中长期电力负荷的灰色残差修正组合预测  

Gray System Theory and Residual Corrected Model Combination Forecasting of Long-term Power Load

作　　者：卢建昌[1] 裴乐萍[1] 

机构地区：[1]华北电力大学经济管理系,河北保定071003

出　　处：《电气技术》2013年第7期7-10,共4页Electrical Engineering

摘　　要：中长期电力负荷走向趋势曲线往往呈S型,直接运用灰色GM(1,1)模型预测会造成较大的偏差。针对负荷曲线固有的发展趋势,根据曲线发展的不同阶段分阶段预测,采用GM(1,1)模型进行第一阶段的预测,采用残差修正模型进行第二阶段的预测。在分阶段预测的基础上,根据每一阶段的预测精度分配权重,进行组合预测。该方法将数据本身发展趋势和模型固有优点相结合,预测结果表明该方法在提高精度上有可行性。The trend curve of long-term power load tends to be S type, so if it uses the GM（1,1） model directly, it will cause a large deviation. According to the inherent development trend of load curve, it can forecast the load in stages following the different development stages of curve, and in the first stage, it uses GM（1,1） model, in the second stage, it uses residual corrected model. On the basis of forecasting in stages, it assigns weights according to prediction accuracy of each stage to conduct combination forecasting. This method combines the data itself development trend and the model inherent advantages, and the forecasting results show the method is feasible to improve forecasting accuracy.

关 键 词：负荷预测 灰色GM(1 1)模型 残差修正 组合 

分 类 号：TM714[电气工程—电力系统及自动化]