Intelligent Detection of High Impedance Fault using Extreme Learning Machine

Electrical energy is one of the most promising forms of energy. The power system needs to be in a healthy condition to look after the continuous supply of electrical energy. High Impedance Faults (HIFs) couldn't be detected through conventional protection schemes present in the power systems. In this paper, Discrete Wavelet Transform (DWT) and machine learning techniques are utilized for efficient detection of HIF. The feature extraction is performed using DWT. Ranking of the features is done using t-test class separability criterion. Classification is performed using one of the randomization networks called Extreme Learning Machine (ELM) with four most significant features. Multi-layer perceptron (MLP) and Support Vector Machine (SVM) is also used for comparison. Current signal data for HIF and Non-HIF events are acquired from a standard IEEE 13 bus distribution system. The simulation result shows that the selected method gives good performance accuracy.