基于多阶段近邻排序法与实体匹配法的电能质量关联分析方法  

Power Quality Association Analysis Method Based on Multi-Stage Sorted Neighborhood Method and Entity Matching Method

作　　者：周刚 肖斐[2] 艾芊[2] 谢善益 王玥[2] ZHOU Gang;XIAO Fei;AI Qian;XIE Shanyi;WANG Yue(Electric Power Research Institute,Guangdong Power Grid Corporation,Guangzhou 510006,China;School of Electronic Information and Electrical Engineering,Shanghai Jiaotong University,Shanghai 200240,China)

机构地区：[1]广东电网有限责任公司电力科学研究院,广东广州510006 [2]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《电器与能效管理技术》2019年第14期14-20,33,共8页Electrical & Energy Management Technology

基　　金：广东电网公司科技项目资助(GDKJXM20162540)

摘　　要：在复杂电网环境下,电能质量监测系统检测到的扰动事件与设备动作之间可能具有关联特性。为分析挖掘扰动事件的关联特性,基于电能质量扰动数据从数据清洗、关联分析、时空分布可视化展示方面开展研究。针对扰动能量传播特性,提出一种多阶段近邻排序数据预处理方法,有效划分扰动时段并剔除重复数据。其次利用实体匹配方法融合同一扰动时段的电能质量数据和电网动作事件数据,实现电能质量扰动事件关联分析。在此基础上,运用ArcGIS10.1中的热度分析工具,以可视化的表达形式揭示出:电能质量扰动事件发生次数在短时间尺度范围内存在空间相似性。同时为电能质量数据可视化展示提供了一种尝试。In the complex power system,the disturbance events detected by the power quality monitoring system may be associated with device actions.In order to analyze these association properties,this paper investigated on data cleaning,association analysis and temporal-spatial distribution visualization based on power quality disturbance data.First,a multi-stage sorted neighborhood data preprocessing method was proposed to divide disturbance periods and delete duplicate data effectively.Then,the entity matching technology was used to integrate power quality data and power system equipment action data within the same disturbance period to analyze the correlation between power quality events and device actions.Based on that,with the application of the hot-spot analysis tool in ArcGIS10.1,this paper demonstrated that the spatial similarity of the number of power quality disturbance events exists in short time scale.In addition,the proposed method provided a new attempt for power quality data visualization.

关 键 词：电能质量 关联分析 时空分布可视化 实体匹配 

分 类 号：TM711[电气工程—电力系统及自动化]