Knowledge graph (KG) is a database that stores the information and relationship between entities in the form of graph. This paper presents a knowledge graph construction method for grid substation knowledge graph based on multi-source data. The device connection, alarm information and ledgers are contained in KG, which benefits the substation operation and maintenance. First, entities and relationships are selected according to the actual composition of substation. Then, triples are extracted through domain dictionaries and conditional random field. Entity disambiguation is carried out before establishing the knowledge graph based on corresponding data sources. Finally, knowledge fusion is used to remove the redundancy in the database to complete the construction of the KG. Compared to traditional relational database, graph database can more intuitively show the connection relationship between devices. By adding more links to knowledge graph, it takes less time for monitoring personnel to locate accidents when dealing with alarm information.

knowledge graph;grid substation;multi-source data;named entity recognition;knowledge fusion