High Impedance Fault Detection using Wavelet Transform and Artificial Neural Network
The detection of High impedance faults (HIFs) on distribution system is most difficult problem faced by electric utility system. These faults remain undetectable. Therefore maintenance personnel will not identify fault until a hazard is reported. When one of the phases of the transmission line makes electrical contact with a semi-insulated object like a tree, pole, road surface, gravel, concrete, dry land, etc., the fault path has a very high resistance, which is known as a fault with high impedance (HIF). The fault current values ranges from 0 to 75 amperes and cause arcing and flashing at the point of contact, poses the greatest risk of public electrical shock or fire for HIF. As a result, the public and reliable operation view the detection as more significant. An empirical solution to fault detection using Discrete Wavelet Transform (DWT) and Neural Network is presented here. This is achieved by training the Artificial Neural Network using the features (standard deviation values) extracted from the fault current signal by DWT technique for different conditions of fault with different fault resistance values in the system.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.