电力运营基础资源数据信息采集及挖掘技术研究  

Research on Data Collection and Mining Technology of Basic Resources of Electric Power Operation

作　　者：赵强 高振峰 于卓智 熊艳 刘路峰 ZHAO Qiang;GAO Zhenfeng;YU Zhuozhi;XIONG Yan;LIU Lufeng(State Grid Information&Telecommunication Co.,Ltd.,Beijing Branch,100031,Beijing,PRC)

机构地区：[1]国网信息通信产业集团有限公司北京分公司,北京100031

出　　处：《江西科学》2021年第6期1088-1093,共6页Jiangxi Science

基　　金：国网信息通信产业集团有限公司两级研发项目“面向多站融合及能源大数据中心商业化运营的电力基础资源管理平台及网络运营安全管控平台研发”(52680020002W)。

摘　　要：针对电力运营基础资源数据量较大、异构数据较多,难以在短时间内获得有效信息的问题,提出基于归一稳态平衡的信息数据采集与挖掘算法。分析电力运营资源数据的分布特点,得出运营数据源间的共性,根据样本数据源的族群变化,将资源数据看作节点族群采集计算。将资源数据空间按照维度划分为多个深度局域,在局域中融合所有数据源,计算整体数据源的模态向量,提取出差异性较大的向量数据,作为标准化挖掘矩阵的代入因数,以在时间序列关系下的斜差异率作为判定阈值,完成数据挖掘。仿真实验证明,所提方法的通信开销较小,电力运营基础资源数据信息采集的协同性较好,挖掘偏差较小。Aiming at the problem of large amount of basic resource data and heterogeneous data in power operation,it is difficult to obtain effective information in a short time,and an information data collection and mining algorithm based on normalized steady-state balance is proposed.Analyze the distribution characteristics of the power operation resource data,and draw the commonalities between the operation data sources.According to the ethnic changes of the sample data sources,the resource data is regarded as the collection and calculation of the node ethnic group.The resource data space is divided into multiple depth local areas according to the dimensions,all data sources are merged in the local areas,the modal vector of the overall data source is calculated,and the vector data with large differences is extracted as the substitution factor of the standardized mining matrix.The slope difference rate under the time series relationship is used as the judgment threshold to complete data mining.The simulation experiment proves that the communication cost of the proposed party is small,the data collection of the basic resources of electric power operation has good coordination,and the mining deviation is small.

关 键 词：电力运营 异构数据 族群 维度划分 模态向量 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]