R&D PULSE
A new method could help engineers avoid defect formation.
Written by Kayt Sukel
SINCE THE EARLY 1900s, manufacturers have relied on friction stir welding (FSW), a solid-state joining process leveraging the combination of frictional heat and mechanical force, to bring together materials like aluminum with high strength and less distortion, without problems like hot cracking or porosity. Yet, despite its many advantages, it has been challenging for manufacturers using this technique to visualize potential defects and avoid costly rework.
“This technology holds a lot of pedigree as it takes away a lot of the issues involved with fusion welding,” said Hemant Agiwal, who recently received his doctorate at the University of Wisconsin-Madison in Frank Pfefferkorn’s laboratory but now works as a senior scientist in mechanical research at Raphe mPhibr Pvt. Ltd. in Noida, India.
“But because, with this technique, everything happens at the surface of the material, both the scientific and the industrial community have struggled to visualize how or when defects might come,” he said. “Anyone trying to do this with a new material or a new machine or new alloy will have the problem of understanding if the weld will work or if it will have defects.”
“This research gives you a roadmap in understanding, regardless of what material you use, there will always be transition regions. You need to understand those transition regions, and where the transition is coming in, so you can know the areas where you’ll find a good build.”
—Hemant Agiwal, a doctoral candidate in the Mechanical Engineering Department at the University of Wisconsin-Milwaukee
Video: University of Wisconsin-Madison
Agiwal said there has been decades of work trying to address this issue, and most rely on in situ visualization using cyber-physical systems (CPS) or high-power X-rays. Unfortunately, these costly techniques are not available in standard laboratories.
“These technologies are very expensive. You cannot use them for everyday research, and you definitely cannot use them in a production line,” Agiwal explained.
To come up with an easier (and more replicable) way to understand material flow in FSW applications, Agiwal, Pfefferkorn, and colleagues looked for a new method that could be done by anyone, even in more limited resource environments. The research team relied heavily on the work of W.J. Arbegast, who had detailed a critical model of the FSW process, stating that understanding any mass imbalances caused by insufficient material flow can help determine the quality of the weld.
“That work showed there’s always a transition where you go from a good weld to a bad one,” Agiwal said. “Based on that, we wanted to see if we could find some sort of signatures in images or in forces to identify those transitions. We started to look at 2D images of our weld areas to gain a quantifiable, volumetric understanding of what might be happening in the weld.”
Agiwal and colleagues relied on macrographs, or extreme close-up images, of welded samples to provide new insights into material flow. Agiwal said the images provided “interesting variations” to help welders know if any defects might occur.
“Our results agreed with previous postulations and hypotheses regarding understanding this technology from a defect standpoint,” he said. “This research gives you a roadmap in understanding, regardless of what material you use, there will always be transition regions. You need to understand those transition regions, and where the transition is coming in, so you can know the areas where you’ll find a good build.”
The team’s work, “Empirical Analyses of Weld Zones to Understand Material Flow and Defect Formation During Friction Stir Welding,” was recently published in the Journal of Manufacturing Science and Engineering.
Video: University of Wisconsin-Madison
This study demonstrates the importance of always going back to “first principle physics” when trying to solve problems, Agiwal said. And, today, as part of a research team at Raphe mPhibr Pvt. Ltd, Agiwal is still taking this approach as he seeks to understand new materials.
“It was such an amazing thing to see that differences in color patterns in these images, something that we never really gave any thought to, can translate into understanding the physics of a process,” he said. “When you go back to those first basic principles and think about the best way to visualize the physics, you’ll find you don’t always need heavy equipment or X-ray or other things like that. By just focusing on the physics predictions, you are trying to make, you can find new ways to achieve that goal.”
Kayt Sukel is a technology writer and author in Kansas City.
© 2026 The American Society of Mechanical Engineers. All rights reserved.