基于相似日负荷修正算法的短期负荷预测  

Short-term load forecasting based on algorithm of similar days'load modification

作　　者：刘晶[1] 朱锋峰[1] 林辉[2,3] 郝志峰[4] 吴广潮[4] 

机构地区：[1]华南理工大学理学院,广东广州510640 [2]华电集团公司 [3]惠州供电分公司,广东惠州516000 [4]华南理工大学计算机科学与工程学院,广东广州510640

出　　处：《计算机工程与设计》2010年第6期1279-1282,共4页Computer Engineering and Design

摘　　要：为实现准确、快速预测电力系统短期负荷的目的,综合考虑气象、日类型和时间对负荷的影响,提出了基于相似日负荷修正算法的预测模型。首先建立相似度量化模型,具体用灰色关联分析法计算气象相似度,兼顾"近大远小"和"周期性"原则来量化时间相似度,二者乘积作为总体相似度,依此选取若干相似日;然后基于"日类型"和"时间跨度"修正相似日负荷;最后用加权平均法预测负荷。短期负荷预测的实例结果表明了该算法的可行性。In order to make short term load forecasting more accurate and faster, the algorithm based on the similar days＇load modification is proposed by taking into consideration of the effects that meteorology, day type and time have on the short-term load forecasting.Firstly, several models are established to evaluate the similarity degree.In particular, the similarity degree in meteorology is obtained with grey correlation analysis method while the similarity degree in time reflected＂near greater far smaller＂and＂cyclical＂principal.The product of these two is to evaluate the overall similarity, based on selecting the similar days.Moreover, the loads of the similar days are modified on the basis of the day type and the time span.Finally, the load is forecasted by weighted average method.An example of short-term load forecasting is presented to verify the feasibility of the method.

关 键 词：相似日 短期负荷预测 负荷修正 气象 日类型 时间跨度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]