Data drives new AE professor's simulations

7/21/2015 Susan Mumm, Media Specialist

Looking at data, Maciej Balajewicz simplifies computational modeling.

Written by Susan Mumm, Media Specialist

Assistant Prof. Maciej Balajewicz
Assistant Prof. Maciej Balajewicz
Assistant Prof. Maciej Balajewicz
More and more, scientists use simulations to model physical systems. Often, they model smaller and smaller pieces of the system, resulting in billions of equations and a plethora of data.

“That’s fine if you’re interested in something very specific,” believes Maciej Balajewicz, new assistant professor for Aerospace Engineering at Illinois. “The problem is, for a lot of applications, you don’t know exactly what configuration you want to solve. You know you want to build something, but how is that going to look? The design keeps evolving and for a computer, every time (design) changes, you have to resolve the equations. Especially to achieve many answers for different changes of parameters, it becomes too expensive.”

Balajewicz’s research in developing theoretical and computational tools for low-dimensional and low-rank models of multi-scale and multi-physics problems examines questions in reverse, using data to derive which equations are necessary for the simulation, which can be changed and simplified, and which can be discarded.

“Instead of looking at the equations, you look at the data so it can tell you something about the equation you’re solving. That way, you can reduce the number of equations” and the complexity, he said.

Modeling tubulent fluid flows.
Modeling tubulent fluid flows.
Modeling tubulent fluid flows.
Balajewicz uses the approach to study aerodynamic challenges such as aeroelasticity and turbulence. “How does the air, itself, behave? The equations are simple; the solutions are very complicated. It can take weeks and weeks and weeks to do one simulation of a turbulent flow. But we want to get that information quicker with changes of design and purpose incorporated.

“The reality is now that these models aren’t being used routinely because they’re too complicated and expensive,” Balajewicz maintains. “I work to reduce and simplify these complex high-fidelity models to the point at which only a desktop computer is required to solve them.

“The cool thing is that what I do really applies to a lot of systems,” he said. “For example, several colleagues and I are trying to develop model reduction for stock pricing equations. (My work) can be applied to any physical system that is governed by partial differential equations.”

Unsteady transonic aeroelastic simulation.
Unsteady transonic aeroelastic simulation.
Unsteady transonic aeroelastic simulation.
He believes his research in data reduction will be beneficial in collaborations with many of AE’s faculty, including Larry Bergman, Harry Hilton, Marco Panesi, Dan Bodony and Phil Ansell. Using Ansell’s experimental work in dynamic stall as an illustration, Balajewicz said, “In experiments, you collect a lot of data, and you have to get something out of it. Model reduction can provide a smaller, simpler kind of model, and tell other things like ‘What are the coherent structures that live in the flow?’ ‘What are the large vortices and how do they interact?’”

The combination of AE faculty working in computational and experimental aspects played a big role in Balajewicz’s decision to come to Illinois. “The idea of working with both parts of the problem was a major attraction,” he said.

In addition to his research, Balajewicz will teach this fall, starting with AE 451, Aeroelasticity, which is both an on-campus and on-line course offering.

Balajewicz comes to Illinois from Stanford University where he was a postdoctoral research fellow in aeronautics and astronautics. He earned his PhD in mechanical engineering and material science at Duke University in 2012. He earned a bachelor’s and a master’s degree in mechanical and aerospace engineering in 2004 and 2007, respectively, from Carleton University in Ottawa, Canada.

 


Share this story

This story was published July 21, 2015.