基于聚类算法的电网分布式增值业务数据集成系统

Data Integration System for Distributed Value-added Services in Power Grid Based on Clustering Algorithm

为了缓解电网数据存在的孤立与异构问题,提升电网数据集成效果,基于聚类算法设计了电网分布式增值业务数据集成系统。依托Spark平台,建立由应用层和基础平台层构成的电网分布式增值业务数据集成系统;利用B/S结构和MVC模型内前端开发工具包完成界面设计;设置在后端服务器,负责聚类分析与权限查询等功能;以基础平台层为系统提供计算服务。利用Hibernet框架,联合Canopy算法与k-medoids聚类算法,实现电网分布式增值业务数据集成。实例测试结果表明,该系统受线程数量影响较小,聚类能力好,提高数据集成能力。

In order to alleviate the isolated and heterogeneous problems of power grid data and improve the effect of power grid data integration,a distributed value-added service data integration system based on clustering algorithm is designed.Based on Spark platform,a distributed value-added service data integration system for power grid is established,it is composed of application layer and basic platform layer.The interface design is completed by using B/S structure and front-end development kit in MVC model.It is set in back-end server,responsible for clustering analysis and authority query.The basic platform layer provides computing services for the system,and the Hibernet framework is used.Combined with Canopy algorithm and k-medoids clustering algorithm,the distributed value-added service data integration of power grid is realized.The results show that the system is less affected by the number of threads,has good clustering ability,and improves the ability of data integration.