Bryony DuPont joined the School of Mechanical, Industrial, and Manufacturing Engineering as an assistant professor in 2013. She is now one of seven faculty members who make up the largest academic mechanical engineering design research group in the nation.
It’s called the OSU Design Engineering Lab, and DuPont brings her computational expertise in design automation with a laser focus on the long-term environmental impacts of early design decisions. Although these impacts often don’t manifest until very late in a product’s lifecycle, early design decisions can play a critical role in everything from recyclability to water conservation to energy consumption.
DuPont and her students tap artificial intelligence, machine learning, and algorithms to develop methods and computational tools for improving the design process so products are better for the planet – from cradle to grave.
“All our work is computational – system optimization, algorithm development, and machine learning, and the main focus is tackling green problems,” DuPont said. “We do a lot of work in renewable energy and energy systems and a lot on the environmental impact of consumer products – how people use and interact with consumer products and how this affects environmental sustainability.”
One of DuPont’s research projects is aimed at helping design engineers understand the long-term environmental implications stemming from the decisions they make very early on in the design process. This focus on early design decision-making for environmental impact is relatively new, for several reasons. Not only has environmental design often taken a back seat to profits, but reliable data has been lacking, or non-existent.
“Most engineers don’t know where to start when it comes to designing to reduce environmental impacts, because there are currently no methods to help you during the early design phase – right when you’re getting started,” DuPont said. “So we’re creating some of the first data sets and computational tools that will change that.”
One of the tools is a web-based, quiz-like decision engine that asks engineers a series of key questions early on – questions ranging from power supply to whether or not plastic parts can be made from materials that qualify for the most common recycling symbols (1 and 2).
“If you’re using 12 different materials but only some are recyclable, or you can’t disassemble the product to extract the recyclable materials, or if batteries will need replacement every few weeks, our system will call that out.,” DuPont said.
DuPont and her students are also developing a repository to address the lack of data available to design engineers.
“We’ve created a repository of 47 products with 26 different environmental impact metrics, and we’re adding to it all the time,” DuPont said. “It’s a component-by-component analysis of what products are made of, how they’re made, and the environmental impacts of each component.”
DuPont is using machine learning to find the correlation between the decisions a designer makes and the environmental impacts that result from those decisions.
She’s also applying her computational expertise to improve the ability of the power grid to more efficiently accommodate renewable energy and to determine if offshore energy systems, like floating wind farms, might work well on the Pacific coast.
During a $160,000 research project sponsored by the U.S. Dept. of Energy’s National Energy Technology Lab, DuPont optimized the Oregon and Washington power grid and pointed out ways of managing the power that have the potential to cut the cost of energy by almost 18 percent.
Some of her students are working on another research project that is simulating floating wind systems off the coast in order to analyze the cost and biological impacts of floating wind turbines and the potential impact on energy costs in Oregon.
“That one is a fun and very, very challenging problem, but these students are at the forefront of it,” said DuPont, who is seeing an uptick in interest from women in this area of engineering. “I have so many great students exited to do this work, in part because they can see how they can apply the work to big, save-the-world issues.”
— Gregg Kleiner