Research On Power And Line Loss Analysis Based On Clustering Algorithm And Spark Framework

Line loss level is an important comprehensive index to judge and evaluate the production technology and management level of power supply enterprises.Starting from the actual demand,the article develops research and analysis on quantity of electricity and line loss.The main research contents are as follows:Firstly,at the condition of the quantity of electricity data of complex,incomplete now researching on various clustering algorithms,a limit learning algorithm based on One Class SVM algorithm is proposed.Detection and recognition of abnormal data based on limit value learning,And the density based clustering algorithm DBSCAN algorithm is analyzed to realize outlier detection,proposed the abnormal data identification algorithm based on limit value learning;Combining with local anomaly factor(LOF)algorithm to analysis and research,the multidimensional data outlier identification based on density clustering algorithm is studied.Secondly,the framework of power grid line loss calculation based on CIM is studied.Through the detailed study of CIM standard,the mapping relation between the calculation of power grid line loss and CIM is determined,and the CIM model is extended to applied to the calculation of power grid line loss.In addition,on the basis of theoretical research on the line loss calculation method,a theoretical line loss calculation model of full voltage level is established,and various line loss models are studied and designed and the analysis and research on the residual electricity model,a new model is designed for the residual power.Finally,a parallel distributed line loss calculation method based on Spark framework is proposed.First to Spark the basic structure and ecosystem of big data technology and general big data analysis platform architecture are studied,and then to the function of the traditional line loss analysis system and the line loss analysis system based on the technology of big data demand analysis,through the line loss calculation module and data processing and data integration and management to improve the optimization,implements the parallel distributed line loss calculation method based on the framework of Spark.