SurfMon: a learning-based optical in-process surface monitoring system for topographical defects

Area of Science:

• Manufacturing Metrology
• Sustainable Manufacturing
• Optical Surface Monitoring

Background:

• Metrology is crucial for manufactured part quality, especially with increasing sustainable manufacturing demands.
• Early defect detection in manufacturing processes saves materials, energy, and costs, contributing to sustainability.
• Existing metrology systems may lack flexibility for in-process monitoring.

Purpose of the Study:

• To present SurfMon, a portable, in-process optical surface monitoring system.
• To demonstrate the system's capability in detecting topographical defects using laser speckle and machine learning.
• To evaluate SurfMon's effectiveness in real-world manufacturing environments.

Main Methods:

• SurfMon utilizes laser speckle pattern measurements and a machine learning algorithm (Principal Component Analysis).
• The system comprises a laser, camera, microcontroller, and optics, powered by a rechargeable battery.
• It is designed for in-process measurements, mountable on machine tools to monitor surfaces post-machining.

Main Results:

• The system successfully captured reflective laser speckle patterns from manufactured surfaces.
• A Principal Component Analysis model was trained to identify anomalies in surface topography.
• Demonstrated effectiveness on a CNC machine indicates promise for real-time defect detection.

Conclusions:

• SurfMon is a promising solution for in-process surface monitoring in manufacturing.
• The system's flexibility and learning capability support sustainable manufacturing practices.
• Early detection of topographical defects is achievable with this laser speckle-based approach.