Blind Extraction of Spot Weld Features for Real-Time Resistance Spot Weld Process Monitoring

Ultrasound-based spot weld monitoring systems primarily focus on extracting key characteristics needed to assess the quality of spot weld nuggets and monitor the spot weld process. However, factors such as intense temperature variation, electromagnetic noise, mechanical vibrations from servo motors and moving parts, and air bubbles in the coolant can significantly degrade the quality of the received A-scans. Consequently, the efficiency of real-time ultrasonic inspection systems can be severely compromised. This paper introduces a new method for effectively extracting desired features to recover A-scan signals from substantial noise interference. The effectiveness of the proposed method is demonstrated through experiments using real data collected by commercial spot weld analyzers under different environmental conditions.