摘要
分析了实际电力系统中负荷异常数据的主要成因,并针对2类主要的坏数据各自的特点,分别使用不同的方法处理负荷预测样本数据。针对自动化系统故障造成的坏数据,提出了具有负荷预测应用特点的总加值动态多源处理技术,从而能够充分利用采集设备或网络通道对负荷总加值而言的多重冗余配置;针对大负荷的突发性偶然波动造成的坏数据,采用对电网终端负荷的逐一扫描辨识,部分避免了对单一总加数据预处理的误判和漏判。
The major reasons for the bad data of real power system loads are analyzed and a bad data identification and correction method for two major kinds of bad data is presented. The method dredges the resource of energy management system (EMS) fully. In order to resolve the had data resulting from automatic system fault of electric power systems, a new method of dynamic multi-source data disposal for the sum of electric load is proposed. The redundancy of the monitoring and transfer systems is exploited to insure the validity of the sum of load. Aimed at the identification of bad data caused by the big load accident, the method scans the ultimate big load one by one to prevent the misidentification of had data. The new method has the advantages of high efficiency and validity for bad data identification and correction.
出处
《电力系统自动化》
EI
CSCD
北大核心
2006年第15期85-88,共4页
Automation of Electric Power Systems
关键词
能量管理系统(EMS)
负荷预测
异常数据辨识
多源数据
energy management system (EMS)
load forecasting
anomalous data identification
multi-source data