智能电网云数据储存平台次月留存数据聚类方法

Clustering Method of Retained Data in the Next Month of Smart Grid Cloud Data Storage Platform

为了优化智能电网业务质量,提出智能电网云数据储存平台次月留存数据聚类方法。业务数据处理模块以FPGA芯片作为核心,设计内部CLB连接形式,优化数据存储器的相关参数,并选择合适的中央控制芯片对平台硬件加以控制。在计算次月留存数据特征提取匹配度的基础上,确定智能电网云数据储存平台次月留存数据聚类中心。利用增益函数初始化Spark程序,通过设计次月留存数据聚类算法,实现了智能电网云数据储存平台次月留存数据的聚类。实验结果表明方法次月留存数据聚类效率高于90%,对于不同的数据集具有较高的聚类效率。

In order to optimize the service quality of smart grid,a clustering method for the next month's retained data of smart grid cloud data storage platform is proposed.The business data processing module takes FPGA chip as the core,designs the internal CLB connection form,optimizes the relevant parameters of data memory,and selects the appropriate central control chip to control the platform hardware.On the basis of calculating the feature extraction matching degree of the next month's retained data,the next month's retained data clustering center of the smart grid cloud data storage platform is determined.The spark program is initialized with the gain function,and the clustering algorithm of the next month's retained data of the smart grid cloud data storage platform is designed to realize the clustering of the next month's retained data.The experimental results show that the clustering efficiency of the proposed method is higher than 90%for the next month's retained data,and it has high clustering efficiency for difierent data sets.