The phase-field method is a powerful and versatile computational approach for modeling the evolution of microstructures and associated properties for a wide variety of physical, chemical, and biological systems. However, existing high-fidelity phase-field models are inherently computationally expensive, requiring high-performance computing resources and sophisticated numerical integration schemes to achieve a useful degree of accuracy. In this paper, we present a computationally inexpensive, accurate, data-driven surrogate model that directly learns the microstructural evolution of targeted systems by combining phase-field and history-dependent machine-learning techniques. We integrate a statistically representative, low-dimensional description of the microstructure, obtained directly from phase-field simulations, with either a time-series multivariate adaptive regression splines autoregressive algorithm or a long short-term memory neural network. The neural-network-trained surrogate model shows the best performance and accurately predicts the nonlinear microstructure evolution of a two-phase mixture during spinodal decomposition in seconds, without the need for “on-the-fly” solutions of the phase-field equations of motion. We also show that the predictions from our machine-learned surrogate model can be fed directly as an input into a classical high-fidelity phase-field model in order to accelerate the high-fidelity phase-field simulations by leaping in time. Such machine-learned phase-field framework opens a promising path forward to use accelerated phase-field simulations for discovering, understanding, and predicting processing–microstructure–performance relationships.
Advances in machine learning (ML) have enabled the development of interatomic potentials that promise the accuracy of first principles methods and the low-cost, parallel efficiency of empirical potentials. However, ML-based potentials struggle to achieve transferability, i.e., provide consistent accuracy across configurations that differ from those used during training. In order to realize the promise of ML-based potentials, systematic and scalable approaches to generate diverse training sets need to be developed. This work creates a diverse training set for tungsten in an automated manner using an entropy optimization approach. Subsequently, multiple polynomial and neural network potentials are trained on the entropy-optimized dataset. A corresponding set of potentials are trained on an expert-curated dataset for tungsten for comparison. The models trained to the entropy-optimized data exhibited superior transferability compared to the expert-curated models. Furthermore, the models trained to the expert-curated set exhibited a significant decrease in performance when evaluated on out-of-sample configurations.
The local atomic structure, local chemistry, and stoichiometry of grain boundaries control in part the strength and fracture toughness of silicon carbide components. The predictions of the structure and properties of these grain boundaries are generally limited to their ground-state configurations. We investigated the tensile strength behavior of metastable grain boundaries in silicon carbide using high-throughput atomistic simulations combined with machine learning techniques. We analyzed and compared the ∑5 ⟨100⟩{120} and ∑9 ⟨110⟩{122} tilt grain boundary metastable configurations to identify structural and chemical attributes that dominate their tensile strength. We characterized these metastable grain boundaries using a set of microscopic descriptors representing the local grain boundary atomic structure and the local grain boundary stoichiometry and chemical-bound types. We used a boosted regression tree surrogate model for the successful prediction of metastable grain boundary strength as a function of these descriptors. Our results show that the tensile strength of generic (i.e., any random grain boundary from the entire grain boundary population), metastable grain boundaries is primarily dominated by the grain boundary excess free volume, closely followed by the type of structure composing the boundary and the amount of C–C bonds. The 5% strongest metastable grain boundaries have particular characteristics with a low amount of free volume and the highest density of C–C bonds. Our results reveal that the 5% strongest and weakest metastable grain boundaries are most sensitive to the local stoichiometry, regardless of the local atomic structure composing the grain boundary as compared to any other generic metastable grain boundaries. We show that the strongest and weakest metastable grain boundary configurations can be identified as specific regions in a low-dimensional-representation space of their microscopic descriptors. Taken together, these findings showcase the effectiveness and validity of using a low-dimensional representation of the grain boundary structure and machine-learned surrogate models to rapidly assess metastable grain boundary strength without the need to perform actual tensile simulations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.