Strengthening Resilience: An Agent-Based Model for Poststorm Restoration and Grid Hardening in Power Distribution Systems

Storm-related power outages cause substantial economic losses each year, affecting millions of customers for days or even weeks. To support effective preparation for extreme weather events, utility operators require reliable predictions of system performance metrics, such as the number of affected customers and restoration time. This study develops a unified agent-based power system restoration framework that explicitly links storm hazard modeling, component-level fragility, and postevent restoration processes within a single simulation environment. By integrating physics-informed fragility modeling with agent-based restoration processes, the proposed framework provides a unified agent-based modeling that links damage and recovery, while also enabling the joint evaluation of crew allocation decisions and infrastructure hardening strategies and their combined effects on system recovery. The framework is demonstrated using a case study of the Connecticut power distribution system under Tropical Storm Isaias (2020) and is validated using observed outage and restoration data, achieving less than 13% error in predicted restoration time. Further, the findings quantify the potential benefits of deploying more crews, improving crew efficiency, and upgrading the physical infrastructure by using higher-class poles or replacing older poles.