An Ultrasonic Signal Processing Technique for Extraction of Arrival Time from Lamb Waveforms

Abstract

Travel time Lamb wave tomography has been shown to be an effective nondestructive evaluation (NDE) technique for plate-like structures. The methods used previously to extract the arrival times of the fastest or multi-mode Lamb wave modes are mostly based on enveloping techniques, and various time-frequency methods such as Wigner-Ville distributions, short-time Fourier transforms, and recently explored wavelet transform (WT). Frankly speaking, the uses of signal processing methods improve the accuracy of the arrival time extraction to a great extent relative to interpret the Lamb waveforms directly. Hilbert-Huang transform (HHT) is also an efficient way for analyzing and processing non-stationary signals. The resolving power of time and frequency is restricted from Heisenberg principle in wavelet analysis, while in HHT time resolving power is precise and steady, and frequency resolving power is adaptive according to signal intrinsic characteristics. Conclusion can be made that the HHT method is more adaptive than WT analysis in analyzing non-stationary signals. Based on the above analysis, HHT is attempted to extract multi-mode arrival times from Lamb waveforms in this paper. Comparison with the wavelet transform shows the new method offers higher temporal resolutions in extracting arrival times of multiple Lamb wave modes.