What if we fully understood the basic structure of materials and could prevent fracturing from happening before deterioration ever began? Materials are all around us and while researchers have a reasonable understanding of their overall composition, there is plenty still to understand, particularly at the microscopic level.
Levi McClenny, a doctoral candidate in the Department of Electrical and Computer Engineering at Texas A&M University who holds a fellowship from the Data-Enabled Discovery and Design of Energy Materials (D3EM) training program, is working with Dr. Ulisses Braga-Neto on a research project using state-of-the-art machine learning tools to gain a better overall picture of what happens at the microstructure level in materials. Moreover, he is pursuing a full understanding of the fracturing process in order to predict when breakage or deterioration would occur in military vehicles and other structures and ultimately prevent this from happening in the first place.
Recently, McClenny spent two weeks at the Army Research Lab (ARL) in Aberdeen, Maryland, to establish
One of the vehicles being studied with these applications is the very aircraft I fly.
“A vehicle is comprised of many components, all of which are at individual states in their individual lifecycles,” McClenny said. “If we can get an overall system state from these component states, we can report to the driver or the pilot the overall state of his or her vehicle or aircraft in real time. The idea here is to engineer vehicles that can begin to detect their own deterioration.”
This project is uniquely impactful to McClenny as he is a Blackhawk pilot in the United States Army Reserve.
“One of the vehicles being studied with these applications is the very aircraft I fly,” he said. “This is by happenstance, but that makes the project all the more exciting for me. In particular, the vehicle application research for the military could find its way into the force, and knowing that I can impact soldiers in my unit or in future units with this work is extremely exciting.”
McClenny and his collaborative team are seeking to discover how certain microstructure properties relate to material properties by using images from a microscope lens to observe distortions and other inconsistencies within a material. These images could hold a wealth of information on how the materials function in the presence of some sort of stress.
The team is hoping to reverse engineer these observations, essentially trying to generate images from data, with the original images as a “roadmap.” However, these images would have specific desirable properties that the researchers can control.
“If we can generate images with the desired properties then we can potentially determine the processing parameters to generate those materials in real life," McClenny said. "With this approach, we could generate
McClenny is utilizing machine learning and artificial intelligence to research the factors that cause materials to fracture, cracks to propagate and eventually break.
“We want to understand these microstructure interactions, modeled using machine learning approaches, to better leverage their properties for more efficient materials,” McClenny said. “Once an understanding is gained, there are numerous applications, such as ‘smart’ vehicle technology and many others.”
“Levi started as my undergraduate researcher on an NSF project on Boolean dynamical systems, went on to earn a Master of Science in the same area, and is now pursuing his doctorate in the applications of machine learning in microstructure informatics, all the while balancing his duties as an Army Blackhawk pilot and his academic work,” said Braga-Neto, electrical and computer engineering associate professor. “I am gratified to see how far he has come and how much is ahead for him. The collaboration with D3EM and with the ARL have the potential to produce real breakthroughs in the characterization of material properties using artificial intelligence techniques.”