电力系统短期负荷预报的几种改进手段  

SOME IMPROVED APPROACHES OF POWER SYSTEM SHORT TERM LOAD FORECASTING

作　　者：侯志俭[1] 吴际舜 张琦雨 万亮 潘龙兴 

机构地区：[1]上海交通大学电力系

出　　处：《电力系统自动化》1996年第7期27-31,共5页Automation of Electric Power Systems

摘　　要：采用时间序列法（ＡＲＭＡ）对电力系统进行短期负荷预报，着重研究了负荷样本伪数据的处理以及如何建立从自动搜索定阶到节假日预报一整套的程序化模型，所编制程序在ＨＰ４８６ＰＣ机上获得通过。采用负荷数据为上海某供电局１９９５年５月份小时负荷报表，预报结果的日平均误差为１％～３％，最大误差不超过５％。The paper adopts the time series approach (ARMA) to forecast short-term load of power system. It mainly deals with the bad data and builds a program for the model which can automatically search for the order and forecasl holidays' load. The program has been passed on the compulcr of HP486/PC. We have used the hour-load data table in May. 1996 provided by a power system department in Shanghai. According to the approaches of this paper. the results show that the mean error is 1%￣3% per day and the maximum error of hour-load is less than 5%.

关 键 词：电力系统 负荷预报 专家系统 

分 类 号：TM715[电气工程—电力系统及自动化]