Aero-acoustic response of an acoustic liner

Abstract

In this thesis, the resistance of a single degree of freedom liner is investigated experimentally. Semi-empirical resistance modelling fails at higher Mach numbers (M₀>0.3) and sound pressure levels (SPL>130dB), and experimental work is missing in these conditions. An experiment combining in-situ impedance measurements and phase-locked particle image velocimetry is selected. This combined approach allows to evaluate the resistance, in-orifice velocity, friction velocity, vorticity and turbulence intensity.

The high grazing velocity and low acoustically induced velocity form a wide range of velocity scales that the PIV measurement must resolve. This sets a requirement for the dynamic velocity range that typical PIV with advanced interrogation techniques is not expected to achieve. A multi-frame approach is selected, which combines measurements with different pulse separation times. A second benefit of such a multi-dt approach is that information from different pulse separation times allows to decompose the PIV measurement noise from the turbulent fluctuations. Limits on the pulse separation time are seen in the freestream velocity, acoustic frequency, correlation strength and spatial filtering. The resulting timing ranges are evaluated before the experiment, and experimentally confirmed to be appropriate.

The decomposition between PIV measurement noise and turbulence is shown to be not applicable to the current data, as the measurement noise is correlated with the pulse separation time. This is caused by turbulence intensity and laser sheet reflections near the liner's face sheet. A DVR of 600 is achieved, and an improvement with respect to a grid refinement approach is seen. The difference between multi-dt and grid refinement vanishes towards the wall, and the DVR itself is also seen to decrease, due to the increase in measurement noise towards the wall. It is shown that the DVR must exceed the expected range of velocity scales by approximately 25% to accurately resolve phase-dependent quantities.

Results for the friction velocity found from a momentum-integral method and from a log-law fit, modified for surface roughness, transpiration and inclusion of the outer layer, agree well with direct drag measurements. The boundary layer is seen to develop in a periodic manner over the orifices, and the influence of the aero-acoustic interaction at the liner on the local friction velocity can not be neglected. The in-orifice velocity fluctuation RMS is consistent between the in-situ impedance and PIV measurements. A lumped-element model suggested in literature is shown to accurately predict the in-orifice velocity scale, but it requires an accurate value for the orifice quality factor. The quality factor is associated to the reduction in effective open orifice area, known as a vena contracta. It is thereby inversely proportional to the friction velocity, and the in-orifice velocity scale. The current results on the quality factor agree qualitatively and quantitatively to statements and suggestions from literature. The resistance is confirmed to be a function of the friction velocity and the in-orifice velocity \gls{rms}, with both velocity scales associated to vena contracta. A velocity scale associated to the intensity of vortex shedding from the orifices is introduced by integrating the acoustically induced vorticity near the wall. The onset of non-linear resistance is confirmed at a ratio between the in-orifice and friction velocity scales of 2, and is clearly linked to the intensity of vortex shedding.