This is an excerpt from an article published on FAB Shop Magazine written by Len Zapalowski
As automation progresses, the challenge for both vendors and buyers is to balance the advance of innovation with practical applications that make real-time sense for users and offer a high return on investment. To overcome those challenges, automation companies are developing a thorough understanding of the established approaches that are implemented on factory floors as well as the specific pain points that each manufacturer faces. In doing so, they have the knowledge to engage the appropriate technology at the right time for the right problem. This has never been more true than in the application of (AI) to robotic welding applications.
Robotic welding is still evolving, but is now being enhanced with vision and AI. The magic of these four technologies – welding, robotics, vision and AI – and the fact that vendors now have more technologies to draw from than ever before is promising. But, the industry should still take the time to understand the next layer of improvements that manufacturers can attain.
Infusing AI
Among the automation companies looking to lead in this new era of AI-enhanced welding is Novarc Technologies, Inc., a full-stack robotics company specializing in the design and manufacture of robots and computer vision AI solutions for automated welding applications. The company developed NovAI, its parent technology, on the back of two decades of research and practical industrial grade deployments of its award-winning Spool Welding Robot (SWR). In this time, Novarc has transferred automation lessons from high-grade welds on pipes to high-quality welds of structural steel.
NovAI Autonomy, Novarc’s latest AI-powered system, brings both vision and real-time adaptation to articulated robotic and mechanized welding, which holds great potential for various applications that can benefit from AI in robotic welding. It allows operators and fabricators to intuitively address part fit-up variations and distortions during the weld by providing real-time welding and robotic parameter adjustments.
Historically, automated welding systems have struggled with part fit-up variations, distortion and weld seam distortions, which leads to costly rework, inconsistent quality and downtime. These automated systems rely on pre-scanning and manual adjustments that slow production and require skilled welders.
On the flipside, however, Novarc learned that by combining machine vision with machine learning, robots can see and adapt in real time like an expert welder – without any operator intervention. Furthermore, NovAI Autonomy eliminates the need for pre-scanning, manual adjustments and rework and, therefore, speeds up manufacturing time while also increasing the quality of the welds.