ANN-based fault classification and localization with optimized PMU deployment for transmission systems

Accurate and timely Fault Detection (FD) remains a fundamental challenge in transmission systems due to the dynamic operating conditions of modern power grids, the diversity of fault types, and fast reclosure events. Conventional fault diagnosis techniques are often constrained by assumptions of static load profiles, dependence on high-rate sampling, or extensive data requirements, which hinder their real-time deployment. To address these limitations, this study proposes a novel Artificial Neural Network (ANN)-based framework for Fault Classification (FC) and localization using only bus voltage measurements. The distinctive contribution lies in the use of physically meaningful features rather than purely statistical ones, thereby ensuring robustness and interpretability. Simulation studies conducted on the IEEE 14-bus system show that the suggested method achieves FC accuracy above 98% and localizes faults with an error margin of less than 2% of line length, outperforming existing data-driven techniques. Furthermore, a new Phasor Measurement Unit (PMU) placement strategy is introduced, which improves observability while reducing hardware requirements, enabling reliable performance even under severely compromised measurement conditions. Achieving 0.89783 accuracy and notable reductions, ANN showed up to 5.66% higher accuracy and over 30% lower error rates, confirming its superior FC capability. These findings underline the potential of the suggested methodology as a scalable and resilient solution for real-time fault management in modern transmission networks.