Risk Mitigation Strategy for Power Grid Construction Accidents Based on Knowledge Graph Representation Learning

Abstract

In order to solve the problems of lack of accident relation, unstable response results and insufficient diversity in power grid construction accident risk handling, a method of power grid construction accident risk handling based on knowledge graph representation learning was proposed. First of all, for enhancing the correlation between accidents, the power grid construction accident report was taken as the data source, and the information graph structure was constructed according to the accident elements and the relationship between the elements through knowledge graph technology. Secondly, for improving the retrieval speed and accuracy, Bert model and GraphSAGE model were used to complete the knowledge graph representation learning and the accident feature vector was obtained. Finally, for having more comprehensive knowledge coverage, the accident case database was established to compare and match the new accident knowledge map. The risk control and emergency treatment of power grid construction accidents we carried out. The results show that the method combined with GraphSAGE model has better processing effect on knowledge graph, the stability of output data is improved by 12.3%, and it can make rapid response, forming an efficient accident risk processing system. It is concluded that the designed method can significantly improve the efficiency of power grid construction accident handling, and provide reference and support for power grid construction accident risk handling.