Machine learning for composite materials
- PDF / 765,381 Bytes
- 11 Pages / 612 x 792 pts (letter) Page_size
- 103 Downloads / 249 Views
Artificial Intelligence Prospective
Machine learning for composite materials Chun-Teh Chen, Department of Materials Science and Engineering, University of California, Berkeley, CA 94720, USA Grace X. Gu, Department of Mechanical Engineering, University of California, Berkeley, CA 94720, USA Address all correspondence to Grace X. Gu at [email protected] (Received 3 February 2019; accepted 28 February 2019)
Abstract Machine learning (ML) has been perceived as a promising tool for the design and discovery of novel materials for a broad range of applications. In this prospective paper, we summarize recent progress in the applications of ML to composite materials modeling and design. An overview of how different types of ML algorithms can be applied to accelerate composite research is presented. This framework is envisioned to revolutionize approaches to design and optimize composites for the next generation of materials with unprecedented properties.
Introduction Designing novel materials with superior tailored properties is the ultimate goal of modern engineering applications.[1–5] In the past few decades, with rapid advances in high-performance parallel computing, materials science, and numerical modeling, many essential properties of materials can now be calculated using simulations with reasonable accuracy. For example, the chemical reactivity and stability of molecules can be estimated using density functional theory (DFT).[6,7] Molecular dynamics (MD) and the finite element method (FEM) can be applied to simulate a wide range of mechanical behaviors of materials at the nano-scale and continuum-scale, respectively.[8,9] Nowadays, simulations of material properties can be performed on a laptop, workstation, or computer cluster, depending on the computational cost. In general, performing simulations to predict the properties of a material is much faster and less expensive than synthesizing, manufacturing, and testing the material in a laboratory. Moreover, simulations offer very precise control over environments and offer more detailed information of material behavior and associated mechanisms under different conditions, many of which cannot or would be very difficult to be observed using experiments. For instance, the stress field of a composite material under fracture and the motions of each molecule in an organic material under loading can be predicated in simulations but are difficult to measure in experiments. It is this reason that many studies in the literature have focused on advancing computational tools and methods to model various types of materials.[10–18] Compared with solely predicting properties of known materials, designing new materials to achieve tunable properties is a more important problem for scientific and engineering purposes. In fact, predicting materials’ properties and designing materials are quite different problems, in which the former is often referred to as a forward modeling problem and the latter
is an inverse design problem. For a forward modeling problem, the structure (e.g., atomic co
Data Loading...
