A HYBRID METHOD FOR FAULT DETECTION AND CLASSIFICATION IN ELECTRICAL NETWORKS: COMBINING DISCRETE WAVELET TRANSFORM AND ARTIFICIAL NEURAL NETWORK

In modern electrical power systems, the prompt detection and accurate classification of faults are critical for maintaining operational reliability and safety. This study introduces a hybrid fault detection methodology that combines Discrete Wavelet Transform (DWT) with Artificial Neural Networks (ANN) to enhance the performance of fault diagnosis in electrical networks. Traditional fault detection techniques, such as Fourier transforms, often fall short in analyzing non-stationary and transient signals, leading to ambiguous results and delayed responses. The proposed framework utilizes DWT for effective feature extraction from electrical signals, decomposing them into various frequency components that capture essential characteristics of both normal and fault conditions. Subsequently, an ANN is trained using these extracted features to classify fault types, including phase-to-ground, line-to-line, and three-phase faults. Simulations conducted on the IEEE 9 Bus System validate the effectiveness of the hybrid approach, demonstrating significant improvements in detection accuracy and response time compared to conventional methods. The integration of DWT and ANN not only improves fault identification rates but also provides precise diagnostic capabilities, allowing operators to make informed decisions swiftly. The findings confirm that this hybrid model is a robust solution for modern power systems, particularly as they become more complex with the integration of renewable energy sources. This methodology represents a significant advancement in fault detection techniques, offering a reliable framework for enhancing the stability and safety of electrical distribution networks.