Dynamic Network Clustering for Distribution System Resilience Against Typhoon Events

Abstract

Natural disasters intensified by climate change stimulated the recent efforts to develop power system resilience both in the structural and in the operational sense. This study proffers an operational resilience framework that uses distributed generation (DG) and switches in order to pick up critical loads through dynamically forming DG-powered clusters in a radial distribution system damaged by a typhoon onslaught. As an innovation to consider a typhoon event’s spatiotemporal conditions in cluster formation, we use wind-based line fragility curves to simulate line damages, and introduce a load priority update scheme based on government-mandated flood hazard levels. A mixed-integer optimization problem is also formulated to maximize the priority-weighted load pickup during clustering, without violating power system operation requirements. MATLAB-based implementations use the IEEE 37-node radial distribution test system, modified to include DG units and geographically mapped lines and nodes. Numerical results suggest effectiveness of the proposed scheme in terms of dynamically forming clusters optimized according to load pickup priorities and system operation requirements. Program runtime is also found out to be affected by the system size and the load flow convergence settings.