Search for: All records

Abstract Graph neural networks (GNNs) have excelled in predictive modeling for both crystals and molecules, owing to the expressiveness of graph representations. High-entropy alloys (HEAs), however, lack chemical long-range order, limiting the applicability of current graph representations. To overcome this challenge, we propose a representation of HEAs as a collection of local environment graphs. Based on this representation, we introduce the LESets machine learning model, an accurate, interpretable GNN for HEA property prediction. We demonstrate the accuracy of LESets in modeling the mechanical properties of quaternary HEAs. Through analyses and interpretation, we further extract insights into the modeling and design of HEAs. In a broader sense, LESets extends the potential applicability of GNNs to disordered materials with combinatorial complexity formed by diverse constituents and their flexible configurations. 

Abstract We introduce a novel digital twin framework for predictive maintenance of physical systems with long term operations. Using monitoring tire health as an application, we show how the digital twin framework is used to enhance automotive safety and efficiency, while overcoming technical challenges using a three-step approach. Firstly, for managing the data complexity over a long operation span, we employ data reduction techniques to concisely represent physical tires using historical performance and usage data. Relying on this data, for fast real-time prediction, we train a transformer-based model offline on our concise dataset to predict future tire health over time, represented as Remaining Casing Potential (RCP). Based on our architecture, our model quantifies both epistemic and aleatoric uncertainty, providing reliable confidence intervals around predicted RCP. Secondly, we incorporate real-time data by updating the predictive model in the digital twin framework, ensuring its accuracy throughout its life span with the aid of hybrid modeling and the use of a discrepancy function. Thirdly, to assist decision making in predictive maintenance, we implement a Tire State Decision Algorithm, which strategically determines the optimal timing for tire replacement based on RCP forecasted by our transformer model. This three-step approach ensures that our digital twin not only accurately predicts the health of a system, but also continually refines its digital representation and makes predictive maintenance decisions for removal from service. Our proposed digital twin framework embodies a physical system accurately and leverages big data and machine learning for predictive maintenance, model update and decision making. 

At fast-spreading centers, faults develop within the axial summit trough (AST; 0 to 250 m around the axis) primarily by diking-induced deformation originating from the axial magma lens (AML). The formation of the prominent abyssal-hill-bounding faults beyond the axial high (>2,000 m) is typically associated with the unbending of the lithosphere as it cools and spreads away from the AST. The presence of faults is rarely mapped between these two thermally distinct zones, where the lithosphere is still too hot for the faults to be linked with the process of thermal cooling and outside of the AST where the accretional diking process dominates the ridge axis. Here, we reveal a remarkable vertical alignment between the distinct morphological features of the magma body and the orientation of these faults, by comparison of 3-D seismic imagery and bathymetry data collected at the East Pacific Rise (EPR) 9°50’N. The spatial coincidence and asymmetric nucleation mode of the mapped faults represent the most direct evidence for magmatically induced faulting near the ridge axis, providing pathways for hydrothermalism and magma emplacement, helping to build the crust outside of the AST. The high-resolution seafloor and subsurface images also enable revised tectonic strain estimates, which shows that the near-axis tectonic component of seafloor spreading at the EPR is an order of magnitude smaller than previously thought with close to negligible contribution of lava buried faults to spreading. 

Recent advances in machine learning (ML) are expediting materials discovery and design. One significant challenge facing ML for materials is the expansive combinatorial space of potential materials formed by diverse constituents and their flexible configurations. This complexity is particularly evident in molecular mixtures, a frequently explored space for materials, such as battery electrolytes. Owing to the complex structures of molecules and the sequence-independent nature of mixtures, conventional ML methods have difficulties in modeling such systems. Here, we present MolSets, a specialized ML model for molecular mixtures, to overcome the difficulties. Representing individual molecules as graphs and their mixture as a set, MolSets leverages a graph neural network and the deep sets architecture to extract information at the molecular level and aggregate it at the mixture level, thus addressing local complexity while retaining global flexibility. We demonstrate the efficacy of MolSets in predicting the conductivity of lithium battery electrolytes and highlight its benefits in the virtual screening of the combinatorial chemical space. Published by the American Physical Society2024