A new FDOST entropy based intelligent digital relaying for detection, classification and localization of faults on the hybrid transmission line

The paper presents a new digital relaying for detection, classification and localization of faults on the hybrid transmission line consisting of an overhead line and an underground cable. The entropy principle together with fast discrete orthogonal S-transform (FDOST) represented by window dependent bases is utilized for feature extraction and the support vector machine (SVM) classifier model & support vector regression (SVR) model are employed for pattern recognitions to predict the types and locations of faults. After modelling and simulation of the transmission system in Electromagnetic Transient Program (EMTP) software, three phase fault current signals are recorded at one end of the line to extract entropy of FDOST coefficients from each of the three current signals of half cycle duration after fault initiation. The proposed relaying technique is tested on a single-junction and a multi-junction hybrid transmission lines under different fault conditions and is found fast and accurate independent of fault type, fault section, fault resistance, fault inception angle (FIA) and load angle. Another important aspect of the method is that it needs no prior identification of the faulty section for the estimation of fault location. The immunity of the proposed method to noise is also established by testing it with fault current signals impregnated with white Gaussian noise of level 30 dB signal to noise ratio (SNR).