John Deere enters smart manufacturing
28 April 2016
In a pilot project, agricultural machinery manufacturer John Deere is transitioning its production workflows to smart manufacturing with IBM.
Logistics processes control themselves; the Watson AI platform helps workers by giving precise instructions. It’s a model for the future of manufacturing.
Every one of the John Deere's huge tractors is made-to-measure. No customisation request is too outrageous. That is an immense challenge for manufacturing systems, and one that requires processes to be both perfectly attuned and highly flexible - no one can afford bottlenecks here. For the agricultural machinery expert, paving the way to digital factories was the obvious choice.
A pilot project based on IBM’s smart manufacturing platform has been underway in the largest John Deere plant in Mannheim, Germany since January. The demonstration at Hannover Fair showed how a tractor drivetrain is manufactured, starting from order picking. The components needed are displayed to the worker directly from the ERP system; the worker confirms the withdrawal and it is reported to the central platform in real time. An IT system ensures that the materials required are available at all times. Documentation is produced automatically, making the dream of a paperless factory a reality.
The thinking, talking computer
IBM and John Deere are currently working on integrating cognitive maintenance: If there ever is a manufacturing problem, IBM's Watson deep-learning platform comes into play. The worker takes a photograph of their workstation, and the artificial intelligence uses an image recognition algorithm to determine the cause of the fault. One voice command is all it takes for Watson to explain how to fix it. Need a specialist? The system checks the schedules and suggests the best time for maintenance, taking possible synergy effects into consideration at all times.
In the factory of the future, many processes can be automated fully. For example, production managers can use rules to define which processes are triggered in the event of deviations. This is critical for autonomously controlled production workflows that significantly reduce human workloads in factories.