Fault Detection Method Using ANN for Power Transmission Line

• Published in: 2020 10th IEEE International Conference on Control System, Computing and Engineering (ICCSCE) pp. 79 - 84
• Main Authors: Leh, Nor Adni Mat, Zain, Fatihah Mohd, Muhammad, Zuraida, Hamid, Shabinar Abd, Rosli, Anis Diyana
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-08-2020
• Subjects: ANN; Artificial neural networks; Backpropagation algorithm; Circuit faults; Fault detection; Fault detection and classification; Feed-forward ANN; Impedance; Neurons; Power transmission lines; Training; Transmission line
• DOI: 10.1109/ICCSCE50387.2020.9204921

Electric power transmission line is an important medium to supply the electric power continuously. However, the exposed environment of transmission line leads to the presence of faults. Therefore, this paper presented a method to overcome the problem in order to improve the quality of power system. The objective was to provide an accurate method for fault detection and to classify the fault that occurs in power transmission line using Artificial Neural Network (ANN). 14-bus system has been tested through MATLAB programming to create the fault using impedance technique. After carrying out the fault simulation, the result of faulty voltage and current was taken as an input to the neural network. The study implemented feed-forward ANN with backpropagation algorithm to develop the fault detection and classification. The performance of the detection and classification is measured using the Mean Square Error (MSE), linear regression and confusion matrix. The detection achieved a tolerable MSE of 5.5614e-8, a correlation of 1 and an accuracy of 100% proving the satisfactory performance of the system. While the classification achieved a tolerable MSE of 0.43699, a correlation of 0.83955 and an accuracy of 70% showing that the performance of the system is acceptable.