Fault Localization on the Transmission Line Using FDOST and RBFNN

This paper presents a fault localization technique based on fast discrete orthogonal Stockwell transform (FDOST) and radial basis function neural network (RBFNN) on the transmission line. A part of the transmission network of WBSETCL, West Bengal is designed and simulated in MATLAB Simulink for the fault investigation. The fault current signals are recorded at one end of the transmission line with a sampling frequency of 50 kHz, and FDOST energy is extracted as fault feature from each of the three fault current signals. These features are fed to the RBFNN for fault localization on the transmission line. The proposed algorithm is found accurate for different types of faults, fault resistances and fault inception angles (FIA) at different locations on the transmission line.