Aeroacoustic Measurements in Open-jet Wind Tunnels - An Evaluation of Methods Applied to Trailing Edge Noise

Abstract

Open jet wind tunnels have become commonplace in aeroacoustic testing, as having an open-jet test section in an anechoic chamber can provide a near-anechoic environment when attempting to measure acoustic field levels due to an aeroacoustic source. Additionally, reduced flow is present over installation microphones, so SNR gains are observed versus equivalent closed test section experiments. However, open-jet wind tunnels and their acoustic treatment introduce additional noise sources which may contaminate a given signal of interest. To overcome this limitation, multiple-microphone processing techniques involving coherent power and/or beamforming are often leveraged. These techniques can isolate the behavior of an aeroacoustic noise source using different types of assumptions. However, depending on the facility background noise characteristics and measurement setup these assumptions may be violated. A comparison of these techniques is conducted using a NACA 63-215 Mod-B airfoil in UFAFF, and shows that when the signal of interest, trailing edge noise, is not the dominant noise source, facility flow noise can introduce aberrant behavior to the coherent power and array processing techniques. This behavior leads to correspondingly large uncertainty bounds in output power spectra. Modifications to the wind tunnel facility and experimental setup are proposed to mitigate these problems, and a new experiment set with a smaller NACA 0012 airfoil is designed. Results show that modifications appear to improve the behavior of coherent power methods, specifically those tailored to trailing edge noise measurements, significantly. Major disagreement between coherent power techniques and delay-and-sum integration is found even with significant mitigation of facility contaminating noise sources. Uncertainty analysis does not account for the difference visible between the methods, suggesting that further analysis involving deconvolution and coherent source behavior is warranted.