Novel filter based ANN approach for short-circuit faults detection, classification and location in power transmission lines

This paper presents a novel approach for detecting, classifying and locating short-circuit faults in power transmission lines. Based on the proposed approach, a hybrid framework consisting of a proposed two-stage finite impulse response (FIR) filter, four support vector machines (SVMs), and eleven support vector regressions (SVRs) is implemented in Proteus 6/MATLAB environments. The proposed two-stage FIR filter together with the SVMs are used to detect and classify short-circuit faults while the SVRs are utilized to locate short-circuit faults. The implemented framework needs few training samples for training the SVMs and SVRs. As will be shown, for a power transmission line with the length of 50 km, only 6 training samples are needed to train each SVR. The trained hybrid framework carries out the processes of fault detection, classification and location only during 1 cycle which is strictly shorter than the faults clearing time. It means that the proposed hybrid framework can rapidly detect, classify and locate short-circuit faults in power transmission lines before power outage carried out by protection relays. An actual three-phase 230 kV, 50 Hz power transmission line with the length of 50 km is simulated to validate the theoretical results and to verify the proposed technique accuracy.