeatable and reliable. A reduction of sound pressure level (SPL) of about 16.3dB in the duct and 21.3dB in the cavity was achieved by using the optimal control parameters. 2) During the generation of the trailing edge vortex shedding for the semi-circular leading edge test model, a pressure pulse was generated and feedback to the leading edge shear layer to influence the formation of the next vortex shedding. The surface perturbation technique generates a small local perturbation in the space between leading edge shear layer and trailing edge, which disturbs the pressure pulse. The change of the pressure pulse alters the generation of next vortex shedding that can lead to a reduction in the vortex shedding strength. This mechanism is referred to as 'Pressure Disturbance Mechanism'. Furthermore, the surface perturbation technique changed the geometry of the up surface of test model and shifted the shear layer which was attached to the surface. The shift disturbed the shear layers around the test model, and thus resulted in a clear disturbance on the formation of the trailing edge vortices, leading to a significant reduction in the vortex strength in the wake of the model. 3) A frequency shift phenomenon was observed in both open-loop and closed-loop control scheme, and its effect on suppressing acoustical resonance inside the cavity was discussed. A formula was derived to predict this frequency shift in the effective control region. It was shown that the applied perturbation brings about a small increase in the effective height of the control plate, resulting in a frequency shift. If the shedding frequency shift exceeded the resonance bandwidth of the downstream acoustic cavity, the sound reduction inside the cavity would be larger than that in the duct. This phenomenon along gave a rise to a further sound pressure reduction of 5 dB for open-loop control and 4.4dB for the closed-loop control inside the acoustic cavity, respectively. 4) Closed-loop tests were proposed for a semi-circular leading edge test model, along with a down-sampling control algorithm. It was observed that the closed-loop control could achieve a better control performance than that of the open-loop control. At the optimum control voltage and control phase delay, a noise reduction of 17.5 dB in the duct and 22.6 dB inside the cavity was obtained. In particular, the phase delay of control actuation could be optimally tuned so that the strength of vortex shedding energy could be minimized effectively, leading to a satisfactory noise reduction in the duct and cavity. This process was evident from the spectral phase shift results, where the vortex traveling time has been delayed at downstream of the test model. 5) For the square leading edge test model, the developed surface perturbation technique was also effective. However, the control mechanism was found to be different from that of the semi-circular leading edge test model. The path of the leading edge vortex shedding started from the leading edge and then propagated downstream. The vertical velocity generated by the perturbation played the key role in the vortex strength reduction. More significant reductions can be achieved when the velocity disturbance was near the propagation path of the vortex shedding. 6) The control strategies for two primary leading edge geometries have been investigated. The optimal control strategy for the semi-circular leading edge test model consists in arranging the perturbation at the position where the pressure pulse was the smallest. This way, the targeted noise reduction can be achieved by using perturbation within a relatively small area at a low control voltage. While for the square leading edge test model, the optimal control strategy requires applying the perturbation near the propagation path of the LEVS for achieving sufficient vortex strength abasement. 7) The energy distribution at vortex shedding frequency along the up surface of test model measured by the hot wire was used to identify the existence of the pressure pulse for the semi-circular leading edge test model. It was found that the pressure pulse had strong directivity characteristics; therefore, a small disturbance generated by the surface perturbation can change the direction of the pressure pulse and then influence the generation the vortex shedding.