摘要
为充分利用并挖掘电力系统中海量的用电数据进行用电异常状态的识别,本文基于计量自动化系统智能电能表所采集的用电大数据,用随机矩阵理论对用电异常状态、异常时间段及异常地点的辨识进行研究。首先,介绍了用电大数据构造高维随机矩阵的方法,并分析了高维随机矩阵的协方差矩阵特征谱分布规律;然后,根据矩阵的特征值统计特性变化规律提出基于用电大数据矩阵的用电异常状态辨识及定位方法;最后,以贵州各大行业实际用电数据为例,对不同的行业进行了仿真验证,仿真结果表明该方法能准确识别用电异常并判别异常时间段和异常地点,不仅能满足电网对可视性、时效性、可靠性和安全性的迫切要求,而且为数据驱动用电环节智能化、可视化监控提供了新思路。
In order to fully utilize and excavate the massive power data in the power system to identify the abnormal state of electricity consumption,this paper uses the random matrix theory to analyze the abnormal state of the electricity consumption based on the power big data collected by the intelligent watt-hour meter of the measurement automation system,Meanwhile,the identification of abnormal time period and abnormal location is studied.Firstly,the method of constructing high-dimensional random matrix by using big data is introduced,and the distribution law of characteristic spectrum of high-dimensional random matrix is analyzed.Then,according to the statistical characteristic law of the matrix eigenvalue,the paper puts forward the method of identifying and locating the abnormal state of power consumption based on the power big data matrix.Finally,the paper takes the actual situation of various industries in Guizhou province as case studies,the simulation results of different industries show that the method can accurately identify the abnormal power consumption and distinguish the abnormal time period and abnormal location,which can not only meet the urgent requirements of visibility,timeliness,reliability and security,but also provides a new idea for the intelligent and visual monitoring of data driven electricity.
作者
张秋雁
岑远洪
安静
丁超
邵峥
王蓝苓
ZHANG Qiuyan;CEN Yuanhong;AN Jing;DING Chao;SHAO Zheng;WANG Lanling(Electric Power Research Institute of Guizhou Power Grid Co.,Ltd.,Guiyang 550001 Guizhou,China;Xingyi Power Supply Burean of Guizhou Power Grid.Co.,Ltd.,Xingyi 562400 Guizhou,China)
出处
《电力大数据》
2019年第5期41-48,共8页
Power Systems and Big Data
关键词
用电异常状态
大数据
随机矩阵理论
状态辨识
abnormal state of electricity consumption
big data
random matrix theory
state identification
