REVOLUTIONIZING FAULT DETECTION IN HIGH-VOLTAGE TRANSMISSION LINES THROUGH ANN MODELS
Keywords:
Artificial Neural Network, Symmetric Fault, Unsymmetric Fault, Transmission Line, Multilayer Perceptron, Neural Network Pattern Recognition Tool (NPR TOOL)Abstract
Transmission lines, critical to ensuring a constant flow of electricity, are susceptible to various problems because they are on the surface. Faults occurring in the system after relays are used to determine the type of the fault and its location. However, failures of the relays occasionally happen, and these failures can significantly impact the operation of the power system, resulting in noticeably delayed fault recovery. Advanced, efficient, and sensitive fault detection systems are necessary to overcome this issue. This study aims to propose an intelligent and automated fault detection method for a 117 km, 500 kV power system. The transmission line is simulated using MATLAB and Simulink to generate an extensive dataset. In this sense, the dataset is used to train, validate, and test an Artificial Neural Network (ANN) designed for fault detection based solely on instantaneous voltage and current measurements, as well as a multilayer ANN trained using a backpropagation algorithm. The Mean Square Error (MSE) and confusion matrix were utilized to evaluate the system's performance, achieving an MSE of 2.2498e-9 and 100% accuracy, thereby showcasing the effectiveness of the proposed NN-based algorithm in a practical application, such as transmission lines.
