Machine learning for metallurgy V: A neural-network potential for zirconium

Liyanage, Manura; Reith, D.; Eyert, V.

doi:10.1103/physrevmaterials.6.063804

Cited by 12 publications

(4 citation statements)

References 92 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The lowest energy surface predicted by the DP model is surf-0001, consistent with the results of DFT and Ref. [51], and the DP model has high accuracy for surfaces (surf-10 11) and (surf- 11 22). The DP model shows that the surface energy is consistent with the DFT result, and can also predict the energy order of the surface of zirconium.…”

Section: Basic Physical Propertiessupporting

confidence: 85%

“…For α-Zr, the lattice parameters (a and c), bulk modulus (B), elastic constants (C 11 , C 12 , C 13 , C 33, and C 44 ), and surface energy (surf-0001, surf-10 11, surf-11 22) are listed in Table 2. DFT NNP [51] EAM [50] a (HCP) 3.241 3.239 3.231 [57] 3.231 [58] 3.231 3.223 c/a 1.597 1.599 1.603 [57] 1.598 [58] 1.602 [59] 147.78 [58] 141.5 157.3 C 12 69.75 62.24 67.2 [59] 70.66 [58] 67.2 98.9 C 13 69.37 65.54 64.6 [59] 65.51 [58] 67.2 77.6 C 33 159.33 170.01 172.5 [59] 168.29 [58] 168.7 181.4 C 44 26.24 24.61 36.3 [59] 25.41 [58] 28.9 30.3 Surf-0001 1589.45 1638.12 2000 [60] 1580 [51] 1593.8 1127 Surf- 10 10 1657.40 1662.25 2050 [61] 1664 [51] 1631.9 1273 Surf- 11 22 1724.48 1718. 34 In general, the results from the DP model are in better agreement with the DFT results for these properties in comparison with those from EAM [50] and demonstrate that its performance is comparable to NNP's [51] in terms of DFT results.…”

Section: Basic Physical Propertiesmentioning

confidence: 99%

See 1 more Smart Citation

Properties of radiation defects and threshold energy of displacement in zirconium hydride obtained by new deep-learning potential

Wang 王,

Tang 唐,

Jiang 蒋

et al. 2024

Chinese Phys. B

View full text Add to dashboard Cite

Zirconium hydride (ZrH2) is an ideal neutron moderator material. However, radiation effects significantly alter its properties, subsequently impacting its behavior and the lifespan of the reactor. The threshold energy of displacement is an important quantity of the number of radiation defects produced, which helps us to predict the evolution of radiation defects in ZrH2. Molecular dynamics (MD) and ab initio molecular dynamics (AIMD) are two main methods for calculating the threshold energy of displacement. MD simulations with empirical potentials often fail to accurately depict the transitional states that lattice atoms must surpass to reach an interstitial state. Additionally, the AIMD method does not afford large-scale calculation, posing a computational challenge beyond the scope of density functional theory simulations. Machine learning potentials are renowned for their high accuracy and efficiency, making them an increasingly preferred choice for molecular dynamics simulations. In this work, we developed an accurate potential energy model for the ZrH2 system by using the deep-potential (DP) method. The DP model has a high degree of agreement with first-principles calculations for the typic defect energies and mechanical properties of the ZrH2 system, including the basic bulk properties, formation energy of point defects, as well as diffusion behavior of hydrogen and zirconium. By integrating the DP model with Ziegler-Biersack-Littmark (ZBL) potential, we have predicted the threshold energies of displacement of zirconium and hydrogen in ε-ZrH2.

show abstract

Section: Basic Physical Propertiessupporting

confidence: 85%

Section: Basic Physical Propertiesmentioning

confidence: 99%

Properties of radiation defects and threshold energy of displacement in zirconium hydride obtained by new deep-learning potential

Wang 王,

Tang 唐,

Jiang 蒋

et al. 2024

Chinese Phys. B

View full text Add to dashboard Cite

show abstract

“…[173][174][175][176] Machine learning based potentials (MLPs) have recently emerged as a promising method of accurately modelling the properties and dynamics of several systems and reactions. [177][178][179][180][181][182][183][184][185] As a result, MLPs are iteratively trained to 'learn' the potential energy surface of the system (based on limited DFT data) and can serve as a viable substitute for QM/MM and QM/QM schemes and apply to systems of arbitrary size at almost DFT level of accuracy. Consequently, MLPs can help to perform high-throughput screening of several catalyst configurations in multiple zeolites 179 if desired, in addition to being able to model system dynamics.…”

Section: Frontier Dalton Transactionsmentioning

confidence: 99%

Zeolite encapsulated organometallic complexes as model catalysts

Iaia,

Soyemi,

Szilvási

et al. 2023

Dalton Trans.

View full text Add to dashboard Cite

show abstract

“…Zong et al adopted a kernel ridge regression model to capture the martensitic phase transformations in Zr, but this resulted in a low melting point . Nitol et al modeled the transformations among α, β, and ω phases of Zr and Ti using artificial neural networks as regression models, and Liyanage et al developed an artificial neural network potential for hcp Zr that focused on extended defect properties. These sophisticated neural network architectures can possess up to tens of thousands of parameters, fitted on over ten thousand DFT configurations.…”

Section: Introductionmentioning

confidence: 99%

A Set of Moment Tensor Potentials for Zirconium with Increasing Complexity

Luo,

Meziere,

Samolyuk

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

Machine learning force fields (MLFFs) are an increasingly popular choice for atomistic simulations due to their high fidelity and improvable nature. Here we propose a hybrid small-cell approach that combines attributes of both offline and active learning to systematically expand a quantum-mechanical (QM) database while constructing MLFFs with increasing model complexity. Our MLFFs employ the moment tensor potential formalism. During this process, we quantitatively assessed the structural properties, elastic properties, dimer potential energies, melting temperatures, phase stability, point defect formation energies, point defect migration energies, free surface energies, and generalized stacking fault (GSF) energies of Zr as predicted by our MLFFs. Unsurprisingly, the model complexity has a positive correlation with prediction accuracy. We also find that the MLFFs were able to predict the properties of out-of-sample configurations without directly including these specific configurations in the training dataset. Additionally, we generated 100 MLFFs of high complexity (1513 parameters each) that reached different local optima during training. Their predictions cluster around the benchmark DFT values, but subtle physical features such as the location of local minima on the GSF energy surface are washed out by statistical noise.

show abstract

Machine learning for metallurgy V: A neural-network potential for zirconium

Cited by 12 publications

References 92 publications

Properties of radiation defects and threshold energy of displacement in zirconium hydride obtained by new deep-learning potential

Properties of radiation defects and threshold energy of displacement in zirconium hydride obtained by new deep-learning potential

Zeolite encapsulated organometallic complexes as model catalysts

A Set of Moment Tensor Potentials for Zirconium with Increasing Complexity

Contact Info

Product

Resources

About