Advanced materials are engineered to exhibit enhanced properties, such as increased strength, fatigue life, or corrosion resistance to fuel future technological progress in the aerospace, automotive, and energy industries. 

However, their complex behavior presents significant challenges, often requiring extensive and costly experimental testing to fully understand their performance. 

Mechanical Engineering Assistant Professor Andrew Gross has spent the past three years developing a new approach to address this issue. His method combines digital imaging with computational modeling to reduce both the time and cost associated with characterizing the mechanical properties of advanced materials. 

Gross’s work is supported by a Defense Advanced Research Projects Agency (DARPA) Young Faculty Award, initially granted in 2023. The project focuses on streamlining material property testing, particularly for advanced materials like those produced through additive manufacturing, which typically require far more testing than conventional materials. 

Gross describes his work as part of a broader shift toward “material testing 2.0,” which seeks to replace methods dating back to the Industrial Revolution, which are increasingly inadequate for modern advanced materials. The traditional testing methods evaluate one load state at a time, often requiring around 20 separate experiments to calibrate state-of-the-art material models. Gross aims to replace this approach with a more efficient framework capable of extracting complex material behavior from a single experiment that probes multiple loading states simultaneously. 

“We have this novel framework for how we're going to take any efficient experiment and pull out this complex material behavior from it,” Gross says. “But before applying our framework, we needed to come up with an experiment test to activate that complex material behavior.”

To pack multiple traditional tests into a new, highly efficient one, Gross first needed to develop a metric to evaluate how many load states could be represented in a single test and whether those states could be effectively measured. 

“All existing work in this area of research has been focused on 2D mechanical tests,” Gross says. “We needed to lift specimen performance measures up to the 3D world we live in and ensure that our new measure is readily applicable across a wide array of advanced metals.”

The complex varied conditions expected across the new test specimens drove Gross to employ an eight-camera 3D digital image correlation (DIC) rig to resolve these local variations. DIC is a non-contact optical measurement technique used to analyze the deformation motion and shape of solid objects by comparing images. DIC methods were invented by Mechanical Engineering Distinguished Professor Michael Sutton in 1982 and have become an indispensable technology used in mechanical testing. The new 3D specimen performance metric developed by Gross also needed to consider the measurability of the experimental data.

“In mechanical testing, what’s happening throughout the inside of the test specimen influences all of the data,” Gross says. “A test specimen may generate different load states, but if you can't measure it on the surface with standard lab equipment, then it's not useful.”

With the new specimen performance metric in hand, Gross used engineering intuition to create a design space ripe with candidate test specimen designs that rank competitively. 

“We know if you take a structure and twist it, you get the maximum deformation on the surface,” Gross says. “The same also happens if you bend that structure. You get maximum tension on one surface and maximum compression on the other. By combining twisting and bending, a variety of multiaxial strains are generated in one experiment.”

He then unleashed a computationally intensive shape optimization algorithm to design a test specimen with maximum performance. The resulting design—a curved beam—intentionally creates a range of multiaxial strain conditions within one experiment that are readily measured with DIC. The result is the most efficient experiment for elastoplastic model calibration ever made.

Cost efficiency is central to Gross’s approach. Rather than relying on expensive, specialized equipment, he designed the system to work with standard laboratory setups. This allows tests to be conducted for a few thousand dollars instead of the hundreds of thousands typically required for comprehensive material characterization campaigns. 

“If we say, ‘here’s a new test that only costs $2,000 to run, but you need to buy this million-dollar piece of equipment,’ it defeats the purpose,” Gross says. “We want something that's inexpensive to run on the equipment that is already standard.” 

Gross believes this pragmatic approach is necessary for material testing 2.0 to gain traction and change the state of practice for engineers. He highlights that a gap has grown over the last 70 years between the state of engineering science and practice. For example, engineers regularly use theories from more than 100 years ago, even though they are less accurate for today’s advanced materials. 

“The high cost of the experimental test campaigns required to use the more accurate modern material models is their main impediment from use in practice,” Gross says. “If we can make calibration of advanced material models cost competitive with the classical theories then engineers can readily incorporate the most contemporary science in their material modeling. This will increase the accuracy of engineering predictions that drive component design and manufacturing, leading to cost savings and higher performance.”

Gross aims for the rest of the model calibration framework, which combs through the wealth of data generated in the new experiments, to be published later this year.