Dynamic UAV-Assisted Power Grid Restoration With Rolling-Horizon Crew Dispatch for Post-Disaster Resilience

The increasing frequency and intensity of natural disasters, such as typhoons and hurricanes, pose severe threats to the resilience of power distribution systems, leading to prolonged outages, economic losses, and disruption of critical services. This paper presents a novel unmanned aerial vehicle (UAV)-assisted restoration framework that integrates probabilistic failure analysis, dynamic optimization, and real-time dispatch strategies to enhance post-disaster recovery. The proposed methodology minimizes the cost of interruption (COI) by simultaneously optimizing UAV-based inspections and repair crew (RC) scheduling. The UAV inspection task is formulated as a modified Traveling Salesman Problem (TSP) with adaptive prioritization of critical nodes using a dependency matrix. Additionally, a rolling-horizon optimization strategy dynamically coordinates UAV inspections with real-time RC dispatch, ensuring parallel execution of fault detection and restoration. Both CPLEX and genetic algorithm (GA) were employed to solve the problem, where GA enabled solving the full unsegmented system with reduced computational time. Simulation results on a 33-bus test system demonstrate that the proposed framework significantly outperforms conventional restoration methods, achieving up to 56% reduction in COI and up to 32% reduction in system restoration time by prioritizing disaster-affected nodes, reducing inspection delays, and improving system resilience. The integration of UAV-based inspections, battery-aware path optimization, and adaptive RC scheduling ensures a scalable and efficient power system restoration strategy for large-scale disaster scenarios.