Motor current signature analysis via four-channel FIR filter banks

Abstract

Motor current signature analysis (MCSA) is capable of providing continuous monitoring of induction motors in a non-intrusive manner. Fourier based techniques have been used widely in processing of stator current but these techniques have a shortcoming in processing non-stationary signals such as the stator current. Recently, wavelet packet decomposition (WPD) has become popular in such applications since it gives better results in the case of non-stationary signals. The latter approach has much higher computational complexity limiting its use in motor diagnostics applications. In this study, the use of four-channel FIR filter banks is proposed to provide lower computational complexity. Four-channel filter banks employ higher level of parallel processing than currently used two-channel filter banks. FPGA implementation of the proposed algorithm would result in even further reduction in overall computation time.