Study On Spatial Load Forecasting Method Based On Support Vector Machine And Multilevel Clustering Analysis

Spatial load forecasting (SLF) is predicting the load value in the future andstudying the spatial distribution of the load. As the premise and matting of urbandistribution network plan, the accuracy of SLF is directly related to funding,equipment distribution pattern and economic reliability for planning. Appropriateselection of the type and layout of power equipments such as transformer substationand transmission line is an important content of urban distribution network planning,SLF provides reference and basis for the planning. Improving the prediction accuracyand effect of SLF are among the emphasis and difficulty in the research field, thereasonable selection and perfect improvement can fully reflect the development trendof the cell. In this paper, in order to satisfy the demand of urban power grid planning,the actual operation mode is combined with support vector machine (SVM) theory,and the forecasting results and spatial distributions of load are analyzed andexpounded.In the past much research work is focused on proposing and improving SLFprediction methods, rarely from the perspective of historical load characteristic, trendextrapolation is one of the methods usually used in SLF, but rarely considers involvesthe power load characteristic analysis. Therefore, in the paper firstly the loadforecasting methods in the urban power grid is classified and analyzed one by one;then according to different classification principle the power load is classified and theconcepts of power space resolution and power supply district division rules areintroduced to analysis the characteristics of urban electricity power space load; thenSVM theory is applied to SLF to put forward the SLF method based on SVM, and byusing the connection between cellular load and properties to improve the predictionmethod, so SLF method based on multilevel clustering analysis and support vectormachine (MCA-SVM) is proposed, the feasibility and effectiveness of the method isverified as an example of Chuanying district Jilin; Finally, the important research tools and platforms of spatial load forecasting are introduced---geographicinformation system (GIS), and by using the MapInfo software to make a introductionof the functions of geographic information system in the SLF such as data andgraphical information and secondary development program.