Beamforming by Means of Inverse Variance Weighting

Conventional beamforming is one of the standard methods for localization and quantification of aeroacoustic sources in wind tunnel testing. It offers a robust indicator for the sound source distribution. The conventional beamformer minimizes the deviation between the synthetic and measured cross correlation matrix with respect to the Frobenius norm. This approach imposes the assumption that the measured cross correlations are superposed with uncorrelated white noise. In real-life applications this idealized noise model is often violated. We consider a slightly relaxed noise model where the noise is still uncorrelated but non-white. Assuming non-white noise we end up with a modified beamformer that minimizes the deviation with respect to a weighted Frobenius norm. The application of this approach on experimental data of a wind tunnel test shows that the resolution and dynamic range of the source maps can be enhanced compared to conventional beamforming.