摘要
目前研究的电力系统用电数据异常监测系统的精度较差,导致监测效率较低。为了解决上述问题,设计了基于极限学习机的用电数据异常动态监测系统,系统分别设计硬件区域和软件区域。硬件区域设计了服务器接收器、芯片和接口零件、调度中心站、数据通信器以及监测器,硬件区域内所有器件的性能都是目前各个领域性能最佳的设备,以便达到提高基于极限学习机的用电数据异常动态监测系统运行速度的目的。在软件区域分别讨论了极限学习机的工作原理和极限学习算法,提高数据异常动态监测系统的监测训练速度和规范性,增加了格兰杰因果检验流程,对系统需要监测的数据进行检验,增加系统监测的精度。实验结果表明,基于极限学习机的用电数据异常动态监测系统的监测性能满足系统应用的指标标准,达到了系统设计的目标,可以进行推广应用。
The current research on the power system data abnormality monitoring system has poor accuracy,resulting in low monitoring efficiency.In order to solve the above problems,a dynamic monitoring system based on extreme learning machine for abnormal electricity data was designed,and the hardware area and software area were respectively designed for the system.The hardware area is designed with server receivers,chips and interface parts,dispatching center stations,data communicators,and monitors.The performance of all devices in the hardware area is the best⁃performing equipment in various fields at present,so as to improve the performance based on extreme learning machines.The purpose of dynamic monitoring of system operation speed when power consumption data is abnormal.In the software area,the working principle and the extreme learning algorithm of the extreme learning machine are discussed separately,to improve the monitoring training speed and standardization of the data abnormal dynamic monitoring system,increase the Granger causality test process,test the data that the system needs to monitor,and increase the accuracy of system monitoring.The experimental results show that the monitoring performance of the dynamic monitoring system for abnormal electricity data based on the extreme learning machine meets the index standards of the system application,reaches the goal of the system design,and can be promoted and applied.
作者
孙志杰
张艳丽
王利赛
刘继鹏
SUN Zhijie;ZHANG Yanli;WANG Lisai;LIU Jipeng(Metering Center of Jibei Power Grid Co.,Ltd.,Beijing 100045,China)
出处
《电子设计工程》
2022年第15期81-85,共5页
Electronic Design Engineering
关键词
极限学习机
用电数据
异常动态
异常监测
extreme learning machine
electricity consumption data
abnormal dynamics
abnormal monitoring