基于回归分析的频域分解短期负荷预测方法  

Short-term load forecasting based on frequency analysis of regression

作　　者：姜玉山[1] 刘嘉婧[1] 卢毅[1] 

机构地区：[1]东南大学电气工程学院,南京210096

出　　处：《电力需求侧管理》2011年第5期12-16,共5页Power Demand Side Management

摘　　要：基于回归分析的频域分解短期负荷预测方法通过对负荷序列进行频域分解,将负荷序列分解为周期负荷分量、低频负荷分量以及随机负荷分量。采用了外推法、一元线性回归分析和加权平均法,分别对周期负荷分量、低频负荷分量和随机负荷分量进行预测。最后通过对某地区电网日负荷进行预测,表明该方法应用于短期负荷预测具有较好的预测精度。Based on regression analysis frequency domain de- composition of short-term load forecasting, by a sequence of load frequency domain decomposition, load sequence will be bro- ken down into components of the load cycle, low-frequency load component and random load component. The article uses extrap- olation method, a linear regression analysis and weighted aver- age method, respectively forecasts the cycle load component, low -frequency load component and random load component. Finally by forecasting grid daily load in some area, numerical results show that the method applied to short-term load forecasting has better prediction accuracy.

关 键 词：回归模型 频域分解 负荷预测 预测精度 

分 类 号：TM715[电气工程—电力系统及自动化] F407.61[经济管理—产业经济]