Neural Network Analysis of Dynamic Fracture in a Layered Material
- PDF / 962,046 Bytes
- 9 Pages / 432 x 648 pts Page_size
- 10 Downloads / 257 Views
MRS Advances © 2018 Materials Research Society DOI: 10.1557/adv.2018.673
Neural Network Analysis of Dynamic Fracture in a Layered Material Pankaj Rajak1,2, Rajiv K. Kalia2,3,4,5, Aiichiro Nakano2,3,4,5,6, Priya Vashishta2,3,4,5 1
Argonne Leadership Computing Facility, Argonne National Laboratory, Argonne, IL 60439, USA; Collaboratory for Advanced Computing and Simulations, 3Department of Physics, 4Department of Computer Science, 5Department of Chemical Engineering and Material Science, 6Department of Biological Sciences, University of Southern California, Los Angeles, CA 90089 2
Abstract
Dynamic fracture of a two-dimensional MoWSe2 membrane is studied with molecular dynamics (MD) simulation. The system consists of a random distribution of WSe2 patches in a pre-cracked matrix of MoSe2. Under strain, the system shows toughening due to crack branching, crack closure and strain-induced structural phase transformation from 2H to 1T crystal structures. Different structures generated during MD simulation are analyzed using a three-layer, feed-forward neural network (NN) model. A training data set of 36,000 atoms is created where each atom is represented by a 50-dimension feature vector consisting of radial and angular symmetry functions. Hyper parameters of the symmetry functions and network architecture are tuned to minimize model complexity with high predictive power using feature learning, which shows an increase in model accuracy from 67% to 95%. The NN model classifies each atom in one of the six phases which are either as transition metal or chalcogen atoms in 2H phase, 1T phase and defects. Further t-SNE analyses of learned representation of these phases in the hidden layers of the NN model show that separation of all phases become clearer in the third layer than in layers 1 and 2.
Downloaded from https://www.cambridge.org/core. Iowa State University Library, on 12 Jan 2019 at 15:27:53, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1557/adv.2018.673
INTRODUCTION Molecular dynamics (MD) simulation of various physical and chemical phenomena of materials requires complex data analysis of the simulation results so as to identify different phases, chemical reaction and defects generated during the simulation. For example, during nanoindentation simulation, plastic deformation occurs inside the material due to dislocation nucleation.[1] Similarly, under stress, materials like SiC, AlN show phase transformation.[2, 3] Identification of different phases and defects generated during MD simulation requires complex structural analysis. These analyses range from calculation of nearest neighbours to shortest circuit analysis of atoms. In general, there is no single evaluation technique that can be applied to all simulation data since these analyses are system dependent. However, we can observe that structures generated in MD simulations are complex functional forms of their local environment. Hence, we can create a machine learning (ML) model that c
Data Loading...
