Nanotechnology

Machine learning and 3D printing produce strong foaming materials of steel.

From left to right: A complete grid -shaped image is lined up with 1875 million cells floating in the bubble. Credit: Peter Cells /Toronta University Engineering

Researchers at the University of Toronto University Applied Science & Engineering have designed a nano -amterior material with light Styrofoam’s strength using machine learning.

A new paper published in the Advanced Materials explains how the team led by Professor Tobin Filleter has created an exceptional strength, lightweight, and a combination of inconsistent combinations. This approach can benefit a wide range of industries, from automobiles to aerospace.

“Nano architetlet material is a nanoscale size that uses the” smaller one “effect, and combines high -performance shapes, such as creating a bridge from a triangle, achieving some from the highest strength and rigidity to rigidity. I will do it. “The weight ratio of all materials” says Peter Cells, the first author of the new paper.

“However, the standard lattice shape and shape used tends to have sharp intersections and corners, leading to stress -concentration issues. This causes local failure and damage to the material. The possibilities are limited.

“When I thought about this issue, I realized that it was a perfect problem to work on machine learning.”

Nano archites are made of small building blocks or repeated units that measure several hundred nanometers. If more than 100 exceed 100, patterns are lined up to reach the thickness of the human hair. In this case, these building blocks, which are composed of carbon, are located in a complex 3D structure called nanorattis.

To design improved materials, Serles and Filleter cooperated with Professor Seunghwa Ryu and the Ph.D. Jin Wukuyo is a student of the South Korean High Science and Technology Research Institute (KAIST). The partnership was launched through the Toronta University’s international doctoral program cluster program. This program supports a doctoral program through research with international collaborators.

The KAIST team has adopted a multipurpose Bayisian optimization machine learning algorithm. This algorithm has been learned from a simulated geometry to strengthen stress distribution and predicts the best geometry to improve the ratio of nano arched design strength and weight.

Later, Serles created a prototype for experimental verification using a 2 -optical multiple 3D printer (Craft) contained in the center for research and application of fluid technology. This additive manufacturing technology enables 3D printing with micro and nanoscale, creating an optimized carbon nanorattes.

These optimized nanolatis withstands 2.03 mega pascal stress for each cubic meter per cubic kilogram, with a 2.03 mega pascal stress, about 5 times the titanium.

“This is the first time that machine learning has been applied to optimize the Nano Architect Material, and I was shocked by improvement,” Serless says. “It’s not just a duplication of successful geometry from the training data. The shape changes worked, learned what did not work, so that you could predict a completely new lattice shape.

“Machine learning is very very data -consolidated, and it is difficult to generate a lot of data if high quality data is used from finite element analysis. More than 20,000 people are required.

“These new material design ultimately, ultimately lightweight weight ingredients in aerospace applications such as airplanes, helicopters, and spaceships that can reduce fuel demand in flights while maintaining safety and performance. I hope it will be connected, “says the filet. “This ultimately helps reduce high -flight carbon dioxide emissions.”

“For example, if you replace components made of airplan titanium with this material, you will consider saving 80 liters per year for each replacement material,” Serless says.

Other contributors to this project include Professor Juzu, Chandra Via Sin, Janehae, Charles Jia, and Germany, Massachusetts, Massachusetz, Includes the Technical Research Institute (MIT) and international collaborators at Rice University. US.

“This was a multifaceted project that linked various elements of material science, machine learning, chemistry, and mechanic that helps understand how to improve and implement this technology.” Serus, a fellow, says. Caltech.

“Our next step focuses on further improving the scale of these material designs and enabling expensive macroscale components,” Filleter adds.

“Furthermore, we will continue to explore new designs that push the material architecture to even lower density while maintaining high strength and rigidity.”

Details: Ultra -high strength due to Peter Serless et al, Bayesian Carbon nanolattices, and Advanced Materials (2025). Doi: 10.1002/ADMA.202410651

Provided by Toronto University

Quoted: Machine learning and 3D printing yield steel strong form light material (January 24, 2025) January 24, 2025 https://phys.org/news/2025-01-01-Yield-Steel Acquired from -stel-strong.html

This document is subject to copyright. There is no part that is reproduced without writing permission, apart from fair transactions for private research and research purpose. Content is provided only by information.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button