同步压缩短时傅里叶变换和时频脊线增强识别结构固有频率

Structural natural frequency recognition based on synchrosqueezing short-time Fourier transform and time�frequency ridge enhancement

根据实测结构振动响应准确地提取结构模态特征是评估结构健康状态的关键一环。将同步压缩短时傅里叶变换和时频脊线增强方法相结合,提出了一种新的结构自振频率识别方法。该方法首先对振动响应信号进行了同步压缩短时傅里叶变换,锐化了信号的时频表达,然后采用脊线增强方法对锐化后的时频图提取了峰值序列,获得了结构最可能的能量波动路径。通过对包含间断信号和模态混叠信号的仿真信号以及包含不同信噪比噪声的仿真信号进行验证,证明了同步压缩短时傅里叶变换的可靠性和准确性。接着,通过对比一个4自由度系统在地震作用下的结构响应,分析了不同变换方法的脊线增强效果,进一步验证了脊线增强方法对理解结构能量耗散路径和识别结构自振频率的有效性。

Accurately extracting the modal characteristics of structures from the measured vibration response is a key step to evaluate the health status of structures.A new method for natural frequency identification of structures was proposed by combining synchrosqueezing short-time Fourier transform and time-frequency ridge enhancement.Firstly,synchrosqueezing short-time Fourier transform was used for the vibration response signal to sharpen the time-frequency expression of the signal.Then,the peak sequence was extracted from the time-frequency diagram by ridge enhancement method to obtain the most likely energy fluctuation path of the structure.The reliability and accuracy of synchrosqueezing short-time Fourier transform were verified by simulation signals including discontinuous signals,modal aliasing signals and simulation signals including different signal-to-noise noises.Then,by comparing the structural response of a four degrees of freedom system under seismic load,the effectiveness of the ridge enhancement method for understanding the energy dissipation path of the structure and identifying the natural vibration frequency of the structure was verified.