Real-Time Electrical Fault Classification in Transmission Power Lines Based on Advanced Signal Processing Techniques

This project addresses the critical challenge of rapid fault location and classification in transmission systems to reduce power outage durations. Common faults, such as short circuits (single-phase, two-phase, three-phase), often lead to extended restoration times due to the lack of fault locators in existing infrastructure. The proposed solution focuses on processing the altered signals during a fault to apply Artificial Intelligence techniques, specifically Machine Learning (ML). A data-driven approach is employed, training models with real or simulated fault patterns (L-G, L-L, L-L-G). Specifically, the supervised k-Nearest Neighbor (KNN) algorithm is implemented to identify and classify the fault type with high accuracy, demonstrating low sensitivity to variations in parameters like fault magnitude and resistance. The methodology includes collecting and simulating data from transmission circuits in MATLAB®, creating robust databases, and conducting comparative tests to evaluate the accuracy and performance of the proposed methods under various operating conditions and fault resistance variations.<br/><br/><b>Goal</b>: <br/>The main objective is to classify real-time electrical faults on transmission lines using advanced signal processing techniques and machine learning algorithms.<br/><br/><b>Research lines</b>: <br/>Optimization in electrical systems