With support from a three-year, $576,943 National Science Foundation (NSF) grant, materials science and engineering assistant professor Ryan Sills and industrial and systems engineering assistant professor Aziz Ezzat are collaborating to develop a model for predicting the failure of ductile metals such as aluminum and steel.
For principal investigator (PI) Sills, the award – his first from the NSF – is a major career milestone. “It’s also an opportunity to collaborate with Co-PI Aziz, who I greatly enjoy working with and learning from.”
“The award will enable our research team to explore challenging – and exciting – new avenues right at the interface of data and materials sciences,” says Ezzat.
“Ryan and I started our academic careers here at the same time in late 2019, and have since shared a well-aligned and complementary research vision for bringing data and materials science closer,” says Ezzat. “I’m extremely happy and excited about NSF enabling us to realize our vision, expand our research teams, accelerate our research progress in this area, and build on our exciting – and I hope – long-lasting collaboration that might serve as an exemplar of how data science can intertwine with materials science, and vice versa.”
Formulating a Predictive Model
Used by industries ranging from automotive and aerospace to defense and energy, ductile metals are part of everyday life. Yet because the reasons they fail are so poorly understood, it is virtually impossible to accurately predict when fractures might occur. The result is higher costs, diminished energy efficiency, and an inability to develop materials for energy and defense applications able to reliably resist fracture.
The professors’ interdisciplinary project will devise a physics-informed, machine-learning enabled model for ductile fracture prediction that paves the way for lower product costs and greater energy efficiencies.
“Our goal is to combine Ryan’s expertise in micromechanics and computational modeling with my research group’s strength in physics- and domain-informed data science modeling so we can better explain and predict ductile fracture in engineering materials,” says Ezzat.
The benefits are clear, explains Sills, “Right now all models used to predict damage in ductile metals when designing things like cars, bridges, and aircraft are best guesses. With our approach, no guessing will be necessary, because we are directly simulating the specific processes leading to damage.”
According to Sills, while the project will focus on a few specific aluminum and steel alloys, their findings should be applicable to the many different metals used in energy, defense, and aerospace applications – and eventually lead to greater cost savings and efficiencies.
Ezzat and Sills will work “hand-in-hand” on the project. “My PhD student and I will perform simulations and assemble datasets, which Aziz and his PhD student will then apply machine learning techniques to. Then, they will feed their results back to us as we develop a new model for damage and fracture,” Sills notes.
Promoting Undergraduate Learning
While the award supports a PhD student in each of the professors’ departments, it also provides opportunities for undergraduates through research projects and enriched classroom experiences.
Sills, for example, plans to integrate some MSE senior lab projects so that undergraduates can join the team’s efforts, while Ezzat will design in-class data challenges involving materials failures inspired by the project that students in his industrial informatics course can solve with data-science-based tools.
High school students will also benefit through new modeling and machine learning modules that will be piloted in a high school physics classroom through an innovative collaboration with a high school teacher.