中文摘要：

针对超声导波检测应用中模态识别难的问题,提出一种基于导波频散特征的模态识别方法。通过估计导波信号的频散量来识别导波模态,而导波的频散量与该模态导波波数在激励频率处的二次泰勒展开系数有关。根据导波信号的频散特点,构造带时间斜变的Chirplet匹配原子库,基于该Chirplet原子库,对导波信号进行匹配追踪分解,并算得该导波信号波数在激励频率处二次泰勒展开系数的最优估计值,再根据预先算得的各模态导波的波数-频率频散关系,区分导波信号的不同模态。数值模拟和试验验证都表明,该导波模态识别方法是准确且有效的。而该方法的不足之处是尚不能识别混叠严重的导波信号。研究结果有助于提高人们对复杂导波检测信号的解析能力,并推动导波检测技术的推广应用。

英文摘要：

Mode identification of ultrasonic guided waves, which is always encountered in guided wave inspections, is a difficult problem. A dispersion-based mode identification approach is introduced, which distinguishes guided wave modes by their degree of dispersion. The degree of dispersion of some mode is found related with the second coefficient of Taylor expansion of this mode＇s wave number at the excitation frequency. Based on this dispersion feature of guided waves, a Chirplet dictionary with the time shear is designed. And then the guided wave signals are analyzed with the matching pursuit method based on this Chirplet dictionary. The optimal estimation of the second coefficient of Taylor expansion of wave number can be obtained by the matching pursuit method, and combined with the wavenumber-frequency dispersion relation calculated in advance, the guided wave modes can be recognized. The numerical simulation and experimental verification all show that the proposed mode identification method is effectively and correctly. The deficiency of this method is that it still cannot distinguish the severely overlapped guided wave modes. The study results help to improve the interpretation of complicated guided wave inspection signals, and generalize the application of guided wave inspection.