Integrating knowledge graph and network topology: An innovative construction method of power safety accident case think tank

In this paper, combined with knowledge graph technology and network topology analysis, an innovative construction method for a power safety accident case think tank is proposed. A knowledge model for power safety accident cases was constructed, including five major entity categories and four types of relationships. Topology analysis indicators were formulated, and a construction method for an accident case think tank was proposed. Finally, the key links in power safety accident prediction and management were determined. The results indicate that the frequently-occurring accident categories were identified as “fall from height,”“electric shock,”“tripping,” and “fire.” The higher cause accessibility indexes of the causes of falling from heights and electric shock accidents were human’s unsafe behaviors, and their consequences were usually relatively large accidents. The operation types that had a greater influence on them were construction operations and inspection operations. Falls from height often occur in power production units, while electric shock accidents often occur in power supply units. For tripping and fire accidents, the higher cause accessibility indexes of the causes were the unsafe state of equipment. Through the accident case think tank, the high-incidence risks and potential hazard points of power safety can be understood in a targeted manner.