The research in this thesis aims at improving aeroacoustic testing in acoustically disturbed environments, with particular focus on closed test section wind tunnels. Ideally, it should become possible to perform fully reliable aeroacoustic experimental campaigns without the need of anechoic open—jet facilities.

The investigation of wall cavities for microphone placement was done with Computational Fluid Dynamics simulations, which in turn were validated experimentally. Cavities covered with a mesh cover were investigated, since previous literature shows that these reduce hydrodynamic noise while allowing for the transmission of waves to the microphones. The numerical simulations show that covering microphone cavities with a mesh cover results in a stagnant flow inside the cavities. As consequence, the only source of pressure fluctuations at the cavity bottoms are acoustic waves. The hydrodynamic pressure fluctuations from the wall’s boundary layer still propagate acoustically to the cavity bottoms. The findings show that increasing cavity size, by increasing the cavity opening diameter with respect to the length of the eddies in the turbulent boundary layer, leads to a lower propagation of spurious pressure fluctuations to the cavity bottoms. This in turn leads to an increased signal to noise ratio of acoustic measurements recorded at closed test section wind tunnels.

Wind tunnel wall liners have been characterized based on their viscous resistivity, inertial resistivity and roughness. Several porous liners have been tested experimentally. The aim was to analyze their impact on the aerodynamic properties of the wind tunnel boundary layer, on the generation of spurious noise, and on the absorption of acoustic reflections. The results show that the ideal choice of liner consists of a liner: with high viscous resistivity, which leads to high acoustic absorption; with low roughness, to reduce the impact on the wind tunnel wall’s boundary layer; and with low inertial resistivity, to reduce the generation of spurious noise. The best lining material tested was melamine foam.

The acoustic propagation and acoustic interference in a closed wind tunnel test section were predicted with a FEM acoustic solver. The propagation was modelled on a baseline test section, with fully reflective walls, and on test sections with lined walls. The numerical results were found to give a very accurate prediction of the acoustic experimental tests. It was possible to use the numerical results to improve the post—processing of experimental data. The Green’s function used to process experimental microphone data with beamforming was corrected, using the numerical results. Beamforming with the Green’s function corrected for the acoustically disturbed environment led to a higher beamforming spatial resolution. In addition, the estimated noise levels are more accurate when the correction is used. This improved approach was shown to work for post—processing experimental measurements of a monopole sound source placed at the center of the test section, with and without free—stream flow.