Cause Analysis of Substation Faults Based on Multimodal Fusion Detection and Knowledge Graph

The expansion of the grid scale puts forward higher requirements for the stability of substation operation, so it is necessary to detect faults in time and analyze possible fault causes through intelligent inspection. However, commonly used fault detection methods based on image processing or single-mode data such as voltage and temperature are likely to cause misjudgment of faults and cannot realize the function of fault cause analysis. Based on this, this paper proposes a substation fault cause analysis model based on multi-modal data fusion and knowledge graph. The cause of the fault is inferred based on the substation fault cause knowledge graph. For the first time, this solution comprehensively uses target detection technology, audio recognition neural network and knowledge map, and achieves a fault cause analysis accuracy rate higher than 90% on the substation fault data set we constructed.