A Fast Bayesian High-Resolution Localization Method for Rotating Blade Noise

The high-resolution blade noise identification problem of axial fans is studied based on the Modal Composition Beamforming (MCB) method in this paper. To simulate blade defects, a vortex generator (VG) is put on the fan blade. Concomitatively, Phase Average Beamforming (PA-BF) and Rotating Source Identifier (ROSI) methods are also applied to process experimental signals to compare and analyze the sound source identification performance of the MCB method. In addition, Subspace Variational Bayesian (SVB) method is integrated into the above three rotating sound source localization methods, and the resolution of the localization results is significantly improved, and the performance of high-resolution localization methods based on MCB is further discussed. According to the location results of MCB-SVB, the blade noise mechanism of the five-blade axial fan used in the experiment is analyzed, and the suggestion of noise reduction is given. This study not only discusses how the MCB-SVB method is used to reduce fan blade noise, but it also illustrates the potential for using this method to identify axial fan faults.