Modern Approach for Diagnosing and Detecting Faults on Overhead Transmission Lines using Artificial Neural Networks

The high voltage transmission over long distance is mainly focused on safety and economic. If necessary, to transmit certain amount of power through long area should care about power regulation, efficiency and losses. Any time fault may occur if any deviation in voltage and current reaches above normal range. The faults in power system causes over current, under voltage, unbalance of the phases, reversed power and high voltage surges. Three-phase symmetrical faults are known to be the most severe in a power system due to large fault currents. However, single phase, phase to ground faults are more common faults that occur. If not checked in due time, these faults may grow to symmetrical fault which is uncommon but most severe. Trial and error method is usually practices for detecting the fault location on transmission line. In this method, the supply feeds at the single end at a time by dividing that transmission line into two parts and detect the fault up to that limit. This paper provides a modern approach for detecting a fault and diagnosing overhead transmission lines through the implementation of DWT and ANN controllers. Voltage signals are found from the sending end for each phase, the decomposition using DWT to obtain a detail coefficient of up to 2 stages.