Treffer: Ultrasonic Rayleigh wave microscopic shape reconstruction of surface and near-surface defects using geometrical full waveform inversion.

Ultrasonic Rayleigh waves are widely used to detect surface/subsurface defects. However, complete shape reconstruction of defects with deep subwavelength details in both the lateral and depth directions remains challenging. This study introduces a level-set-based geometrical full waveform inversion (LSGFWI) method to achieve super-resolution defect reconstruction using surface-measured Rayleigh wave signals. The GFWI approach employs the level-set function to represent defect boundaries, enabling simultaneous reconstruction of multiple arbitrarily shaped surface/near-surface defects. To mitigate slow convergence caused by uneven Rayleigh wave insonification with depth, a memoryless Broyden–Fletcher–Goldfarb–Shanno (BFGS) method is used to approximates the inverse Hessian matrix, preconditioning the inversion process. All the wave physics, including the evanescent waves induced at the defect boundaries, are directly measured and exploited to reconstruct the defects. Numerical evaluations are conducted for various defect types, showing the shape reconstruction resolution to achieve tens of micrometres. In addition, investigations are conducted to study the effects of measurement set-ups (e.g., measurement interval and area) on shape inversion. Ultrasonic experiments are also conducted to reconstruct a near-surface void, further showing the feasibility of the method for practical use. This approach significantly enhances defect characterization capabilities, offering high-resolution defect geometric information critical for engineering applications and materials science. [ABSTRACT FROM AUTHOR]