Mode identification of fan tonal noise in cylindrical duct based on Bayesian compressive sensing

Accurate identification of in-duct acoustic modes is crucial for comprehending fan noise generation mechanisms, propagation characteristics, and control strategies. The number of fan noise modes increases with frequency, and using a sufficient number of microphones for higher-order mode identification becomes challenging due to cost considerations. A Bayesian compressive sensing method for mode identification is proposed in this paper to address the issue of insufficient microphones. The sensing matrix is constructed by randomizing the arrangement of a small number of microphones. The acoustic field is characterized using a Bayesian probabilistic model, and the inverse problem is formulated as the estimation of mode coefficients within the Bayesian compressed sensing framework. The proposed method accurately identifies acoustic modes in the presence of adaptive parameters and achieves more precise magnitude recovery than previous methods. The effectiveness and robustness of the proposed method under various parameters is demonstrated by comparing the simulation results of different mode identification methods. The effectiveness of proposed method is further substantiated by experimental validation using a 1.5-stage axial flow compressor, demonstrating accurate identification of target modes with fewer microphones.