Research And Application Of Knowledge Graph Of Power System Based On Graph Database

With the advent of power big data era,a lot of technologies represented by the Internet of Things,cloud computing and artificial intelligence have been gradually used in the field of power,and the information age of the power industry has arrived.Knowledge graph technology has become a new driving force for innovation and development in all walks of life.However,the research on knowledge graph in power field is still very rare.Therefore,it is urgent to introduce knowledge graph technology into power field and build knowledge graph for power field.Through the research and application of the power knowledge graph,it is convenient for the power personnel to inquire and analyze the power information and improve the utilization rate of the power grid information.The main research content of this study is the research and application of power domain knowledge graph.Based on the Neo4j graph database,the power domain knowledge graph is stored,the power domain knowledge graph is built,and the search engine based on the power system knowledge graph is finally realized.The specific research work completed is as follows.First of all,in the work of knowledge extraction of power system.For power entity extraction,my study uses LSTM-CRF model to complete power system entity extraction.By comparing the results of single neural network model,such as recurrent neural network model,it is proved that the LSTM-CRF model adopted in this study has better effect on entity extraction of power system.For power entity relationship extraction,GRU-PCNN combined model is adopted in this study to achieve the extraction of power entity relation.By comparing and analyzing other relational extraction models such as CNN,RNN and other network models,the experimental results show that the combined model selected in this study is superior to the single neural network model in both accuracy rate and recall rate.The power knowledge is extracted by extracting the power entity and the power entity relation.Secondly,for the construction of the power system knowledge graph work.For knowledge storage of power knowledge extraction results,Neo4j graph database is selected in this study through comparison and analysis.As the most popular graph database in recent years,Neo4j has powerful functions and provides good technical support for the storage of power knowledge.Therefore,this study builds knowledge graph of power system based on Neo4j graph database.By using the visual interface of Neo4j,power information can be queried by Cypher language provided by Neo4j.Finally,in the search engine realization based on the power knowledge graph,the B/S framework is adopted for the design of the search engine,and the realization of the power search engine based on the power knowledge graph is completed through the coding of the front and back end of the search engine and the test of the search engine.