Evaluation of Bonding Quality with Advanced Nondestructive Testing (NDT) and Data Fusion

Area of Science:

• Materials Science
• Mechanical Engineering
• Non-Destructive Testing

Background:

• Adhesive bonding is critical for composite structures.
• Evaluating bond quality non-destructively is challenging.
• Advanced NDT methods offer potential for improved assessment.

Purpose of the Study:

• To quantitatively compare NDT techniques for adhesive bond evaluation.
• To assess the effectiveness of data fusion for defect detection.
• To investigate different data fusion algorithms for NDT data.

Main Methods:

• Investigated composite-epoxy single-lap joints with artificial debonding defects.
• Employed advanced ultrasonic testing and induction thermography.
• Extracted feature matrices using interface reflection amplitude and PCA.
• Applied basic and statistical data fusion algorithms, including Dempster-Shafer rule.

Main Results:

• Quantitatively evaluated defect detection performance using receiver operating characteristics (ROC).
• Demonstrated that data fusion algorithms improve the detectability of debonding defects.
• Compared the performance of various fusion techniques.

Conclusions:

• Data fusion significantly increases the detectability of artificial debonding in single-lap joints.
• Combined NDT approaches provide a more robust evaluation of adhesive bond quality.
• The study provides a quantitative basis for selecting NDT and data fusion strategies.