An Improved RAPID Imaging Method of Defects in Composite Plate Based on Feature Identification by Machine Learning

Area of Science:

• Materials Science
• Non-destructive Testing
• Artificial Intelligence

Background:

• Lamb wave-based defect detection methods like RAPID (reconstruction algorithm for probabilistic inspection of defect) are crucial for locating flaws in composite plates.
• Accurate visualization of defect features (type, size, direction) remains a challenge for existing methods.

Purpose of the Study:

• To propose an improved RAPID imaging method enhanced with machine learning (ML) for precise defect localization and feature identification in composite plates.
• To refine the RAPID method's accuracy by integrating ML-driven analysis of defect characteristics.

Main Methods:

• Utilizing multiple machine learning models to analyze Lamb wave detection signals for identifying specific defect features.
• Modifying the RAPID method's scaling parameter (β) based on identified defect features.
• Assigning weights to critical detection paths correlated with defect characteristics to enhance imaging precision.

Main Results:

• The proposed ML-enhanced RAPID method successfully visualizes defect locations and associated features.
• Simulation results demonstrate an intuitive characterization of defect information.
• Significant improvement in the accuracy of defect imaging was achieved.

Conclusions:

• The integration of machine learning with the RAPID method offers a powerful approach for accurate defect characterization in composite materials.
• This enhanced technique provides more intuitive and precise defect imaging compared to traditional methods.
• The findings contribute to advancing non-destructive testing techniques for composite structures.