摘要
针对电能质量复合扰动特征量识别难的问题,通过分析复合扰动特点,提出一种基于奇异值分解的多级SVD特征提取方法。在单级SVD分解原理的基础上,利用2行Hankel矩阵实现SVD多层次递推分解,以多分辨率来展现信号不同空间的近似和细节信号。给出多级SVD的分解和重构算法及数学框架。研究结果表明,在对含有多种复合扰动的电能质量信号处理中,经多级SVD分解及相应幅值谱分析后,清晰提取到各种扰动的特征信息,相应地与单级SVD和小波变换进行比较,证明多级SVD在电能质量复合扰动信号处理领域中具有应用前景。
Aiming at the feature extraction of power quality mixed disturbances,a feature extraction method based on multi?level singular value decomposition(SVD)is proposed to solve the difficulty. Based on the decomposition principle of single?lev?el SVD and two?line Hankel matrix,SVD multi?level recursion decomposition was realized to reveal a series of approximation and detail signals with different resolution. The decomposition and reconstruction algorithm of multi?level SVD is given,and themathematical frameworks of this method is provided. The results show that the multi?level SVD can extract the accurate charac?teristics of power quality mixed disturbances. The comparative study carried out with single?level SVD and wavelet transform demonstrates that multi-level SVD has good application prospect in signal processing of power quality mixed disturbances.
作者
刘嫣
汤伟
LIU Yan;TANG Wei(College of Electrical and Information Engineering,Shaanxi University of Science & Technology,Xi’an 710021,China)
出处
《现代电子技术》
北大核心
2017年第4期32-35,共4页
Modern Electronics Technique
基金
陕西省教育厅专项科研计划项目(14JK1083)
关键词
电能质量
复合扰动
多级SVD
特征提取
power quality
mixed disturbance
multi.level SVD
feature extraction
