Model-based imaging of damage with Lamb waves via sparse reconstruction.

Area of Science:

• Materials Science
• Mechanical Engineering
• Signal Processing

Background:

• Ultrasonic guided waves are increasingly used for structural health monitoring (SHM) and nondestructive evaluation (NDE) of plate-like structures.
• Current methods often involve comparing baseline signals to detect changes, but struggle with multiple damage sites.

Purpose of the Study:

• To develop a novel sparse imaging technique for SHM that leverages the assumption of mostly damage-free structures.
• To improve the detection and localization of structural damage, especially in scenarios with multiple damage sites and noisy data.

Main Methods:

• Utilized sparse signal decomposition of differential ultrasonic guided wave signals.
• Applied sparse reconstruction techniques including basis pursuit denoising and orthogonal matching pursuit.
• Evaluated a hybrid reconstruction method and tested with simulated and experimental data on an aluminum plate.

Main Results:

• Demonstrated the efficacy of sparse reconstruction methods in identifying damage locations accurately.
• Achieved very sparse indications of damage even with model mismatch and significant noise.
• Successfully localized artificial damage on an aluminum plate using the proposed techniques.

Conclusions:

• The proposed sparse imaging approach offers a promising alternative for SHM and NDE of plate-like structures.
• This method is robust to noise and model inaccuracies, outperforming traditional techniques in complex scenarios.
• The technique effectively exploits the sparsity of structural damage for improved flaw detection.