Multimodal lamb wave identification using combination of instantaneous frequency with EMD.

Lamb wave is guided plate wave that has potential in industrial inspection for non-destructive evaluation. However, the multimodal nature of Lamb wave obstructs the extraction and interpretation of information in signal. This work proposes a Lamb wave mode identification method combining instantaneous frequency (IF) with empirical mode decomposition (EMD). Firstly, EMD is implemented on multimodal signal and separates it into individual mode functions. Then the individual functions are analyzed by high-resolution time-frequency distribution. Finally, IF curve is obtained in the time-frequency domain and corresponding time information is used to calculate wave velocity. Lamb wave mode is determined by comparing the wave velocity with theoretical value. To verify the effectiveness of the proposed method, simulation presents the situation of boundary reflection and experiment presents the mode conversion by defect. Both simulation and experiment demonstrate the good performance of the mode identification method