灰色自适应模型及在负荷短期预测中的应用  

GREY-ADAPTABILITY MODEL FOR SHORT RUN PREDICTION OF THE LOAD

作　　者：施泉生[1] 奚辉龙 

机构地区：[1]上海电力学院

出　　处：《电力系统自动化》1993年第2期16-18,共3页Automation of Electric Power Systems

摘　　要：本文针对电力负荷变化的非平稳性,首先建立起灰色模型分析用电负荷的趋势项,用灰色模型值与原始数据值进行比较,得出一系列误差值,作为二级数据,进一步应用自回归AR(2)模型对灰色模型产生的误差进行适应性调整。以1991年上海市用电负荷为例,预测二周用电负荷的平均相对误差为1.65％,对于特殊的节假日负荷,根据其自身的特点,本文则采用二元线性拟合模型进行预测,以避免产生大的预测误差。采用本文的系统具有一定的自适应性,对于非平稳性的电力负荷具有很好的预测效果。In view of the non-stationarity of electric load,the authors first set up the grey model for the analysis of the load trend.By comparing the grey model values with the original ones,a series of errors is got.Then the auto-regression AR(2)model is applied to the adaptability modification of the errors. Taking as an example the electric load in Shanghai in 1991,the author predicts that the average relative error of a two-week load is 1.65%.As for a special holiday load which has its own features, the binary linear rearession model is adopted in order to avoid large errors. This system has self-adaptability and can make satisfactory prediction for non-stutionary electric load.

关 键 词：灰色模型 电力负荷 预测 

分 类 号：TM714[电气工程—电力系统及自动化]