Artificial intelligence for resilient power system:Motivations,advances,and challengesOA

The increasing frequency of extreme events,including natural disasters and man-made attacks,poses significant threats to power systems,highlighting the urgent need for resilience research.This research is generally divided into three phases:pre-disaster,in-disaster,and post-disaster.The pre-disaster phase centers on impact prediction,enabling risk warnings and proactive resource deployment.During the in-disaster phase,the emphasis shifts to real-time event monitoring and operational adjustments to minimize adverse effects.The post-disaster phase is dedicated to coordinating emergency resources to rapidly restore power loads and infrastructure.Enhancing power system resilience necessitates the integration of multi-source data and high-efficiency resource management.Artificial intelligence(AI),with its advanced computational capabilities,has emerged as a key enabler,driving substantial growth in AI-based resilience research.However,a comprehensive approach to AI integration remains underexplored.This article reviews recent advancements in resilient power systems and evaluates AI-based approaches across three critical areas:extreme event modeling,resilience assessment and early warning,and loss mitigation and restoration.Finally,the paper identifies future challenges and opportunities in this rapidly evolving field.