Research On Key Technologies Of Graph Database For Grid Topology Management

With the rapid construction and development of the global energy Internet,a large number of new energy sources,distributed power sources,and energy storage devices have emerged,the power grid structure have become increasingly complex.The grid system records the characteristics of the grid network.The physical relationships,customer relationships,and asset relationships of the grid form a huge network topology.At the same time,massive grid operation data,condition monitoring data,and smart meter data are generated.The data can be used for user power range analysis,fault research judgment,and safety warning.The current data organization and processing of the power grid is mainly based on traditional relational database management systems.With the increase of data volume and business complexity,severe technical bottlenecks have been encountered in terms of storage scale,query efficiency,scalability,etc.Problems such as the average response time of the database reaches the minute level,the status update during the peak period of business is not timely,and the visualization performance is poor.The graph database has the advantages of quickness and accuracy in terms of storing,updating,and querying data,and can meet the performance requirements of large-scale data processing and large-scale speed of the power grid system.In the graph database,any electrical equipment with real attributes in the power grid can be displayed as a graphic element.By operating the graphic element,the actual operation of the actual physical equipment can be completed.At the same time,the modularization of the attributes of the graphics elements and equipment greatly improves the adaptability and scalability of the program.Therefore,in order to solve the above-mentioned problems in the topology management of the power grid,this dissertation conducts research from three aspects of power graph data modeling,graph data storage,and graph retrieval query.Aiming at the problem of the visualization performance of grid topology data,the data in the relational database needs to be migrated to the graph database.This dissertation uses the graph database technology to build the power graph data model.We extract the power system equipment information as the nodes,and the real physical connection relationships as the edges of the topological graph.With data migrating,we design the basic primitives to form a large power grid topology graph,which can clearly and intuitively show the power equipment connection relationship,and it is easy to complete the power diagram.Aiming at the problem of fast index entry in the index storage of grid topology graphs,based on constructing frequent subgraph indexes,this dissertation proposes a serial storage optimization method based on multi-hash algorithm.This method mainly uses a multi-hash exploration sequence to make keywords stored uniformly in each position in the hash table,which can avoid the conflicts in the storage process and can realize the fast search of the index.Aiming at the query efficiency of grid topology data,this dissertation proposes a node filtering algorithm based on the graph algorithm.This algorithm uses space for time to filter out most nodes that are not in the optimal path from the graph.In order to achieve a fast retrieval query,this dissertation introduces a node blocking strategy based on the width-first search algorithm(BFS)bidirectional search,and uses the results of the single-source shortest path nodes obtained to accelerate the solution of the shortest point-to-point implementation.Finally,the experiment of power system data of a branch of State Grid Corporation was used to verify the effectiveness of the algorithm.