An ultrasonic inspection method for diffusion-welded components with variable curvature and thickness

Automated ultrasonic inspection of diffusion-welded components with variable curvature and thickness is challenging due to difficulties in focusing acoustic waves on curved weld interfaces and the small size of defects, especially when accurate CAD models are unavailable. This study aims to develop a methodology for reliable inspection of such components. An integrated system combining laser profile scanning and multi-axis robotic ultrasonic testing is developed, synchronization among surface conformal data (surface-fitting information describing the geometry of the scanned surface), ultrasonic signals, and robot motion states is resolved in the proposed system design. Adaptive ultrasonic scanning trajectories are generated based on laser profiles to accommodate unknown curved surfaces. Ultrasonic field simulations determine suitable transducer parameters, and the effects of surface curvature and scanning trajectory accuracy on defect echoes are analyzed, ensuring the reliability of inspection results. Multi-axis robotic C-scan experiments are conducted on titanium alloy diffusion-welded curved blade specimens. The proposed methodology enables automated inspection of complex curved diffusion welds. Experimental results show effective detection of internal defects, with improved inspection stability and reliability. This study introduces an integrated laser-ultrasonic inspection approach that explicitly addresses the synchronization of multi-source data and enables adaptive scanning on unknown curved surfaces, providing a reliable approach for ultrasonic inspection of complex curved diffusion-welded components.