Session: 03-11-01: Future of Smart Manufacturing

Paper Number: 145018

145018 - Automated Vision-Based Laser Weld and Braze Quality Systems 

With the widespread use of laser welding and brazing in automotive manufacturing, ensuring the quality of these joints has become increasingly crucial. However, the current reliance on human operators for evaluating weld and braze quality is subjective, labor-intensive, and inadequate for high-volume production demands. Implementing machine vision technology for laser weld quality inspection faces significant challenges due to factors such as part surface reflectivity, discolorations, background light pollution, and variations in surface morphology. This presentation shares insights gained from developing machine vision-based quality inspection systems specifically designed for laser weld and braze applications. These systems incorporate automatic 3D weld data acquisition, data analysis algorithms, and a user-friendly graphical interface to enable fast and accurate detection and evaluation of welds according to industrial standards. Additionally, the presentation will discuss our recent efforts, including the use of graph neural networks (GNN) for laser weld identification in 3D point cloud images and the integration of process signals and post-process machine vision to predict busbar laser weld quality in battery module manufacturing.

Presenting Author: Guangze Li GM

Presenting Author Biography: Guangze is a dedicated researcher working in General Motors R&D, specializing in machine vision-based quality inspection for laser weld and braze applications. With a strong background in Materials Science and Engineering, Guangze obtained his Ph.D. degree from The Ohio State University. Prior to that, he completed his master's and bachelor's degrees at Shanghai Jiao Tong University. With his expertise in quality inspection and his passion for advancing automotive technology, Guangze is committed to driving innovation and ensuring the highest standards of quality in the automotive industry.

Authors:

Guangze Li GM
Hui-Ping Wang GM
Blair Carlson GM
Jorge Arinez GM