关于多模型方法预测电力负荷的研究  

Study on Power Load Forecast with Multi-Model Method

作　　者：俎洋辉[1] 李鸿伯[1] 刘岩[1] 

机构地区：[1]安阳供电公司,河南安阳455000

出　　处：《河南电力》2013年第3期10-12,26,共4页Henan Electric Power

摘　　要：为提高电力系统负荷预测准确度,针对当前负荷特性研究中出现的问题,通过分析多种气象因素与电力负荷的对应关系,提出了一种新型的电力负荷预测的多模型方法。该方法首先采用多个神经网络模型来实现低层的预测;然后通过气象因素修正电力负荷结果,进而构建全局预测模型。通过实例研究,证实了所提方法的有效性。In order to improve the precision of the electric load forecasting, in view of the problems of the load characteristics research ,by analyzing the co-relationships between climatic factors and electric loads, a new multi-model power load forecasting method is presented.The methods use multiple neural networks for the first layer, and revise the results by means of the climatic factors, and then construct a global prediction model.The results of case study verify that the method is effective.

关 键 词：电力负荷预测 多模型 神经网络 

分 类 号：TK012[动力工程及工程热物理]