Fault classification method based on artificial neural networks in power transmission lines

This thesis aims to develop an Artificial Neural Network (ANN)-based model for fault detection in energy transmission lines. Within the scope of the study, four different ANN models were designed, each trained using various training/testing ratios. Diversifying these ratios allows for a more comprehensive evaluation of each model's accuracy and generalization capability.The models were trained using an open-access dataset, and the ANN was employed to accurately classify different types of faults. During training, the network weights were optimized using the backpropagation algorithm, and fault type predictions were performed using the tanh and softmax activation functions.Model performance was evaluated using classification metrics such as accuracy, precision, recall, and F1 score, and results were visualized through confusion matrices. The best-performing model, Model 4, achieved an accuracy rate of 99,46%.The results demonstrate that the ANN-based approach significantly accelerates fault detection and offers higher accuracy compared to traditional methods. This study presents an innovative and effective solution for the early and accurate detection of faults in energy transmission lines.