A machine-learning system created at Sandia National Laboratories might make testing bulk materials faster and less expensive for the aerospace, automotive, and other industries.
The technique was published recently in the scientific journal Materials Science and Engineering: A.
Production stoppages are costly. So, manufacturers screen materials like sheet metal for formability before using them to make sure the material will not crack when it is stamped, stretched and strained as it’s formed into different parts.
According to Sandia scientist David Montes de Oca Zapiain, who is the paper’s principal author, businesses frequently employ commercial simulation software that has been adjusted to the outcomes of various mechanical testing. However, it may take months to finish these testing.
In addition, Montes de Oca Zapiain noted that while some high-fidelity computer simulations can evaluate formability in just a few weeks, corporations need access to a supercomputer and specialized knowledge to conduct them.
According to Montes de Oca Zapiain, Sandia has demonstrated that machine learning may significantly reduce the amount of time and resources required to calibrate commercial software because the algorithm does not require data from mechanical tests.
Nor does the method need a supercomputer. Additionally, it opens a new path to perform faster research and development.
“You could efficiently use this algorithm to potentially find lighter materials with minimal resources without sacrificing safety or accuracy,” Montes de Oca Zapiain said.
The developed algorithm is about 1,000 times faster compared to high-fidelity simulations. We are actively working on improving the model by incorporating advanced features to capture the evolution of the anisotropy since that is necessary to accurately predict the fracture limits of the material.
Hojun Lim
Algorithm replaces mechanical tests
Because metal alloys are composed of tiny, so-called “crystallographic” grains, the MAD3 machine-learning algorithm pronounced “mad cubed” and short for Material Data Driven Design works.
Together, these grains create a texture known as mechanical anisotropy, which makes the metal stronger in some directions than others.
“We’ve trained the model to understand the relationship between crystallographic texture and anisotropic mechanical response,” Montes de Oca Zapiain said. “You need an electron microscope to get the texture of a metal, but then you can drop that information into the algorithm, and it predicts the data you need for the simulation software without performing any mechanical tests.”
In collaboration with Ohio State University, Sandia used a method known as a feed-forward neural network to train the algorithm using the results of 54,000 simulated materials tests.
The Sandia team then supplied the algorithm with 20,000 fresh microstructures to evaluate the method’s precision, comparing the algorithm’s calculations with information acquired from simulations and experiments conducted on supercomputers.
“The developed algorithm is about 1,000 times faster compared to high-fidelity simulations. We are actively working on improving the model by incorporating advanced features to capture the evolution of the anisotropy since that is necessary to accurately predict the fracture limits of the material,” said Sandia scientist Hojun Lim, who also contributed to the research.
In its capacity as a national security lab, Sandia is carrying out more research to determine whether the algorithm can speed up the quality assurance procedures for the U.S. nuclear stockpile, where components must adhere to strict requirements before being approved for use in production.
The Advanced Simulation and Computing program, run by the National Nuclear Security Administration, provided funding for the machine learning study. The user-friendly, graphics-based Material Data Driven Design program was created by a cross-disciplinary team assembled by Sandia to allow other universities to benefit from the technology.
More than 75 interviews with prospective users were conducted as part of its development by the Department of Energy’s Energy I-Corps program.
