High Impedance Fault Detection Using Harmonics Energy Decision Tree Algorithm

In this paper a new pattern recognition based algorithm is presented to detect high impedance fault in distribution networks. In this method the total energy of odd, even and in-between harmonics up to 400 Hz is calculated and is fed to a decision tree as a classifier. The proposed scheme can successfully distinguish the HIFs from normal operations in power system such as harmonic load switching, capacitors switching, and transformer energization. The results show high accuracy of the proposed method in the detection task.