Cooperative self-healing algorithm for ground faults in hybrid overhead-cable active distribution networks

This paper proposes a novel cooperative self-healing algorithm based on multimodal parameter inversion to enhance fault localization accuracy and recovery efficiency in hybrid overhead-cable active distribution networks. The method builds a multimodal equivalence model that integrates transient signals, fault features, and grid topology, accounting for line heterogeneity and dynamic distributed generation. High-reliability feature extraction is achieved via multi-scale time-frequency analysis and adaptive filtering. A deep learning-based inversion algorithm, constrained by physical models, accurately identifies fault locations and key parameters. An intelligent diagnosis system fuses multi-source data for rapid fault recognition, while a differentiated self-healing strategy generates optimal isolation and reclosing schemes tailored to fault types and coordinates with distribution automation. Compared to traditional unimodal, threshold-based methods, the proposed approach offers superior performance in complex scenarios and provides a promising solution for intelligent operation and maintenance.