On-lattice voxelated convolutional neural networks for prediction of phase diagrams and diffusion barriers in cubic alloys
Authors:
Seyedeh Mohadeseh Taheri-Mousavi,
Seyed Sina Moeini-Ardakani,
Ryan W. Penny,
Ju Li,
A. John Hart
Abstract:
Cluster expansion approximates an on-lattice potential with polynomial regression. We show that using a convolutional neural network (CNN) instead leads to more accurate prediction due to the depth of the network. We construct our CNN potential directly on cubic lattice sites, representing voxels in a 3D image, and refer to our method as the voxelated CNN (VCNN). The convolutional layers automatic…
▽ More
Cluster expansion approximates an on-lattice potential with polynomial regression. We show that using a convolutional neural network (CNN) instead leads to more accurate prediction due to the depth of the network. We construct our CNN potential directly on cubic lattice sites, representing voxels in a 3D image, and refer to our method as the voxelated CNN (VCNN). The convolutional layers automatically integrate interaction terms in the regressor; thus, no explicit definition of clusters is required. As a model system, we combine our VCNN potential with Monte Carlo simulations on a Ni$_{1-x}$Al$_x$ ($x$ < 30%) and predict a disordered-to-ordered phase transition with less than 1 meV/atom error. We also predict the energetic landscape of vacancy diffusion. Classification of formation energy with respect to short-range-ordering of Al alloys around a vacancy reveals that the ordering decreases the probability of Ni diffusion. As the width of our input layer does not depend on the atomic composition, VCNNs can be applied to study alloys with arbitrary numbers of elements and empty lattice sites, without additional computational costs.
△ Less
Submitted 3 March, 2021;
originally announced March 2021.
Highly efficient parallel grand canonical simulations of interstitial-driven deformation-diffusion processes
Authors:
Seyed Sina Moeini-Ardakani,
Seyedeh Mohadeseh Taheri-Mousavi,
Ju Li
Abstract:
Diffusion of interstitial alloying elements like H, O, C, and N in metals and their continuous relocation and interactions with their microstructures have crucial influences on metals properties. However, besides limitations in experimental tools in capturing these mechanisms, the inefficiency of numerical tools also inhibits modeling efforts. Here, we present an efficient framework to perform hyb…
▽ More
Diffusion of interstitial alloying elements like H, O, C, and N in metals and their continuous relocation and interactions with their microstructures have crucial influences on metals properties. However, besides limitations in experimental tools in capturing these mechanisms, the inefficiency of numerical tools also inhibits modeling efforts. Here, we present an efficient framework to perform hybrid grand canonical Monte Carlo and molecular dynamics simulations that allow for parallel insertion/deletion of Monte Carlo moves. A new methodology for calculation of the energy difference at trial moves that can be applied to many-body potentials as well as pair ones is a primary feature of our implementation. We study H diffusion in Fe (ferrite phase) and Ni polycrystalline samples to demonstrate the efficiency and scalability of the algorithm and its application. The computational cost of using our framework for half a million atoms is a factor of 250 less than the cost of using existing libraries.
△ Less
Submitted 13 February, 2021;
originally announced February 2021.
