Publications

Results 26–50 of 225

Search results

Jump to search filters

Accelerating Multiscale Materials Modeling with Machine Learning

Modine, N.A.; Stephens, John A.; Swiler, Laura P.; Thompson, Aidan P.; Vogel, Dayton J.; Cangi, Attila; Feilder, Lenz; Rajamanickam, Sivasankaran R.

The focus of this project is to accelerate and transform the workflow of multiscale materials modeling by developing an integrated toolchain seamlessly combining DFT, SNAP, LAMMPS, (shown in Figure 1-1) and a machine-learning (ML) model that will more efficiently extract information from a smaller set of first-principles calculations. Our ML model enables us to accelerate first-principles data generation by interpolating existing high fidelity data, and extend the simulation scale by extrapolating high fidelity data (102 atoms) to the mesoscale (104 atoms). It encodes the underlying physics of atomic interactions on the microscopic scale by adapting a variety of ML techniques such as deep neural networks (DNNs), and graph neural networks (GNNs). We developed a new surrogate model for density functional theory using deep neural networks. The developed ML surrogate is demonstrated in a workflow to generate accurate band energies, total energies, and density of the 298K and 933K Aluminum systems. Furthermore, the models can be used to predict the quantities of interest for systems with more number of atoms than the training data set. We have demonstrated that the ML model can be used to compute the quantities of interest for systems with 100,000 Al atoms. When compared with 2000 Al system the new surrogate model is as accurate as DFT, but three orders of magnitude faster. We also explored optimal experimental design techniques to choose the training data and novel Graph Neural Networks to train on smaller data sets. These are promising methods that need to be explored in the future.

More Details
TYPE SAND Report YEAR 2022
DOIOSTI

Molecular Dynamics of High Pressure Tin Phases I: Strength and deformation evaluations of empirical potentials [Slides]

Lane, James M.; Cusentino, Mary A.; Nebgen, Ben; Barros, Kipton M.; Shimanek, John D.; Allen, Alice; Thompson, Aidan P.; Fensin, Saryu J.

Multi-phase problems have so many more unknowns, we’d like to have a tool to constrain some open questions related to microstructure and twin & dislocation behavior. We want an atomistic scale perspective on aspects of strength. Some multi-scale questions accessible to atomistic study: What lattice-specific behavior influences dislocation production/mobility and/or twinning? Do the phase transformations wipe-out, modify or preserve grain size and orientation? Does plastic strain reset at phase transition? If so under what conditions? Tin is the material chosen for the effort because it is non-hazardous and has multiple accessible solid phases at relatively low pressures.

More Details
TYPE Conference Presenation YEAR 2022
DOIOSTI

Performant implementation of the atomic cluster expansion (PACE) and application to copper and silicon

npj Computational Materials

Lysogorskiy, Yury; Van Der Oord, Van; Bochkarev, Anton; Menon, Sarath; Rinaldi, Matteo; Hammerschmidt, Thomas; Mrovec, Matous; Thompson, Aidan P.; Csanyi, Gabor; Ortner, Christoph; Drautz, Ralf

The atomic cluster expansion is a general polynomial expansion of the atomic energy in multi-atom basis functions. Here we implement the atomic cluster expansion in the performant C++ code PACE that is suitable for use in large-scale atomistic simulations. We briefly review the atomic cluster expansion and give detailed expressions for energies and forces as well as efficient algorithms for their evaluation. We demonstrate that the atomic cluster expansion as implemented in PACE shifts a previously established Pareto front for machine learning interatomic potentials toward faster and more accurate calculations. Moreover, general purpose parameterizations are presented for copper and silicon and evaluated in detail. We show that the Cu and Si potentials significantly improve on the best available potentials for highly accurate large-scale atomistic simulations.

More Details
TYPE Journal Article YEAR 2021
DOIOSTIScopus

$\mathrm{LAMMPS}$ - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales

Computer Physics Communications

Thompson, Aidan P.; Aktulga, H.M.; Berger, Richard; Bolintineanu, Dan S.; Brown, W.M.; Crozier, Paul C.; In 'T Veld, Pieter J.; Kohlmeyer, Axel; Moore, Stan G.; Nguyen, Trung D.; Shan, Ray; Stevens, Mark J.; Tranchida, Julien; Trott, Christian R.; Plimpton, Steven J.

Since the classical molecular dynamics simulator LAMMPS was released as an open source code in 2004, it has become a widely-used tool for particle-based modeling of materials at length scales ranging from atomic to mesoscale to continuum. Reasons for its popularity are that it provides a wide variety of particle interaction models for different materials, that it runs on any platform from a single CPU core to the largest supercomputers with accelerators, and that it gives users control over simulation details, either via the input script or by adding code for new interatomic potentials, constraints, diagnostics, or other features needed for their models. As a result, hundreds of people have contributed new capabilities to LAMMPS and it has grown from fifty thousand lines of code in 2004 to a million lines today. In this paper several of the fundamental algorithms used in LAMMPS are described along with the design strategies which have made it flexible for both users and developers. We also highlight some capabilities recently added to the code which were enabled by this flexibility, including dynamic load balancing, on-the-fly visualization, magnetic spin dynamics models, and quantum-accuracy machine learning interatomic potentials.

More Details
TYPE Journal Article YEAR 2021
DOIOSTI
Results 26–50 of 225
Results 26–50 of 225