摘要
为解决低频振荡在电力系统中实测信号存在噪声干扰、信号处理过程中模态混叠及非线性问题,提出了基于小波无偏风险估计阈值消噪和变分模态分解-希尔伯特黄变化(VMD-HHT)的低频振荡分析的方法。首先,对于含噪的实测信号,采用小波无偏风险估计阈值进行消噪的预处理;其次,通过使用样本熵来确定VMD的二次惩罚因子、使用频谱图来确定分解层数,预处理后的信号经过VMD分解得到IMF(Intrinsic Mode Function)分量;最后,对得到的IMF分量进行希尔伯特黄变换得到模态的参数。通过复合信号测试和IEEE(Institute of Electrical and Electronics Engineers)四机二区域仿真的辨识结果,验证了所提方法的合理性和有效性。同时与TLS-ESPRIT算法、Prony算法和经验模态分解的结果分析对比可知,所提方法在辨识方面更为准确。
In order to solve the problems of noise interference of the measured signal in the power system of low-frequency oscillation,modal aliasing and nonlinearity during signal processing,a low-frequency oscillation analysis has been proposed based on wavelet unbiased risk estimation threshold noise reduction and variational mode decomposition and Hilbert-Huang transform(VMD-HHT).Firstly,in regard to the measured signal containing noise,the wavelet unbiased risk estimation threshold is used for noise cancellation pretreatment;secondly,the secondary penalty factor and spectrogram of VMD are determined by using sample entropy to determine the number of decomposition layer,and the IMF component is obtained after decomposing the preprocessed signal with VMD;finally,the parameters of the modality are available after the Hilbert-Huang transform is performed on the obtained IMF(Intrinsic Mode Function)components.Through the identification results of the composite signal test and the IEEE(Institute of Electrical and Electronics Engineers)quad-machine two-region simulation,the rationality and effectiveness of the proposed method are verified.Meantime,compared with the TLS-ESPRIT algorithm,Prony algorithm and empirical modal decomposition result analyses,it is found that the proposed method is more accurate in terms of identification.
作者
付江涛
杨毅强
宋弘
FU Jiangtao;YANG Yiqiang;SONG Hong(School of Automation and Information Engineering,Sichuan University of Science&Engineering,Yibin 644000,China;Artificial Intelligence Key Laboratory of Sichuan Province,Yibin 644000,China;College of Electronic Information and Automation,Aba Teachers University,Aba 623002,China)
出处
《四川轻化工大学学报（自然科学版）》
CAS
2023年第3期76-85,共10页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)
基金
四川省科技厅项目(2020YFG0178,2021YFG0313)
人工智能四川省重点实验室项目(2019RYY01)。
关键词
低频振荡
小波阈值消噪
样本熵
变分模态分解-希尔伯特黄变化
low-frequency oscillation
wavelet threshold denoising
sample entropy
variational mode decomposition and Hilbert-Huang transform
