Soft Computing Technique for Automatic Detection, Classification and Location of Transmission Line Faults

Proper monitoring and control of power system includes detection of Transmission lines faults. Faults in transmission line is the main factor in power interruption. To ensure a reliable, high quality, uninterrupted power supply to consumers, a smart, efficient, accurate system that can detect, classify, and locate faults is important. Artificial Neural Network (ANN) is one of the best choice for it. However, acquiring real fault data is difficult due to high data processing needs, security issues etc. Therefore, in this paper a transmission line is modelled using MATLAB/Simulink and dataset to train the neural network is generated by simulating the Simulink model for various fault types and different fault location. Three phase RMS voltages and currents measured on the sending end bus is input to the neural network to train it. Later the trained ANN model is tested by applying random unknown set of data. The simulation results exhibit good performance, ANN can effectively detect, classify and locate the faults.