Ultrasonic weld inspection using low-order lamb wave models and machine-learning based signal enhancement and inversion

Automated ultrasonic weld inspection remains a significant challenge in the nondestructive evaluation (NDE) community to factors such as limited training data (due to the complexity of curating experimental specimens or high-fidelity simulations) and environmental volatility of many industrial settings (resulting in the corruption of on-the-fly measurements). Thus, an end-to-end machine learning (ML) workflow for acoustic weld inspection in realistic (i.e., industrial) settings has remained an elusive goal. This work addresses the challenges of data curation and signal corruption by proposing workflow consisting of a reduced-order modeling scheme, diffusion based denoising scheme, and UNET-based segmentation and detection scheme. A reduced-order Helmholtz model based on Lamb wave theory is used to generate a comprehensive dataset that includes varying weld heterogeneity and crack defects, and a signal corruption scheme is introduced to emulate real-world volatility. Conditioned diffusion models transform corrupt signal back to simulation-grade measurements, which are passed to customized UNET modules tasked with identifying cracks and weld heterogeneity. The models are refined through a transfer learning stage using a limited set of full 3-D elastodynamic simulations. Finally, we validate the methodology using an experimental weld specimen. This integrated framework provides an end-to-end solution for automated weld inspection on real data.