Qualitative Knowledge Graph Model Construction Method of Transformer Maintenance

The existing transformer operation and maintenance knowledge resources are scattered, and less intelligent in knowledge management and utilization, so they cannot be applied in the occasions which have high requirements to efficiency and accuracy. Knowledge graph not only have the ability to link multi-source knowledge, but also improve the efficiency of knowledge utilization by graph management technology. Therefore, the paper proposed a method for establishing the knowledge graph of transformer operation and maintenance based on the physical structure and physical characteristics of the equipment. Then, the physical equipment, physical quantity, physical phenomena and equipment applications were used as the ontology. Furthermore, the hierarchy of knowledge entity was constructed according to the physical relationship and application relationship between the ontology. As well as, the relationship between the ontology was described by the structural association and functional association between the various components of the transformer. The structural association reflected the structural relationship of the transformer manufacturing, the installation and fixing etc.. The functional association relationship reflected the qualitative relationship in physical quantities of transformer operation. Finally, an oil immersed transformer knowledge graph was constructed, and a graph search method was illustrated as a search path. Combining with qualitative reasoning technology, the search work was behaved from the ontology of abnormal phenomena to the query target along the road provided by the graph. Conclusively, the efficiency and accuracy of the proposed method is verified.