Inrush and fault current discrimination using wavelet transform and autoregressive modeling

Abstract

Autoregressive modeling, due to its forecasting capability and Wavelet Transform with ability of non-stationary signal analysis, both seem to be proper tools to analyze the systems' transients. Inrush current as a transient, single-handedly is potential to cause problems. Further, wrong discrimination of internal fault and inrush currents may cause wrong operation of protection devices. Differences in protective reactions necessary for internal faults and inrush currents introduce us to the importance of discrimination. This study attempts to scrutinize the performance of Wavelet Transform and AR modeling in the distinction investigation. To accomplish this goal, Fault and Inrush currents spectra have been computed and estimated by AR modeling and WT is applied to the same system. Hence, Fault current and Inrush current spectra and the pertinent coefficients have been computed and estimated by AR modeling and the Multi-Resolution Analysis (MRA) is done for the same system Observing differences between output spectra of AR modeling for inrush and fault cases, have provided an approach for discrimination of these two that makes the decision process trustee, consequently.