Defect formation, Jyoti Mazumder says, is a huge problem in manufacturing.
According to Mazumder, a professor of mechanical engineering and materials science at the University of Michigan and founder of Sen Sigma, hundreds of millions of dollars are wasted every year on defective products in the United States, with a big chunk of those defects occurring in the manufacture of automobiles. An average mid-sized passenger vehicle contains almost 5,000 spot welds, for instance, and each one has the potential to be defective.
Mazumder became interested in developing a computerized way to detect weld defects right on the assembly line more than a decade ago, when he was involved in a research project for Aetna Industries and Toyota. The issue, he says, is that most car companies rely on a spot welding technique in their factories that makes on-the-fly assessment of the weld quality difficult. Detecting defects in a quality control phase later in the manufacturing process increases costs, since it involves stopping the line, delaying production, and scrapping defective materials.
To help solve this problem, Sen Sigma has created something it calls the Smart Optical Monitoring System (SOMS), which is able to perform diagnostic tests and detect weld and material deposition problems in real time. Light is the source of valuable data during the manufacturing process—its color can reveal information about temperature, composition, and defects. Installed on a welding robot and connected to a computer, SOMS measures and analyzes the wavelengths of light emitted by laser-induced plasmas during the welding process, looking for patterns that signal a problem.
Lijun Song, Sen Sigma’s chief technical officer, says that the company currently has four SOMS prototypes that are being used for field testing by manufacturers, and Sen Sigma plans to partner with Caterpillar and ABB Robotics to further test SOMS prototypes. The initial tests have been positive, he says.
In addition to the auto industry, manufacturers of wind turbines and ships are also interested in the technology, Mazumder says. “Any manufacturing process which produces light—we can use that to analyze defects,” Mazumder says. “We’re not touching anything, we’re just looking at the light, and the sensors will detect if there’s anything wrong.”
Officially launched in 2011, Mazumder says Sen Sigma’s business model is to sell SOMS units to manufacturers and teach them how to use it. “We’re debating internally whether we want to also provide service,” he says.
Sen Sigma is Mazumder’s third company—he sold the first two, one of them to Honeywell—and he wants to try to grow this one through revenue and grants. “Maybe I’ll take venture capital when I’m ready to scale,” he says.
The technology behind SOMS is based on basic science that Mazumder, who was named U-M’s Distinguished University Innovator in 2012, has been working on at the university for many years.
“It’s not new, we’re just applying it to modern problems,” he adds.
