As a silicon wafer winds its way through a semiconductor manufacturing plant—from deposition and lithography to etching and dicing—each stop at another machine loaded with sensors also generates information.
And all those numbers, with advanced analysis, could be harnessed to further optimize the production process.
“There are many gigabytes of data for each single product,” says Andi Wang, a researcher whose work applies data science to advanced manufacturing processes. “How do you use that? Industry does not want to discard this data, but they do not know how to use it. This is where our research comes into play.”
Wang is joining the University of Wisconsin-Madison in fall 2024 as an assistant professor of industrial and systems engineering, bringing experience employing a variety of machine learning and other analytical methods to real-world complex manufacturing systems.
He arrives in Madison after spending three years on the faculty of Arizona State University’s School of Manufacturing Systems and Networks, which provided him with insights into emerging manufacturing technologies.
“What I’d like to do is bring that knowledge that I’ve gained and also this experience in the manufacturing field back to industrial engineering,” says Wang, who holds PhDs from Hong Kong University of Science and Technology and from Georgia Tech.
Specifically, Wang is interested in analyzing data to guide design decisions and to improve efficiency and performance in additive manufacturing systems and the semiconductor manufacturing industry. He’s previously analyzed data from steel-rolling plants to uncover process and quality improvement strategies.
Wang has also expanded his research to include nuclear reactor design optimization. He currently leads a project, backed by a $1 million grant from the U.S. Department of Energy, to develop novel data-driven modeling and optimization methodologies (driven by the special characteristics of nuclear reactor core simulations) to shorten design cycles.
He says UW-Madison’s longstanding excellence in nuclear engineering was a draw, along with a growing emphasis on semiconductor research across the College of Engineering.
“I’m impressed by the great research environment here,” says Wang, who’s seen it firsthand.
In 2015, Wang spent six months at UW-Madison as a visiting researcher while pursuing his first PhD, an experience that drove his interest in coming to the United States to continue his career. Several faculty members he met at UW-Madison, such as Professor Kaibo Liu, who shares similar research interests as Wang, have remained his academic mentors.
Wang will start off teaching ISyE 210: Introduction to Industrial Statistics, an introductory engineering analytics course that’s open to all engineering students. But he’s already got plans to develop a course on artificial intelligence for industrial applications in the future.
“Lots of industrial engineering students want to know about the cutting-edge developments of AI, machine learning and data analytics approaches, and how to implement those in the industrial workflow,” he says.
Wang considers himself a versatile instructor, and the same could be said about his research. The ability to apply analytical methods to a whole host of industries and systems was part of what first captured his interest as an undergraduate student at Peking University in his native Beijing.
“A very interesting aspect of industrial engineering,” he says, “is that you can observe that data from different engineering systems, of different applications, sometimes share similar insights, which drive us to develop new models and methodologies. These outcomes can solve a class of problems in various applications.”