Multi-microgrids

highly autonomous and flexibly interconnected

offer a significant approach to enhancing power system resilience. However

they still face challenges from the complex impacts of extreme weather and the need to guarantee the charging demands of electric vehicle (EV) users. Therefore

this paper proposes a resilience enhancement strategy for multi-microgrids with EVs under extreme weather conditions. First

a resilience assessment framework for multi-microgrids is constructed

which includes a decision model for EV charging and discharging patterns. By incorporating the participation factor of EVs

the decision model optimizes the vehicle-to-grid process to guarantee users' charging demands. Second

extreme weather models

including heavy precipitation

extreme high or low temperature

and strong wind

are established to quantify their impact on the resilience of multi-microgrids

which specifically characterize the probability of microgrid disconnection and internal generator shutdowns. Subsequently

to efficiently coordinate power transmission between multi-microgrids and internal resource management within each sub-microgrid

a hierarchical resilience enhancement strategy based on deep reinforcement learning is designed

which incorporates the hindsight experience replay mechanism

effectively improving the convergence of the algorithm in complex tasks to enhance system resilience. Finally

simulation results demonstrate that the strategy significantly enhances the resilience of multi-microgrids

reduces operation costs and guarantees EV users' demands.