Model Based Extended Wavelet Analysis and its application to fault diagnosis of an elevator system

Abstract A new wavelet analysis method is proposed, that is a combination of the wavelet transform, a nonlinear transformation and a correlation analysis. In this paper, it is called ‘Extended Wavelet Analysis (EWA)’ procedure. Furthermore, the ‘Model Based EWA’ method is proposed in which the nonlinear transformation is based on a model of the diagnosed object. It can analyze a short transient signal and it is effective for the correlation analysis between the wavelet spectrum and arbitrary variables in the object. It was applied to the diagnosis of an elevator system. Some simulation results and experimental results reveal the effectiveness of the EWA procedure.