基于聚类分析理论的空间负荷预测方法  

Spatial load forecasting based on cluster analysis

作　　者：郭鹏伟[1] 黄桥林[2] 肖白[1] 

机构地区：[1]东北电力大学电气工程学院,吉林吉林132012 [2]河南省电力公司濮阳供电公司,河南濮阳457000

出　　处：《黑龙江电力》2011年第6期434-437,440,共5页Heilongjiang Electric Power

摘　　要：基于时序分析的空间负荷预测方法通常先把预测区域划分成若干元胞,然后对每个元胞进行基于时序的负荷预测,从而实现对整体的空间负荷预测。目前基于时序分析的负荷预测模型有很多,但其中任意单一模型均难以做到对每个元胞预测都是误差最小的,即无法保证对总区域的空间负荷预测结果的准确性。为此提出了一种基于聚类分析理论的空间负荷预测方法,该方法通过分析元胞运用不同负荷预测模型进行预测得出的预测结果和相对误差。实际算例结果表明、该方法正确、有效、可行。The spatial load forecasting based on time series analysis usually divides the forecasting area into a certain number of structure cells, of which the load forecasting based on time series is then carried out to realize the overall spatial load forecasting. There are lots of load forecasting models at present, any single model, therefore, is difficult to realize the minimum error of every structure cell forecasting, which means that the accuracy of the overall spatial load forecasting cannot be guaranteed. This paper, aiming at this problem, proposes the spatial load forecasting based on cluster analysis, which analyzes the forecasting and relative error of structure cells through different load forecasting models.

关 键 词：空间负荷预测 聚类分析理论 元胞 

分 类 号：TM247[一般工业技术—材料科学与工程]