Fatigue crack monitoring in L-shaped structures using AHBMP and enhanced RAPID imaging

Fatigue cracking in L-shaped aerospace components is a major safety concern. Ultrasonic guided waves can monitor this, but the complex geometry causes multimodal conversion and scattering, which creates noise and hides damage signals. We propose a visualization method combining the Adaptive Hybrid Backward Matching Pursuit (AHBMP) algorithm with an enhanced Reconstruction Algorithm for Probabilistic Inspection of Damage (RAPID) framework. We use AHBMP to decouple overlapping signals and isolate the A0 mode, as it is highly sensitive to cracks. This data is then processed by an enhanced RAPID algorithm that introduces a third focal point to refine the probability distribution. This approach significantly betters damage localization and resolution. Experiments show the method tracks the full fatigue lifecycle, from initiation to propagation. In the stable propagation phase, crack length estimation error remained below 5%, proving its practical engineering value.