Effect of current resampling in Motor Current Signature Analysis

Abstract

Motor Current Signature Analysis (MCSA) is one of the most widely used methods in monitoring condition of induction motors. Traditionally, the stator current is preprocessed by notch filters to suppress line fundamental frequency. Then, the fast Fourier transform is utilized for the spectral analysis of the preprocessed stator current in most applications. However, this approach has a shortcoming in the analysis of non-stationary signals such as stator current under varying load conditions. The use of wavelet transform is suggested for providing better analysis results in recently published studies. Both approaches use some preprocessing of stator current in the analysis that is very sensitive to even slightest variations in sampling frequency. The resampling of current data at an exact integer multiple of line frequency is proposed in this study to improve the overall fault detection performance.