Share content on LinkedIn Share content on YouTube

John Deere’s Will Chemin, tech lead for welding and fabrication for the firms’ Construction & Forestry Division, notes that Deere has been using touch sensing for part ID for several years. What is new is the use of the part-ID data within the MRP system. This prevents errors when reporting production, and automatically updates inventory figures through the MES system. The reporting data also route through to inventory and relieve the child materials present in each weldment. The automated updating of inventory then triggers automated purchasing of outsourced components or drives production orders for parts and assemblies manufactured inhouse.

All of this adds up to improved consistency from the firm’s robotic-welding cells, reduced rework and improved inventory accuracy. “Operator productivity is up,” Chemin adds, “as our operators now perform more value-added work, rather than spending nonvalue-added time on completing reports.”

Accurate AI Weld-Defect Detection

In another application of AI for welding operations, a writeup on the website of industrial PC and automation-device distributor Eforel introduces Taiwanese manufacturer Tron-E, which fabricates electric-bus chassis and battery packs. Tron-E has developed an AI-based system to manage weld inspection during bus-frame fabrication and alleviate bottlenecks. 

The system is based on AI visual recognition of weld appearance and weld-bead sizes. By collecting photos of different defects, classifying them and then feeding them to algorithms, the trained AI model is able to notify personnel when it discovers substandard weld beads. To photograph welded vehicle frames and manage weld inspection, the firm developed a mobile cart equipped with three cameras, an AI-hardware platform and a display screen, powered by lithium batteries. The mobile cart, with top, middle and bottom cameras, will circle the bus frame once to collect pictures of the welding from all angles for identification and comparison. This ensures no defects escape the sharp “eyes” of the AI system. If it discovers substandard welds, repairs are made and the welds are reinspected, until all weld beads meet requirements. The resulting final inspection file is then saved and uploaded to the server as part of the electronic production history.

Weld inspection time has been reduced by 80%, say company officials, to inspect one large bus with hundreds of welds.

The AI platform in use: an Advantech MIC-730AI industrial computer with an onboard AI inference system based on NVIDIA Jetson Xavier (a family of embedded modules and computers designed for AI and high-performance computing).

In addition to the aforementioned AI system in Tron-E factories, the firm also plans to use the Advantech MIC-730AI to develop other AI inspection systems. MF

Industry-Related Terms: Circle, Core, Edge, Fixture, LASER, Model
View Glossary of Metalforming Terms

Technologies: Welding and Joining

Comments

Must be logged in to post a comment.
There are no comments posted.

Subscribe to the Newsletter

Start receiving newsletters.