多谐波互信息特征的自适应Prony提取方法  

Multi-harmonic adaptive Prony method in mutual information feature extraction

作　　者：夏向阳[1] 程莎莎[1] 帅词俊[2] 徐林菊[1] 王欢[1] 周云[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410004 [2]中南大学机电工程学院,湖南长沙410083

出　　处：《中南大学学报（自然科学版）》2013年第10期4081-4087,共7页Journal of Central South University:Science and Technology

基　　金：广东省科技计划项目重大科技专项(2010A080804023);湖南省自然科学基金资助项目(2011JJ5027);湖南省科技计划项目(2011GK3118);衡阳市科技计划项目(2012KG46)

摘　　要：针对电力系统中多谐波源相互影响的谐波具有非持续性和突变性的特征,传统的prony算法具有误差较大、精度不高的缺点以及谐波信号采样值具有高维数据的特性,提出自适应Prony方法提取多谐波相互影响的信号互信息特征,通过方差修正量调整互信息提取的精度;对采样数据矩阵进行奇异值分解,秩后的矩阵中的元素值即为多谐波源互信息提取的数据。研究结果表明:该方法实现了多谐波源互信息特征的提取和谐波频谱的分析且精度较高,验证了该方法的正确性。Considering that the harmonic waves with multi-harmonic source power system are non-persistent and mutant, traditional Prony algorithm has high error, and harmonic signal samples has high-dimensional datacharacteristic, adaptive Prony method （Prony method PM） was proposed to extract information feature with multi-harmonic interaction signal, the accuracy of the information was extracted by the variance of the amount of correction, sampling data matrix singular value was decomposed, the data was multi-harmonic source mutual information extracted data. The results show that the method realizes the extraction of the harmonic source mutual information feature and harmonic spectrum analysis,and the analysis accuracy is high, which verifies the correctness of the method.

关 键 词：多谐波源 自适应Prony方法 方差修正量 互信息特征提取 奇异值矩阵 

分 类 号：TM727[电气工程—电力系统及自动化]