<![CDATA[This article presents a wavelet analysis-singular value decomposition (WA-SVD) based method for precise fault localization in recent power distribution networks using k-NN Classifier. The WA-SVD leverages the slime mould algorithm (SMA) and graph theory (GT) in enhancing the overall accuracy of fault localization. To validate the proposed methodology, extensive tests are conducted on various benchmark systems, including the IEEE 33-bus radial distribution system, the IEEE 33-bus meshed loop unbalanced distribution system, the IEEE 33-bus system with integrated renewable energy sources, and the IEEE 13-bus feeder test system. The results demonstrate a high fault classification accuracy of 99.08%, with an average localization error of just 1.2% of the total line length. The k-NN classifier exhibited a precision of 98.2% and a recall of 99.2%, underscoring the reliability and sensitivity of the proposed method. Additionally, the computational efficiency of the algorithm is evidenced by an average processing time of 0.0764 seconds per fault event, making it well-suited for real-time applications.]]>
Copyrights © 2025
