Transmission Line Fault Classification and Location Using Multi-Layer Perceptron Artificial Neural Network

Most transmission lines are overhead, spanning possibly thousands of kilometers and are exposed to different climatic conditions. and therefore prone to faults. This paper develops accurate, fast and reliable algorithms that can classify and locate faults on transmission lines. These reduce outage time. Multi-Layer Perceptron Artificial Neural Networks (MLP-ANN) are used to implement the algorithms because of their ability to learn during a wide range of conditions. The ANN inputs are the voltage and current signals. These were measured at one terminal of the line. The test system is the 29-bus Great Britain transmission network. The algorithm sensitivity is investigated with varying fault impedance and inception angle, and different locations on the transmission line.