Michigan Engineering News
Battery production technology recently received the S.M. Wu Research Implementation Award from one of the top professional societies in manufacturing.
Experts
A monitoring system for ultrasonic welding, developed by University of Michigan and General Motors researchers for production of the Chevy Volt electric vehicle (EV), identified nearly all bad welds, avoiding an estimated $22 million in potential repairs.
Designed and implemented roughly 15 years ago, the technology was recently recognized with the S.M. Wu Research Implementation Award from SME, formerly known as the Society and Manufacturing Engineers. It’s an award for “outstanding original research presented as a paper at the annual North American Manufacturing Research and, subsequently, upon implementation, had a significant commercial and/or societal impact.”
The inline monitoring system—integrated directly into the production line—had a success rate of 99.99995%, missing two defective welds out of more than 4 million total welds. The technology helped improve the quality of EVs, lower their cost and improve production efficiency.
The Volt, an extended-range EV, was an early foray into mass-produced EVs for Chevrolet. Funding came from the U.S. Department of Energy under the Obama administration, with the aim of helping GM get the Volt into production. The company brought U-M researchers on board to help with the battery assembly process.
“Before we had this system, General Motors was using manual inspection,” said Chenhui Shao, a U-M associate professor of mechanical engineering. “They hired engineers to manually check every single weld once the process was completed.
“The process was very slow, and also inconsistent. Different people may have different opinions on the quality of a weld. Visually, it’s not very quantifiable.”
The system designed at U-M utilizes instruments that provide a large amount of data points generated during the welding process. Ultrasonic metal welding is a solid-state joining process using high-frequency vibrations, as compared to traditional fusion welding using heat.
A high-frequency, ultrasonic microphone monitors the sound vibrations produced. Those vibrations provide information about the weld as it occurs. In addition, a watt meter measured the electrical power being delivered to the welder in real-time.
“We take in those signals and pull the information out of them, and recombine it,” said Jeff Abell, former director and chief scientist for global manufacturing systems at GM, now a professor of engineering practice at U-M. “When we look at the combinations of those factors, we can say ‘Oh, there may be something going awry with the process.”
Abell, who will join the U-M faculty later this month, said the inline monitoring system purposely avoids utilizing vision based evaluations of the welds.
“We focused on pulling information right out of the manufacturing process to do our quality assessment,” he said. “We had a strong philosophy about not using vision. A good weld and a bad weld can look remarkably the same.”
The early work was performed under the leadership of Abell and Jack Hu, a long-time U-M faculty member who is now chancellor at the University of California Riverside. The U-M team would eventually include: Judy Jin, Elijah Kannatey-Asibu Jr., Tae Hyung Kim, Kamran Paynabar, Juil Yum and Grace Guo from U-M, as well as J. Patrick Spicer, Wayne Cai, Debejyo Chakraborty, Jennifer Bracey, and Hui Wang from GM.
Built at GM’s assembly plant in Brownstown, MI, the first generation of the Volt was produced from late 2010 through 2015. Since then, other vehicles have benefitted from the approach pioneered by U-M and GM.
“Fifteen or sixteen years ago, this was all very new,” Shao said. “But automakers are using it all the time now. The concept has been implemented with other products and production lines.”
