High-resolution localization of rotating acoustic sources: An experimental investigation and axial fan application

Ning Chu,Qin Liu,Liang Yu,Qian Huang,Yue Ning,Dazhuan Wu,Ali Mohammad-Djafari

DOI: https://doi.org/10.1016/j.measurement.2022.111149

IF: 5.6

2022-01-01

Measurement

Abstract:This paper presents an experimental investigation of rotating source localization to identify the axial fan noise with high spatial resolution methods. It exploits the on time-domain de-Doppler, and sparsity-based inversion and the convolution model to approximate the power propagation model of the rotating source. This work is also an extension of the previous paper of authors. But the focus of this study is the experimental investigation and application of actual data. To evaluate the performance of the proposed algorithm, a vortex generator is attached to the fan blade, creating discontinuities in the blade to simulate axial fan blade defects (cracks, holes, dust accumulations, etc.). The experimental data are shared as a data benchmark to promote further studies in the field of rotating acoustic localization. Compared with the state-of-the-art rotating Beamforming method (ROSI), the acoustic imaging results at various working frequencies for three different axial fans with different rotating speeds (RPM) are accurately identified and sparsely distributed. In addition, the power propagation model derived in the previous paper can be interpreted by a convolutional approximation model in this paper, and this technique can help to reduce the computational costs of proposed methods for real applications. The accuracy of the aerodynamic noise positions estimated by the proposed method are further proved to match well with the aerodynamic mechanism.