You are here

Ohio State, Michigan engineers advance nanomaterials manufacturing

Interwoven into almost all aspects of our lives, artificial intelligence (AI) has yet to make a substantial impact in the manufacturing field. Currently, the product and manufacturing design process is one of the most time-consuming aspects of the industry. And when that design doesn’t perform as planned? Engineering teams have to iterate on the design, which wastes both time and resources.

With $1.5 million in funding from the National Science Foundation (NSF), Assistant Professor David Hoelzle will develop an autonomous manufacturing framework that learns on the job. He will take a process that is inherently unpredictable and build manufacturing systems that are robust to that variability.

“Whatever design you make is not going to be exactly what you build,” said Hoelzle, who directs the Hoelzle Research Laboratory on campus. “So we’re trying to build manufacturing systems that learn. Instead of saying ‘here is the design I want to make,’ our focus is on the product performances we want.”

Hoelzle will use an NSF Scalable Nanomanufacturing for Integrated Systems (SNM-IS) Award to fund his Manufacturing for Directed Evolution of Materials (MADE-Materials) project. This collaborative award, which is sponsored by the NSF Division of Civil, Mechanical and Manufacturing Innovation, will allow Hoelzle to partner with the University of Michigan’s Kira Barton, associate professor of mechanical engineering, and Max Shtein, professor of materials science and engineering. 

This interdisciplinary team not only wants to develop a framework, they want that system and its algorithm to be replicable and scalable at multiple levels.

“We hope that this project’s lasting impact will give researchers the ability to take our ideas and apply them to the particular manufacturing process they are working on either in the industry or in their research lab,” Hoelzle affirmed.

He plans to use a versatile manufacturing tool incorporating in-situ materials characterization techniques, in addition to using a central AI algorithm to autonomously direct material synthesis processing. This can evolve, as needed, to meet the desired specification. In their approach, the team’s autonomous system will drive the exploration of the material synthesis space, interpret outputs to statistically diagnose defects or deviations, and learn the process-structure-and-material physics. This closed-loop approach is robust enough to process uncertainties, while also enabling the at-scale manufacturing of sensitive nanostructured smart material synthesis processes, including acoustic metamaterials and optical filters.

Hoelzle’s team of researchers in his lab have explored manufacturing learning-based control for more than four years. Their latest research will go a step further to build a manufacturing system that can continuously build a part, test its performance and then revise the manufacturing parameters.

The final result is a process that yields the ideal product performance each and every time.

Interested in working on this project?

Hoelzle is currently looking for a graduate research associate to assist with the development of his Manufacturing for Directed Evolution of Materials (MADE-Materials) framework. If you are interested, contact him at

by Kam King, Dept. of Mechanical and Aerospace Engineering