Intelligent detection of cracks in metallic surfaces using a waveguide sensor loaded with metamaterial elements

Area of Science:

• Materials Science
• Artificial Intelligence
• Non-Destructive Testing

Background:

• Traditional crack detection relies on manual interpretation of microwave sensor signals, which can be subjective and labor-intensive.
• Sub-millimeter crack detection in metallic surfaces is critical for structural integrity and safety.

Purpose of the Study:

• To implement and evaluate an artificial intelligence model for automated detection of sub-millimeter cracks.
• To assess the impact of AI on sensing sensitivity, cost, and automation in crack detection.
• To demonstrate the feasibility of AI for post-processing microwave sensor data for handheld inspection equipment.

Main Methods:

• A real-life experiment was conducted using a waveguide sensor with metamaterial elements to collect data from a metallic plate with cracks.
• An artificial intelligence model was developed and implemented to classify surfaces as either cracked or non-cracked.
• The AI model processed data obtained from the microwave sensor.

Main Results:

• The implemented artificial intelligence model demonstrated good crack classification accuracy rates.
• AI integration showed significant improvements in sensing sensitivity, cost-effectiveness, and automation compared to traditional methods.
• The study validated the effectiveness of AI for post-processing microwave sensor data.

Conclusions:

• Artificial intelligence offers a powerful and efficient solution for automated sub-millimeter crack detection in metallic surfaces.
• AI-driven inspection systems have the potential to enhance the capabilities of handheld test equipment.
• The proposed AI method provides a reliable and accurate approach for non-destructive testing applications.