考虑积温效应的短期负荷组合预测方法  

Combined short-term load forecast with accumulated temperature effect

作　　者：蒋建东[1] 程志豪[1] 朱明嘉[2] 

机构地区：[1]郑州大学电气工程学院,河南郑州450001 [2]济源供电公司,河南济源454650

出　　处：《电力自动化设备》2011年第10期28-31,共4页Electric Power Automation Equipment

摘　　要：给出了积温效应的量化公式,提出了考虑积温效应的小波变换与神经网络负荷组合预测方法。该方法首先通过小波变换,把负荷序列分解为不同频段的趋势性负荷子序列和细节性负荷子序列;采用BP神经网络对各子序列分别进行建模和预测,对趋势性负荷子序列加入了积温系数等特征参数;最后由小波重构得到负荷序列的最终预测结果。该方法考虑了积温效应的影响,充分利用了小波变换与神经网络的优点。算例结果表明了所提出方法能有效提高负荷预测的精度。The quantitative formula of accumulated temperature effect is given and the load forecast combining wavelet transform and neural network is proposed with the accumulated temperature effect. The load series is decomposed into the trend load series under different frequency bands and the detailed load series by wavelet transform,which are modeled and forecasted by BP neural network. The characteristic parameters,such as the accumulated temperature coefficients are added into the trend load series. The final forecasting result of load series is obtained by wavelet reconstruction. Case study shows the proposed method improves the load forecast accuracy effectively.

关 键 词：负荷预测 积温效应 小波变换 神经网络 模型 

分 类 号：TM714[电气工程—电力系统及自动化]