Getting into shapes:
Krishnan Suresh applies CAREER award to optimization
rom automobile parts to microelectromechanical systems, engineers use computer models to optimize components for manufacturing. However, when the components become geometrically complex—for example, having parts of different thickness or shape—modern techniques fail. The National Science Foundation awarded Assistant Professor Krishnan Suresh a 2008 Faculty Early Career Development Award (CAREER) to develop a next-generation optimization framework for synthesizing geometrically complex artifacts.
“The combination of all kinds of shapes is inherent in engineering,” says Suresh. “Just because we can’t handle it doesn’t mean we don’t want to make those components. We’re trying to address that basic limitation.”
Current shape optimization techniques rely on ad hoc processes, which are not robust enough to address the complexity of design or the variety of constraints that design engineers face, says Suresh. Using simplification techniques and concepts such as feature sensitivity and dual representation, he will use his five-year, $400,000 award to build a mathematical framework so fundamental that it can optimize virtually any manufacturable object. Therefore, this research could have a major effect on industry and society, creating more efficient, less expensive and more environmentally friendly components.
“If you give me a part, and say, ‘I want to minimize the thickness but make sure it doesn’t break,’ then I can show you how to do it in a robust and efficient manner, as opposed to current techniques, which will fail simply because this component has both thin parts and thick parts built in,” Suresh says.
Suresh first will apply his framework to specific case studies to establish the fundamental mathematics. To do this, he plans to work with the student hybrid vehicle teams, such as the Formula SAE Team. Every year, the Formula SAE Team designs and builds a formula-style racing car for competition with more than 100 other universities and colleges. Sponsored by SAE International, the competition challenges students not only to build a high-performance car but also to reduce emissions and increase knowledge of alternative fuels. The UW-Madison team currently uses commercial tools for optimizing parts, but Suresh hopes to work with the students to identify challenges inherent in cutting-edge vehicle design and address them with the new framework, while training team members in the underlying concepts of shape optimization.
“Working with them is a win-win situation,” says Suresh. “Hopefully we will help them design better parts, and they will understand the reason behind much of the mathematics.”
The NSF CAREER awards, among the most prestigious given to faculty members who are just beginning their academic careers, are granted to creative projects that integrate research and education effectively.