Scientists at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) have presented a new method for developing high-strength composite materials. By combining simulations, artificial intelligence and 3D printing, they have succeeded in finding microstructures for fiber composites that have an optimal balance between stiffness and toughness.
According to doctoral student Beichen Li, their systematic approach enables the transfer to other areas such as chemistry or fluid mechanics. The focus was on fiber composites, such as those used in vehicles or aircraft. The aim here is to balance conflicting requirements in terms of stiffness and breaking strength.
The researchers first used 3D printing to generate a large number of test specimens with different microstructures from two starting materials. These were then tested mechanically. At the same time, they simulated the material behavior on the computer. With the help of neural networks as surrogate models, an enormous acceleration was achieved. The simulations predicted the results of the real tests very accurately.
An evolutionary algorithm searched the solution space and identified microstructures with almost optimal mechanical characteristics. The whole process runs like a self-correcting system, matching theory and practice step by step.
According to Li, there are still challenges in terms of consistency in 3D printing and scalability. The goal is a fully automated pipeline from design to mechanical testing. This would significantly speed up and improve the development process for new composite materials.
The publication “Computational discovery of microstructured composites with optimal stiffness-toughness trade-offs” in the journal “Science Advances” has already attracted a great deal of interest. The research was supported by the chemical company BASF, among others.