Design And Implement Of Power Grid Fault Prewarning System Based On Heterogeneous Data Fusion

Big data fusion in power systems is the inevitable demand for informatization and intelligent development of the power industry.Under the new situation of the explosive growth of big data in power equipment,traditional data processing technologies have encountered bottlenecks and cannot meet the needs of quickly acquiring knowledge and information from massive data.The use of power big data has received extensive attention.The use of power big data to carry out power grid fault diagnosis has important research value and application prospects.It is of great significance to the stable,safe and reliable operation of the intelligent grid,and has become a research hotspot for the further development of the intelligent grid.Aiming at the application of big data fusion technology in power systems,this thesis focuses on the unified modeling and access technology of multi-source heterogeneous data in power systems,and uses artificial intelligence(AI)algorithms to prewarning equipment faults in severe weather.The specific research results are as follows:1)Based on the actual construction of the power grid,structured and semi-structured the multi-source heterogeneous data model is proposed,by expanding the necessary attributes for various primary equipment.The Hadoop framework is used to establish a global data index function to achieve multi-dimensional unified modeling of data measurement.Further,through the D5000 platform commercial library management tool and the integrated synchronization mechanism of the graph and model library,the integration of multi-source heterogeneous big data access methods enables automatic synchronization of data point changes in the dispatch control system to the big data system,realizing data free of operation and maintenance Point management.2)Based on AI algorithms such as decision trees and random forests,big data of power is used to realize the diagnosis of sudden failures of power system equipment and lines caused by severe weather.Firstly,the typhoon and thunder related information is predicted to realize the bad weather early warning,and then the historical fault information of power grid is combined with the meteorological early warning influence information to realize the fault early warning and diagnosis of power grid lines in such bad weather as typhoon and thunder.3)Based on the Flask framework,a visualization system for equipment fault prewarning under severe weather is further built to verify the above fault prewarning methods.Taking actual typhoon and lightning environmental data as examples,online early warning data sets are generated by analyzing the real-time forecast.Information provided by the big data system,as well as the line location,is measured.The generated prediction sample data is employed into the trained prediction model for online equipment fault diagnosis,and the results are displayed on the visualization platform.