Extract Fault Signal via DWT and Penetration of SVM for Fault Classification at Power System Transmission

Abstract

Power transmission lines are extremely important for the power system to deliver energy of electricity from the plant to the load. The short circuit of fault often occurs in the transmission line and may lead to disconnecting the power supply to the load. This study implements a hybrid technique that is Discrete Wavelet Transformation (DWT) and Support Vector Machine (SVM) for classification of fault in the transmission line. The DWT was created to extract the detailed signal of transient D8 and D9 (order of 4) at 50 kHz for sampling frequency. The value of Root Mean Square (RMS) will be determined by the coefficients D8 and D9 for training and test data using SVM technique. Furthermore, SVM is utilized to detect the fault for each phase and the ground is discovered in the type of fault. The SVM technique has been run using parameter C and kernel scale to achieve the great results of classification of the fault. Type of simulating fault has a variation of location of the fault, fault of resistance and initial angle. The training and test data run for the Test System of Riau, Indonesia. The test result for the classification of fault reaches the highest accuracy of 100%.