Research On Load Simultaneous Factor Of Urban Community Based On Data Mining

Load simultaneous factor as an important parameter of power load forecasting, itsscientific selection is the foundation of rational development of distribution networkplanning. Therefore, based on perfecting distribution planning method, it is necessary toresearch on the method of selecting load simultaneous factor of urban community, toprovide a basis for distribution transformer capacity configuration and cable running ofurban community.This paper is based on data mining technology, puts forward a selection method forload simultaneous factor of urban community which considers multiple effects. Throughthe analysis on the main effect factors of load simultaneous factor, construct indexsystem of simultaneous factors’ influence factors. Then combine the K-means clusteringalgorithm with neural network model, forecast load simultaneous factors. Through apractical example verify the feasibility of the method, and based on this, develop thesoftware system of selecting load simultaneous factor of urban community.Firstly, construct index system of simultaneous factors’ influence factors. Accordingto all kinds of index in the index system, collect sample data of typical community inFoshan area, and do data preprocessing by constructing fuzzy membership function.Then use K-means clustering algorithm to do cluster analysis on the treated samples,through the evaluation of two kinds of clustering validity index—BWP and DB index todetermine the optimal number of clusters.Secondly, according to each kind of community sample after clustering, forecastsimultaneous factors by BP and Elman neural network model. The accuracy of the twomethods’ results is high, the Elman model is better than BP model in convergenceefficiency and prediction accuracy, prediction results of simultaneous factor is morepractical. Then based on the traditional BP model, compared with the method proposedin this paper, the results show the accuracy and superiority.Finally, develop selection system of simultaneous factors by Microsoft Visual C#language. System uses the SQLSEVER2005database to store data, use WPF interfaceframework to develop visual user interface, C#language to develop internal algorithm.By using modular design, the interface phase can be kept separate from code, has thecharacteristics of low coupling and reusable. The system interface is simple and beautiful,has perfect function, can select simultaneous factors, has guiding significance to practicalapplications in engineering project.