Abstract:Encountering extreme natural disasters, traditional methods are difficult to accurately capture the dynamic characteristics and risk evolution laws of disaster events. This article proposes a power grid fault handling framework based on Knowledge Graph (KG). This framework combines power grid topology graph, operational logic graph, and fault case graph, aiming to enhance the situational awareness (SA) capability of the system and enhance the observability and controllability of the information physical fusion system. Through real-time big data mining and the construction of a full process knowledge graph link, advanced perception, scenario description, situation understanding, and evolution trend analysis of extreme natural disaster events have been achieved. This paper also takes "720 extremely heavy rainstorm Heavy Rainfall" as an example to show the whole process of pre disaster early warning, emergency response strategy in disaster and post disaster recovery; In addition, the dynamic update of fault handling KG ensures the continuous accumulation of knowledge and provides effective knowledge guidance support for similar disasters. This research not only improves the response capability of the new power system under extreme natural disasters, but also provides strong theoretical support and practical guidance for future energy Internet disaster response.