基于冗余字典稀疏分解的电能质量扰动信号压缩采样研究  

Compressed sampling of power quality disturbance signal based on sparse decomposition of redundant dictionary

作　　者：欧阳华[1] 邵英[1] 李辉[1] 侯新国[1] OUYANG Hua;SHAO Ying;LI Hui;HOU Xin-guo(College of Electrical Engineering,Naval Univ.of Engineering,Wuhan 430033,China)

机构地区：[1]海军工程大学电气工程学院,武汉430033

出　　处：《海军工程大学学报》2021年第3期14-19,共6页Journal of Naval University of Engineering

基　　金：国家自然科学基金青年科学基金资助项目(51307178)。

摘　　要：针对电能质量监测系统庞大的数据储存和传输问题,采用压缩感知技术实现了电能质量扰动信号压缩采样和非线性恢复。傅里叶变换和小波变换联合构成的冗余字典作为稀疏分解矩阵,实现了扰动信号的稀疏表示。数值仿真表明:与傅里叶变换字典或小波变换字典等单一正交基作为稀疏分解矩阵相比,采用该冗余字典能够更好地匹配多种模式混合的扰动信号,验证了基于冗余字典的电能质量扰动信号压缩采样的可行性;在相同的压缩比下,压缩感知能够取得与传统的小波变换和离散傅里叶变换变换阈值压缩相比拟的压缩性能。In view of the huge data storage and transmission problem of power quality monitoring system,compressed sampling and non-linear reconstruction of the power quality disturbance signal are realized by the compressed sensing(CS)theory.The redundant dictionary composed of Fourier transform and wavelet transform is used as the sparse decomposition matrix to realize the sparse distur-bance signals.Compared with a single orthogonal basis such as a Fourier transform dictionary or a wavelet transform dictionary as a sparse decomposition matrix,a redundant dictionary composed of a Fourier transform dictionary and a wavelet transform dictionary can better match the disturbance with multiple patterns signal.The numerical simulation proves the feasibility of compressed sampling of power quality disturbance signal based on redundant dictionary.The acceptable SNR of recovered signals can be achieved using CS method in the redundant dictionary,comparable to conventional WT and DFT transform compression by certain thresholds.

关 键 词：压缩感知 电能质量 稀疏表示 冗余字典 

分 类 号：TM933[电气工程—电力电子与电力传动] TP274[自动化与计算机技术—检测技术与自动化装置]