基于并行K-Means聚类的配电网台区用户行为分析模型研究及应用

Research and Application of Distribution Network Courts User’s Behavior Analysis Model Based on Parallel K-Means Clustering

当前,电力企业营销部门在电能计量过程中收集的日冻结电量数据仅仅以个体为单位,用于电费计算及线损评估,并未能将海量数据关联起来开展分布式的云计算。日冻结电量的大数据价值还未被挖掘。利用K-means聚类算法,搭建Hadoop大数据分析架构,对电网营销海量日冻结电量数据进行了剖析挖掘,实现了对用户用电行为的高效评判,为电力企业管控低压台区违约用电部署了新的技术手段,充分发挥了电网大数据资源的价值,为电力企业挽回可观的经济损失。

At present,the power enterprise marketing department in the energy measurement process to collect the daily freezing of electricity data only individual units for the calculation of electricity and line loss assessment,and failed to link the massive data to carry out distributed cloud computing.The value of the large data of frozen electricity has not yet been excavated.In this paper,the K-means clustering algorithm is used to build the Hadoop large data analysis structure,and the data of the electricity sales volume of the grid marketing are analyzed and excavated,which realizes the high efficiency judgment of the electricity consumption behavior of the enterprises.The deployment of new technical means,give full play to the value of large data resources for the power companies to restore considerable economic losses.