基于滑动窗宽优化的LMSSGST识别非平稳信号瞬时频率  

Recognition of non-stationary signal instantaneous frequency with LMSSGST based on sliding window width optimization

作　　者：刘景良 苏杰龙 戴逸宸 李宇祖 黄永[2] 郑文婷 LIU Jingliang;SU Jielong;DAI Yichen;LI Yuzu;HUANG Yong;ZHENG Wenting(College of Transportation and Civil Engineering,Fujian Agriculture&Forestry University,Fuzhou 350108,China;College of Civil Engineering,Harbin Institute of Technology,Harbin 150090,China;College of Civil Engineering,Fujian University of Technology,Fuzhou 350118,China)

机构地区：[1]福建农林大学交通与土木工程学院,福州350108 [2]哈尔滨工业大学土木工程学院,哈尔滨150090 [3]福建理工大学土木工程学院,福州350118

出　　处：《振动与冲击》2024年第19期183-193,共11页Journal of Vibration and Shock

基　　金：国家自然科学基金青年项目(51608122);福建省自然科学基金面上项目(2020J01581);福建省住房和城乡建设行业科技研究开发计划项目(2022-K-266)。

摘　　要：为提高非平稳响应信号瞬时频率的识别效果,提出基于滑动窗宽优化的局部最大同步挤压广义S变换(local maximum synchrosqueezing generalized S-transform,LMSSGST)。该方法首先对非平稳响应信号进行广义S变换获得相应的时频系数;其次,利用该响应信号的功率谱密度特征曲线确定局部最大同步挤压算子中滑动窗的宽度;再次,通过局部最大同步挤压算子进行时频重排;最后,采用模极大值改进算法提取瞬时频率曲线。通过两个数值算例、一个滑动窗宽参数分析和一个时变拉索试验验证了所提方法的有效性,研究结果表明:利用功率谱密度特征曲线能够有效确定滑动窗的窗宽和模极大值算法的提取范围。相比局部最大同步挤压变换算法,基于滑动窗宽优化的LMSSGST具有更佳的瞬时频率识别效果。Here,to improve the recognition effect of instantaneous frequency of non-stationary response signals,a local maximum synchronous squeezing generalized S-transform(LMSSGST)based on sliding window width optimization was proposed.Firstly,this method could perform a generalized S-transform for non-stationary response signals to obtain corresponding time-frequency coefficients.Secondly,the power spectral density characteristic curve of response signals was used to determine sliding window width in LMSS operator.Once again,time-frequency rearrangement was performed with local maximum synchronous squeezing operator.Finally,the modulus maximum value improvement algorithm was used to extract the instantaneous frequency curve.The effectiveness of the proposed method was verified with two numerical examples,a sliding window width parametric analysis and a time-varying cable test.The study results showed that using power spectral density characteristic curve can effectively determine sliding window width and the extraction range of the modulus maximum value algorithm;compared to LMSST algorithm,LMSSGST based on sliding window width optimization has the better recognition effect on instantaneous frequency.

关 键 词：瞬时频率 广义S变换 局部最大同步挤压变换 时变结构 功率谱密度 

分 类 号：TN911.6[电子电信—通信与信息系统] TU311.3[电子电信—信息与通信工程]