Successful Enterprise Digital Transformation Cases

The automotive body-in-white is assembled from hundreds of stamped parts via spot welding. Each workpiece has dimensional deviations, and welding robots need to accurately perceive the workpiece pose. This project provides a complete 3D vision-guided upgrade for the body-in-white welding lines at a joint venture OEM's Tianjin plant, covering 36 stations across 4 welding lines. Since commissioning, the system has been operating stably, processing over 15,000 vehicles per month.

The project was officially commissioned in 2022 and is one of the largest-scale body-in-white vision-guided deployments in China. By deep integration with the plant's MES system, it achieves a fully automated closed loop from workpiece arrival, vision inspection, path compensation, to welding execution, completely replacing the original semi-automated mode of manual visual inspection and adjustment.

The body-in-white surface has many reflective areas, and welding generates high-temperature smoke. Traditional 2D vision positioning has insufficient accuracy and poor stability. The customer required the system to operate stably in harsh environments with temperatures above 50°C and smoke visibility below 0.5m, with single-station positioning accuracy better than ±0.15mm, cycle time not exceeding 3 seconds, and seamless integration with the existing KUKA robot control system without modifying the original PLC logic.

We selected the Laser L V3 laser line scan sensor, which, with its strong resistance to ambient light interference, can obtain high-quality point cloud data even in welding smoke and strong reflection environments. Combined with our proprietary high-temperature filtering and smoke compensation algorithms, thermal radiation noise is effectively suppressed. The robot controller receives pose compensation values in real time via a standard TCP/IP interface, and the entire guidance closed loop is completed within 2.1 seconds, meeting the cycle time requirements. The sensor adopts a modular installation design, with maintenance and replacement time <30 minutes, without affecting the main line cycle time.

Positioning accuracy reached ±0.08mm, nearly double the customer's requirement, providing ample margin for weld quality. Welding defect rate decreased by 30%, scrap rate dropped from 2.1% to 1.47%, saving approximately 350,000 CNY per month in material costs. System availability of 99.7%, with no major failures within 24 months of operation, outperforming industry averages.