Research on Fault Diagnosis and Self-Healing Control Model in Power Supply and Distribution Network

This research work innovatively constructed a power supply and distribution network fault diagnosis and self-healing control system with ensemble learning as the core technology. Its core goal is to significantly improve the accuracy of fault identification and the system’s ability to respond immediately to emergencies, while ensuring that an efficient self-recovery mechanism can be realized. By cleverly integrating multivariate machine learning algorithms including random forests and gradient boosted decision trees (GBDT), the constructed model shows excellent performance, can accurately identify various fault modes, and significantly enhances the robustness and environmental adaptability of the system in the context of complex and changeable power supply and distribution networks. Once an abnormality is detected in the system, this model can not only accurately locate the source of the fault, but also rely on the optimization algorithm to intelligently select the optimal recovery path, thereby achieving rapid restoration of power supply. During the research process, people deeply optimized the feature extraction strategy and data integration processing scheme to further improve the accuracy of fault diagnosis and the reliability of system operation. Experimental simulation results show that under a series of fault scenario simulations, the diagnostic accuracy of this model reached an impressive $96.3 \%$. At the same time, the self-healing control response time was shortened by ${1 5. 2 \%}$ compared with the traditional method. This significant improvement directly contributed to a leap in the overall operating efficiency of the power supply network.