Level estimation of extended acoustic sources using a parametric method

Array processing is a powerful toot to extract the noise source characteristics from the acoustic signals measured during wind-tunnels tests. For many years, this measurement technology is commonly being used to study airframe noise. Conventional beam forming methods have proven their efficiency in source localization, but it is well known that they are limited in the estimation of the power levels of extended sources. An approach based on a spectral estimation method (SEM) is proposed in order to determine the actual sound-pressure levels of the acoustic sources found by the localization method. The SEM is based on the array cross-spectral matrix, which contains an indirect measurement of the source power levels of interest and on a modeling of this matrix. Optimal estimates of the actual source power levels can be obtained by minimizing the mean square error between the measurement and model matrices. However, the direct minimization suffers from the sensitivity to noise inherent in the ill-posed nature of the problem. To obtain meaningful results, the problem is regularized by using a priori information. Numerical simulations involving multiple source areas are provided to illustrate the effectiveness of the estimation procedure. The method is successfully applied to compute the power levels of the main sources on a 1/11th-scaled model of the Airbus A320/A321 tested in CEPRA 19 wind tunnel.