Answers are grounded in this paper and cite their sources.
Experimental evaluation demonstrates real-time adaptive fault protection in deeply nested power grids, indicating enhanced grid resilience and reduced operational costs.
Existing protection schemes for drop-out fuses, primarily designed for radial or shallow mesh networks, struggle with the dynamic coordination and real-time adaptation required in deeply nested grids, particularly under complex electromagnetic operating environments with high distributed generation penetration. This paper presents a real-time adaptive operation and maintenance system that integrates a Dynamic Coordination Algorithm (DCA) and a Lightweight Temporal Convolutional Network (LTCN) to address this challenge. The DCA formulates multi-timescale protection thresholds as a mixed-integer programming problem to resolve coordination conflicts in deep topologies, while the LTCN enhances high-frequency transient signal recognition through an optimized cascade of convolutional feature extraction and gated recurrent units. Experimental results demonstrate that the proposed system achieves an average fault trend prediction accuracy of 0.822 (F1: 0.774, AUC: 0.876), reduces the average fault clearance time to 89.3 ms, and limits misoperation incidents to only 2 out of 100 test faults. Furthermore, the operational cost is minimized to 38.2 yuan per scheduling event, with scheduling coverage and fault tolerance rates reaching 96.8% and 89.7%, respectively. Stable edge communication is maintained at 74.24–78.53 Mbps with timing errors of 1.46–2.4 ms and a packet loss rate below 1.11%. The proposed framework improves the reliability of intelligent protection and maintenance for complex distribution systems while providing effective support for electromagnetic infrastructure monitoring and resilient power network operation.
No takes yet. Share an insight, caveat, or question.
Bao et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: