基于小波分析的电力负荷模型辨识数据去噪  

A wavelet-based method for de-noising data of electric load modeling

作　　者：熊传平[1] 曹军杰[2] 陈谦[2] 鞠平[2] 

机构地区：[1]河南省电力公司,河南郑州450052 [2]河海大学能源与电气学院,江苏南京210098

出　　处：《河海大学学报（自然科学版）》2011年第4期470-474,共5页Journal of Hohai University(Natural Sciences)

基　　金：国家杰出青年科学基金(50725723)

摘　　要：为了提高电力负荷模型辨识的准确度,需要对实际扰动数据进行去噪处理.为此,针对有效值动态数据,提出了将平稳阶段噪声过滤掉而暂态阶段数据不加处理的去噪策略,将去噪后的平稳阶段数据与暂态阶段数据合成获得辨识用数据;给出了平稳阶段起始点的判据,提出了基于小波分析的平稳阶段数据去噪方法.实际应用结果表明,将处理过的扰动信号用于负荷模型参数辨识,所得参数波动更小、方差更小.To improve the identification accuracy of electric load modeling,the disturbance data need to be de-noised.A strategy of filtering the noise of the stable stage while not filtering the noise of the transient stage was proposed.The data for model identification was obtained by synthesizing the de-noised data of the stable stage and the data of the transient stage.The position of the starting point was provided.A wavelet-based method was proposed for filtering the noise in the stable stage.The practical applications show that the load model parameters estimated with the de-noised data are stabler and the variances are smaller.

关 键 词：电力负荷模型 噪声 小波分析 辨识 

分 类 号：TM743[电气工程—电力系统及自动化]