A hybrid method for high impedance fault classification and detection

High impedance faults (HIFs) have over the years brought a complex challenge for protection engineers. This complexity is founded of the fact tha a HIF poses characteristics which appear to be difficult for conventional protection schemes to detect their presence in a power system. In this work, we propose a method which makes an attempt to diagnose HIFs effectively. The method uses a feature extraction, classification and regression schemes by applying packet wavelet transform (PWT), support vector machine (SVM) and support vector regression (SVR) respectively. The effectiveness of the proposed method was tested using MATLAB. Furthermore, a practical setup was conducted to test the viability of the proposed method. The results showed good classification accuracy and minimum error of estimation.