Community Gas Pipeline Emergency Response Recommendation Based on Graph Representation Learning

Community gas pipeline accidents are sudden and rapid, and need to be efficiently and comprehensively carried out after the accident discovery report. In this paper, we pioneered the use of graph-structured accident case representation model to realise similar case recommendation, and took the emergency response to the "6·13" gas explosion accident in Shiyan as an example, supplementing the disposal tasks such as probing the edges of gas leakage into confined space, cleaning up and eliminating the explosion conditions. The results show that the case recommendation based on the case-feature graph overcomes the limitations of accident information format and granularity; on the basis of recommending similar environment and type of accidents, the model can also be extended to the evolution of accidents, which plays a positive role in the comprehensiveness of emergency response and the prevention of secondary accidents.