Faulty Feeder Detection Method for Single-Phase Ground Fault in Distribution Networks Based on TVFEMD and Cross Sample Entropy

To solve the problems of weak characteristics of single-phase ground faults (SPGF) in the ungrounded system of the distribution network and difficulties in feeder detection, a faulty feeder detection method based on time-variant filtered empirical mode decomposition (TVFEMD) and cross sample entropy (CSE) is proposed. Firstly, the zero-sequence equivalent network of SPGF is established and the characteristics of reactive power transient of each feeder are analyzed. Secondly, TVFEMD is used to extract the transient components of zero-sequence current (ZSC) of each feeder and bus zero-sequence voltage (ZSV) derivative. Then, CSE is used to calculate the characteristic frequency band (FB) of IMF components and the transient reactive power (TRP) within the characteristic FB is obtained. Finally, based on the directional difference between the TRP of the faulty feeder and that of healthy feeders, the correlation of the TRP of each feeder is detected to determine the faulty feeder. The results show that the proposed method will not be affected by factors such as fault phase angles, locations and noise interferences of the fault, and has strong adaptability to high resistance.