High-resolution Lamb waves dispersion curves estimation and elastic property inversion

Area of Science:

• Materials Science
• Non-Destructive Testing
• Ultrasonic Wave Propagation

Background:

• Ultrasonic Lamb waves are crucial for non-destructive evaluation, but accurately inverting signals to material properties remains challenging due to complex multimodal dispersion and parameter estimation.
• Existing methods struggle with precise signal processing and parameter retrieval for accurate material characterization.

Purpose of the Study:

• To develop a robust strategy for high-resolution extraction of multimodal dispersion curves from guided ultrasonic signals.
• To implement a model-based approach for accurate estimation of elastic properties in plates using advanced signal processing and optimization techniques.

Main Methods:

• Estimation of Signal Parameters via Rotation Invariant Technique (ESPRIT) for high-resolution dispersion curve extraction.
• Particle Swarm Optimization (PSO) algorithm to retrieve optimal model parameters by maximizing an objective function derived from dispersion equations.
• Determination of elastic constants for isotropic and transversely isotropic plates.

Main Results:

• Accurate estimation of elastic properties for steel, aluminum, and composite plates, showing agreement with reference values.
• Low Root Mean Squared Errors (RMSEs) for dispersion curves (0.027-0.102 rad/mm) and less than 1% error in thickness estimation.
• Demonstrated advantages including a sharp objective function and simplified parameter tuning for inverse problems.

Conclusions:

• The proposed model-based inversion strategy effectively addresses challenges in ultrasonic guided wave analysis.
• High-resolution dispersion curve extraction and PSO optimization enable accurate and efficient material property determination.
• This method enhances the non-destructive evaluation capabilities for various plate materials.