Modeling methods of big data for power grid based on graph database

The analysis and utilization of grid big data, big data driven grid analysis and optimization have received extensive attention in recent years, and non-relational databases can handle big data problems efficiently. The graph database is a kind of non-relational database. This paper takes the graph database Neo4j as an example. Based on the domain modeling theory and follows the CIM model, the grid data modeling principle is proposed. Under these principles, some power grid data models and their applicable occasions, and methods for converting these models to each other are proposed. Finally, three graph database models are built using the modeling method proposed in this paper. The query performance of Neo4j and MySQL is compared by three examples. The conclusion is that in the traversal with obvious topological characteristics, such as shortest path traversal, deep traversal and time tree traversal, etc., the graph database model proposed in this paper can greatly improve the retrieval efficiency. It proves the effectiveness of the proposed modeling method and the efficiency of the graph database in specific retrieval.