结合SLMSST和DO提取时变结构瞬时频率  

Instantaneous Frequency Identification of Time-varying Structures by Combing Second-order Local Maximum Synchrosqueezing Transform and Dynamic Optimization

作　　者：李宇祖 刘景良 苏杰龙[1] 吕毓霖 LI Yuzu;LIU Jingliang;SU Jielong;LYU Yulin(School of Transportation and Civil Engineering,Fujian Agriculture and Forestry University,Fuzhou 350108,China;″Digital Fujian″Laboratory of Internet Things for Intelligent Transportation Technology,Fujian Agriculture and Forestry University,Fuzhou 350108,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]福建农林大学“数字福建”智能交通技术物联网实验室,福州350108

出　　处：《噪声与振动控制》2024年第2期50-56,62,共8页Noise and Vibration Control

基　　金：国家自然科学青年基金资助项目(51608122);福建省自然科学基金面上资助项目(2020J01581,2020J01579);福建农林大学科技创新专项基金资助项目(CXZX2020112A);福建农林大学优秀硕士学位论文资助基金资助项目(113/1122YS01005)。

摘　　要：为提升局部最大同步挤压变换估算瞬时频率的精度,本文结合2阶局部最大同步挤压变换(Second-order Local Maximum Synchrosqueezing Transform,SLMSST)和动态规划(Dynamic Optimization,DO)方法提出一种识别时变结构瞬时频率的新方法。该方法首先通过引入2阶瞬时振幅与相位得到精度更高的2阶瞬时频率估算位置。其次,搜索频率方向上时频系数的局部最大值所对应的2阶瞬时频率位置并根据这些位置对时频系数进行重排,从而得到2阶局部最大同步挤压变换后的瞬时频带。再次,运用动态规划法在限定频带范围内提取瞬时频率曲线。通过一组数值算例和一个时变拉索试验验证了所提新方法的有效性,研究结果表明:相比既有的局部最大同步挤压变换算法,2阶局部最大同步挤压变换和动态规划的联合算法不仅具有较好的精度,而且具有更好的时频聚集性。In order to improve the accuracy of instantaneous frequency identified by local maximum synchrosqueezing transform,a new algorithm is proposed by combining second order local maximum synchrosqueezing transform(SLMSST)and dynamic optimization(DO).In this method,the second order phase and amplitude are introduced at first to obtain more accurate second order instantaneous frequency positions.Then,the local maximum values of the time-frequency coefficients are searched in the frequency direction and the corresponding positions of the second order frequencies are found.According to these positions,the time-frequency coefficients are reassigned to obtain the instantaneous frequency band by the SLMSST.Finally,the dynamic optimization method is applied to extract the instantaneous frequency curve inside the frequency band.A numerical case and a test on a time-varying cable are investigated to verify the effectiveness of the proposed method.The results demonstrate that the proposed method behaves better than local maximum synchrosqueezing transform.It not only enhances the accuracy of instantaneous frequency identification,but also has better time-frequency energy concentration.

关 键 词：振动与波 时变 局部最大同步挤压变换 时频系数 瞬时频率 动态规划 

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]