An online stress monitoring strategy based on Wigner–Ville time–frequency energy extraction of single-frequency dual mode Lamb waves

Abstract

In this paper, a new online uniaxial stress monitoring methodology has been developed. Based on the acoustoelastic effect of Lamb waves, multimodal information at a single frequency, in which the approximate linear fitting between group velocity ratio of two modes and stress is obtained, is fused to characterize stress. The transceiver distance can be eliminated, which is essential knowledge for existing methods, in stress coefficient calibration and practical application by using the ratio of acoustic time. The semi-analytical finite element method(SAFE) is adopted to calibrate the stress coefficient. Next, the time–energy assignment algorithm derived from the smoothed pseudo-Wigner–Ville distribution(SPWVD) is evaluated on the ideal dataset obtained by time-domain finite element(TDFE) simulation. The arrival times of different modes are extracted accurately, and the fitting results are consistent with the theoretical analysis. Finally, the laboratory experiments employing a pair of piezoelectric wafer active sensors(PWAS) are conducted to calibrate unknown coefficients. The same calibration process is carried out at three different distances, and almost the same calibration results are obtained, indicating the distance-dependent nature of the stress coefficients. According to the completed calibration, the actual measurements from the same batch show the maximum absolute error of stress measurement is 4.2MPa within the applied stress range from 20MPa to 100MPa and the standard deviation is less than 3MPa, demonstrating good measurement accuracy and consistency. It shows the great potential of the proposed strategy to measure stress at unknown propagation distances.