基于压缩感知的含扰动电能质量信号压缩重构方法

Method Based on Compressed Sensing for Compression and Reconstruction of Power Quality Signals with Disturbances

针对电能质量信号的压缩重构问题,提出了一种应用压缩感知理论对电能质量信号进行压缩采样和非线性恢复的方法。首先对含有扰动的电能质量信号的稀疏性和可压缩性进行了分析,并对现有的典型贪婪恢复算法的特性进行了研究,然后结合有代表性的SP和SAMP两种算法的优点,提出了一种改进的BSMP算法。应用包括BSMP在内的几种贪婪算法对含有谐波、间谐波、电压暂升和暂降等稳态和暂态扰动的电能质量信号的压缩重构性能进行仿真分析,仿真结果说明了BSMP算法恢复混合或单一扰动的电能质量信号的可行性。与现有贪婪算法相比,BSMP无需稀疏度先验,可以用较快的速度和更高的压缩比以100%的概率实现成功重构。

A method based on compressed sensing theory is proposed to realize the compressive sampling and nonlinear recovery. The compressibility and the sparsity of the disturbed power quality signals are analyzed, and the existing iterative greedy pursuit recovery algorithms are reviewed. Combined with the advantages of two typical algorithms SP and SAMP, an improved method named BSMP is proposed. Then the reconstruction performance of several typical greedy pursuit recovery algorithms are simulated and analyzed, using the power quality signals with steady or transient disturbances such as harmonics, interharmonics, voltage swell and sag, etc. Results verify the feasibility of the BSMP in recovering the power quality signals with mixed or single disturbances. Compared with the existing greedy algorithms, BSMP algorithm does not require sparsity level as prior information, and can realize the successful reconstruction absolutely with faster speed and higher compression ratio.