Fault feeder detection method utilized steady state and transient components based on FFT backstepping in distribution networks

Abstract

A novel cut-in point for fault detection issue is introduced by dividing the transient zero sequence current (TZSC) into the steady-state components (SSCs) and transient components (TCs), a comprehensive detection method based on the coordinated SSCs and TCs is presented. For the SSCs criterion, the TZSCs within the second power cycle are used as the research objects, the local SSCs are obtained based on Fast Fourier transformation (FFT), if there is unique phase difference of the local SSCs is larger than Π/4, the correlation coefficients of SSCs are used to detect the fault feeder; Otherwise goes into the TCs criterion stage, the global SSCs based on the backstepping method is calculated, accordingly, the TCs can be obtained by subtracting SSCs, the difference between fault feeder and healthy feeders are enhanced by using the extremum of TCs and bubble sort method, lastly, the cross-correlation distance (CCD) is introduced to detect the fault feeder. The proposed method not only has high detection accuracy but also needs least time length, the correctness and extensive applicability have been verified by the results from the radial distribution network simulation model, arc fault model, improved IEEE-13 node system and field test.