A Novel Approach Based on CEEMDAN to Select the Faulty Feeder in Neutral Resonant Grounded Distribution Systems

Abstract

In this article, a novel approach based on the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is proposed to solve the problems of adaptive extraction of fault signals’ characteristics and the low sensitivity of faulty feeder selection methods. The key idea of this approach is to decompose the feeder’s zero-sequence transient current by CEEMDAN into several modes of different frequency scales. The energy of each mode is calculated in order to obtain the fault characteristic mode. Then, the sum of characteristic mode components and its polarity are calculated. If the polarities of all feeders are the same, then the fault occurs on the bus. On the other hand, if the feeders’ polarities are different, the sum of the characteristic mode components during one cycle is compared, and then, the fault occurs on the largest one. The proposed approach has been applied on different feeder structures, different fault types, and various perturbations for validation. Moreover, the outcomes of the proposed approach have been compared with the traditional EMD algorithm and wavelet transform (WT) method for further verification of its adaptation and sensitivity. The simulation results showed that the proposed approach can adaptively extract the fault characteristics and precisely distinguish the faulty and nonfaulty feeders. Furthermore, the simulation results assured that this approach is adaptable as it is not affected by perturbations, way of network operation, or feeder structure.