Research on Fault Line Detecting Based on Empirical Mode Decomposition (EMD) in Resonant Grounded Systems

Abstract

In resonant grounded power distribution systems, it is difficult to identify the fault line due to weak current and unstable fault arc after single phase-to-earth fault occurring. On the base of analyzing characters of the steady and the transient signal of zero-sequence-current, a new method of fault-line selection by using empirical mode decomposition (EMD) is proposed. Decompose the transient signal of each line into a finite and often small number intrinsic mode decomposition (IMF). Main eigenvalues are contained in each IMF component. And the effective component contents within an IMF is different from others. With Hilbert transform we can calculate instantaneous frequency. Based on the principle that transient information of fault-line contains that of all healthy lines and the mean value of instantaneous frequency in fault-line is the maximum we can detect and select fault circuit. Lots of digital simulation results show that the proposed method is effective and reliable.