Yu-Jun Xia;Lang Zhou;Yan Shen;Diana M. Wegner;Amberlee S. Haselhuhn;Yong-Bing Li;Blair E. Carlson
Realtime quality inspection of resistance spot welding (RSW) is a promising way to actualize intelligent manufacturing of automotive body structures. Apart from nugget diameter, weld penetration is also an important quality index for weld quality assessment. To that point, a novel approach to nondestructive measurement of weld penetration for the robotic RSW process is presented. In-depth research on the electrode displacement signal is carried out based on mechanism analysis and feature extraction. The effectiveness of the signal feature is tested under not only standard but also abnormal manufacturing conditions to investigate the generalization ability. Finally, an analytical model for predicting weld penetration is established via physical modeling to achieve online measurement capability. The error of weld penetration estimation could be limited to ±9.6% (±2σ) of the sheet thickness, and the measurement accuracy is robust to varying welding parameters and abnormal manufacturing conditions.