DT-CWT based event feature extraction for high impedance faults detection in distribution system

In this paper an algorithm for high impedance fault detection is presented. This algorithm uses dual tree complex wavelet transform to extract the features of disturbance signals according to the post- and pre-disturbance data windows. There are also a frequency tracking unit and a disturbance detection unit in this algorithm for enhancing the resolution of features. A trained probabilistic neural network is used to discriminate between the fault and other events. EMTP-RV has been used for simulation of various events with different conditions for training and testing the algorithm. As this algorithm uses the features extracted from the events, the fault detection can be done with more reliability. Results of implementing the algorithm for high impedance fault detection in a distribution test feeder show a high level of dependability and security.