Since John Deere consumes millions of pounds of welding wire every year, it knows a thing or two about weld integrity. But by adding Intel's artificial intelligence technology and processing power with defect identification software in a recently concluded pilot project, John Deere has transformed its manufacturing line to leverage the industrial Internet of Things.
Quality, always the top priority for Deere, was improved even more as defects decreased in tandem. Greater reliability and longer mean time between repairs translates to more satisfied customers for the maker of agricultural and construction machinery.
"Welding is a complicated process, and this AI solution has the potential to help us produce our high-quality machines more efficiently than before," says Andy Benko, quality director, for John Deere Construction and Forestry Division. "The introduction of new technology into manufacturing is opening up new opportunities and changing the way we think about some processes that haven't changed in years."
In the recently completed pilot, Deere and Intel built an integrated system that included a neural network-based inference engine, which logs defects in real-time and automatically stops the welding process. Working with Intel processors and the Intel® Distribution of OpenVINO™ toolkit, the software accelerates defect identification and allows Deere to correct the issue in real-time.
"Weld quality is not unique to John Deere, it's an industry-wide challenge," says Christine Boles, VP, Internet of Things Group, and GM, Industrial Solutions Division at Intel. But what is unique about Deere's approach is how it has embraced AI and machine vision, which automates quality inspection and enables Deere to detect issues as they happen and respond in real-time.
The Nuts & Bolts
Like many big machinery manufacturers, Deere relies on Gas Metal Arc Welding (GMAW) for its mild- to high-strength steel used in its commercial machines and products at 52 factories around the world. Within each factory, hundreds of robotic arms swing into action daily, consuming millions of weld wire pounds each year.
Image credit: John Deere
That volume of welding means Deere has plenty of experience finding solutions to welding issues, as well as the common welding challenge of porosity. Cavities in the weld metal are created by trapped gas bubbles as the weld cools, which weaken the weld strength. Until now, GMAW defect detection has been a manual process requiring highly skilled technicians. Past attempts to deal with weld porosity haven't always been successful. If the flaws aren't found until later in the manufacturing process, they require re-work or even scrapping of full assemblies. That's expensive for Deere and can have a ripple effect across many processes and functions.
"The introduction of new technology into manufacturing is opening up new opportunities and changing the way we think about some processes that haven't changed in years," says Deere's Benko. Use of AI, machine learning, automation and Internet of Things connectivity are helping to make companies more innovative and to compete better--hallmarks of the Industry 4.0 trend.
Industry 4.0 Acceleration
"We're seeing Industry 4.0 adoption accelerating and this is one of the success stories that signals to the rest of the industry that the technology and transformation is viable, and has tangible, impactful results," Boles explains. And deploying solutions based on open architectures can help lower maintenance costs, increase productivity, and create new business opportunities, she adds.
The Deere pilot was powered by Intel® Core™ i7 processors with Intel Movidius™ VPUs. The Intel® Distribution of OpenVINO™ toolkit, implemented with an industrial-strength ADLink Machine Vision Platform and a MeltTools welding camera, rounded out the technology hybrid.
The Deere pilot delivered a big takeaway for other companies looking to add intelligence to existing manufacturing equipment, processes and management. "It's vital you partner with solution providers that can help you solve your challenges today, while ensuring interoperability and scalability to future proof for tomorrow," Boles adds.