MIT Incorporates Artificial Intelligence to Accelerate the Discovery of New Materials for 3D Printing
The increasing popularity of 3D printing for manufacturing a variety of items, from customized medical devices to affordable housing, has increased demand for new 3D printing materials designed for extremely specific applications.
To shorten the time required to discover these new materials, MIT researchers developed a data-driven process that uses machine learning to optimize new 3D printing materials with a variety of properties, such as toughness and compression strength.
By streamlining material development, the system reduces costs and has a positive impact on the environment by reducing chemical waste. Additionally, the machine learning algorithm may stimulate innovation by suggesting novel chemical formulations that human intuition may overlook.
"Materials development remains a largely manual process. A chemist enters a laboratory, manually mixes ingredients, prepares samples, tests them, and arrives at a final formulation. However, rather than having a chemist perform a few iterations over a few days, our system can perform hundreds of iterations in the same amount of time," says Mike Foshey, a mechanical engineer and project manager in the Computer Science and Artificial Intelligence Laboratory's (CSAIL) Computational Design and Fabrication Group (CDFG), and co-lead author of the paper.
Additional authors include Timothy Erps, a technical associate with the CDFG; Mina Konakovic Lukovic, a CSAIL postdoc; Wan Shou, a former MIT postdoc who is now an assistant professor at the University of Arkansas; Wojciech Matusik, a professor of electrical engineering and computer science at MIT; and Hanns Hagen Geotzke, Herve Dietsch, and Klaus Stoll of BASF. The research was published in Science Advances on October 15, 2021.
The researchers developed a system in which an optimization algorithm performs the majority of the trial-and-error discovery process.
A material developer selects a few ingredients, feeds the algorithm information about their chemical compositions, and specifies the mechanical properties of the new material. The algorithm then adjusts the proportions of those components (much like knobs on an amplifier) and evaluates how each formula affects the material's properties before arriving at the optimal combination.
The developer then mixes, processes, and tests the sample to determine the material's true performance. The developer communicates the results to the algorithm, which automatically learns from the experiment and uses the newly acquired knowledge to choose another formulation to test.
"We believe that this method would outperform the conventional method in a number of applications because it relies more heavily on the optimization algorithm to find the optimal solution. "You would not require the presence of an expert chemist to pre-select the material formulations," Foshey explains.
The researchers have developed AutoOED, a freely available open-source materials optimization platform that incorporates the same optimization algorithm. AutoOED is a comprehensive software suite that enables researchers to perform their own optimization.
Producing materials
The researchers validated the system by using it to optimize formulations for a new UV-curable 3D printing ink.
They chose six chemicals to use in the formulations and instructed the algorithm to find the material with the highest toughness, compression modulus (stiffness), and strength.
Manually optimizing these three properties would be particularly difficult because they can be in conflict; for example, the strongest material may not be the stiffest. Using a manual process, a chemist would typically attempt to optimize a single property at a time, resulting in numerous experiments and significant waste.
After testing 120 samples, the algorithm identified 12 materials with the optimal trade-offs between the three different properties.
Foshey and his collaborators were taken aback by the breadth of the materials generated by the algorithm, noting that the results were far more varied than they anticipated based on the six ingredients. The system promotes exploration, which may be particularly advantageous in situations where specific material properties are not readily apparent intuitively.
Additional automation could speed up the process even more. The researchers mixed and tested each sample manually, but in future versions of the system, robots could operate the dispensing and mixing systems, Foshey says.
Additionally, the researchers would like to test this data-driven discovery process for applications other than developing new 3D printing inks in the future.
"This has a wide range of applications across the field of materials science in general. For instance, if you wanted to develop new types of batteries that were more efficient and less expensive, you could do so using this system. Alternatively, if you wanted to optimize paint for a car that was both efficient and environmentally friendly, this system could do that as well," he says.
This work could be a significant step toward realizing high-performance structures because it presents a systematic approach for identifying optimal materials, according to Keith A. Brown, assistant professor in the Department of Mechanical Engineering at Boston University.
"The emphasis on novel material formulations is particularly encouraging, as this is an aspect that is frequently overlooked by researchers working with commercially available materials. Additionally, the team's efficient identification of materials is enabled by the combination of data-driven methods and experimental science. Because experimental efficiency is a concept shared by all experimenters, the methods presented here have the potential to motivate the community to adopt more data-driven practices," he says.
from Scholars Globe - Free Tutorials, HowTo & News in Science, Tech, Business & Lifestyle https://ift.tt/3aLpThR
via IFTTT
No comments:
Post a Comment