Fault Detection and Protection Schemes In Electrical Transmission Networks

Abstract

Reliable transmission-line protection requires fast detection of abnormal conditions, correct identification of the faulted phases, dependable discrimination of internal and external faults, and selective isolation of the minimum network section. This methodology paper develops a hybrid protection framework for high-voltage AC transmission networks by combining conventional impedance-based protection with discrete wavelet transform (DWT) transient features and an artificial neural network (ANN) classifier. A 400 kV, 50 Hz, 150 km two-terminal transmission line is modeled with distributed parameters. Single-line-to-ground, line-to-line, double-line-to-ground and three-phase faults are applied over a wide range of locations, inception angles and fault resistances. Local three-phase voltages and currents are sampled, decomposed using a Daubechies-4 wavelet, and converted into normalized transient-energy and sequence-component features. An adaptive threshold initiates fault detection, the ANN identifies the fault class, and a directional distance element validates zone and direction before issuing a trip. The resulting scheme is designed to retain the security of distance relaying while improving sensitivity to transient signatures. Representative simulation outcomes show that the hybrid logic can achieve sub-cycle detection, high classification accuracy and low fault-location error under the stated test conditions. The study provides a reproducible step-by-step methodology suitable for MATLAB/Simulink or equivalent electromagnetic-transient simulation environments.