Automatic schema construction of electrical graph data platform based on multi-source relational data models

Abstract

Data storage and management in power systems usually adopt relational databases. However, the relational database requires ample storage space and has low data retrieval and query efficiency. An electrical graph data platform can describe the complicated relationships between concepts and entities involved in power systems with the form of an association graph, which provides a better ability to organize, manage, and apply massive amounts of information. Since the construction of the top-level ontology model or schema for a specific field graph data platform is cumbersome, complex, and generally requires lots of association analysis and expert system intervention, it is insufficiently automated, time-consuming, and unable to cope with large-scale electric power knowledge. This paper proposes a method for automatically constructing the schema of an electrical graph data platform, which uses the diverse table structure information from the relational database and SQL language descriptions to extract ontologies to form the ontology candidate set automatically. Then the method utilizes ontology clustering and disambiguation to initial an ontology graph model and automatically update ontology and relationship expressions. Meanwhile, the model layering is used to construct a hierarchical model based on different business needs, and the schema optimization is applied according to expert comments.