Wavelet-Fourier Analysis of Audible Signals to Characterize the Vibrations in LSRA

Abstract

This paper proposes the analysis tools, the methodology, and the experimental setup to support the characterization of vibrations in linear switched reluctance actuators by analyzing the audio signals emitted. The analysis is performed by using two approaches: the windowed Fourier transform and the joint wavelet-Fourier which leads to a hybrid spectrum representation. The spectrum of the audible noise from the vibrations is characterized for different configurations of the actuator, and the spectral components are presented, analyzed and discussed. To evaluate the responses for these configurations, one uses either its frequency components, the spectral cross-correlation, and the Cross-Power Spectrum Density through the Welch periodogram. The reliability and the suitability of this approach are verified by comparing the experimental data with the outcomes from finite elements simulations. The proposed method enables the experimental identification of the critical vibration frequencies in LSRA.