AI Enhanced Robotic Welding System for Multi-Layer-Multi-Pass CJP Groove Welds
Chung-Che Chou Lab. (NTU & NCREE, NIAR)
- 2027 Honoree in Construction & Industrial Tech
An autonomous AI-driven robotic complete joint penetration (CJP) welding platform designed for heavy structural steel fabrication. Integrating multi-axis robotics, digital self-adaptive control, and 3D laser profilometry, the system employs an AI framework to dynamically optimize multi-pass welding parameters. It autonomously compensates for real-world joint assembly tolerances in steel box columns, producing code-compliant (AWS D1.1:2020, AWS D1.8:2021), seismic-grade complete joint penetration welds verified by non-destructive ultrasonic testing.
Complementing the hardware, a vendor-neutral, cloud-based AI expert platform delivers automated parameter planning and process optimization across heterogeneous robotic manipulators, diverse power sources, and multi-brand automated welding systems, ensuring seamless cross-platform interoperability.