11 results
Filters
Settings
Now showing1 - 10 of 11
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , TiAlNb alloy interatomic potentials: comparing passive and active machine learning techniques with MTP and DeePMD(Frontiers Media S.A., 2025-07-30); ; ; ; Intermetallic titanium aluminides are interesting for aerospace and automotive applications due to their superior high-temperature mechanical properties. In particular, γ-TiAl-based alloys containing 5–10 at.% Niobium (Nb) have attracted significant attention. Molecular dynamics (MD) simulations can elucidate and optimize these materials, provided that accurate interatomic potentials are available. In this work, we compare active and passive machine learning approaches for developing TiAlNb interatomic potentials using both deep potential molecular dynamics (DeePMD) and the moment tensor potential (MTP) methods. Our comprehensive evaluation encompasses elastic constants, equilibrium volume, lattice parameters, and finite-temperature behavior, as well as simulated tension tests and generalized stacking fault energy calculations to assess the impact of Nb on the thermo-mechanical properties of γ-TiAl and α2-Ti3Al phases. Active learning consistently outperformed passive learning for both methods while requiring only a fraction of the training samples. Notably, active learning with DeePMD yielded a single potential capable of predicting the properties of both phases, whereas MTP exhibited limitations that necessitated separate training for each phase. Although active learning potentials excelled in predicting high-temperature behavior, their room-temperature property predictions were less accurate due to a sample selection bias toward higher temperatures. Overall, our thermomechanical analysis demonstrates that Nb incorporation enhances ductility while simultaneously reducing strength.Publicationtype: Journal Article38 93 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Comparative analysis of ternary TiAlNb interatomic potentials: moment tensor vs. deep learning approaches(Frontiers Media S.A., 2024-10-03); ; ; ; Intermetallic titanium aluminides, leveraging the ordered γ-TiAl phase, attract increasing attention in aerospace and automotive engineering due to their favorable mechanical properties at high temperatures. Of particular interest are γ-TiAl-based alloys with a Niobium (Nb) concentration of 5–10 at.%. It is a key question how to model such ternary alloys at the atomic scale with molecular dynamics (MD) simulations to better understand (and subsequently optimize) the alloys. Here, we present a comparative analysis of ternary TiAlNb interatomic potentials developed by the moment tensor potential (MTP) and deep potential molecular dynamics (DeePMD) methods specifically for the above mentioned critical Nb concentration range. We introduce a novel dataset (TiAlNb dataset) for potential training that establishes a benchmark for the assessment of TiAlNb potentials. The potentials were evaluated through rigorous error analysis and performance metrics, alongside calculations of material properties such as elastic constants, equilibrium volume, and lattice constant. Additionally, we explore finite temperature properties including specific heat and thermal expansion with both potentials. Mechanical behaviors, such as uniaxial tension and the calculation of generalized stacking fault energy, are analyzed to determine the impact of Nb alloying in TiAl-based alloys. Our results indicate that Nb alloying generally enhances the ductility of TiAl-based alloys at the expense of reduced strength, with the notable exception of simulations using DeePMD for the γ-TiAl phase, where this trend does not apply.Publicationtype: Journal Article41 215 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Hygrothermal aging history of amine-epoxy resins: effects on thermo-mechanical properties(Frontiers Media S.A., 2022-03-02); ; ; Epoxy systems are widely used as matrix resins for fiber reinforced polymers (FRP) and, therefore, often have to withstand harsh environmental conditions. Especially in marine and offshore environments, moisture or direct water contact leads to water absorption into the epoxy resin. As a result, the mechanical properties change during application. Since diffusion at room or colder temperatures is slow, industry and academia typically use accelerated aging methods at elevated temperatures for durability prediction. However, as the water-polymer interaction is a complex combination of plasticization, physical aging, and molecular interaction, all of these mechanisms are expected to be affected by the ambient temperature. To reveal the impact of aging time and temperature on the thermo-mechanical properties of an amine-epoxy system, this publication includes various hygrothermal aging conditions, like water bath and relative humidity aging at temperatures ranging from 8°C to 70°C and relative humidity from 20% to 90%. Thus, it is demonstrated via long-term aging, DMTA and FTIR investigations that, e.g., strength, stiffness, strain to failure, and the glass transition temperature (T g ) can differ significantly depending on aging time and temperature. For example, it can be shown that water absorption at cold temperatures leads to the strongest and longest-lasting reduction in strength, although the maximum water absorption amount is lower than at higher temperatures. For the application, this means that strength differences of up to 26% can be obtained, depending on the aging method selected. Furthermore, it can be shown that conventional prediction models, such as Eyring correlation, which consider the mobility of the molecular structure for the prediction of thermo-mechanical properties, can only be used to a limited extent for prediction in hygrothermal aging. The reasons for this are seen to be, in particular, the different characteristics of the water-polymer interactions depending on the aging temperature. While plasticization dominates in cold conditions, relaxation and strong water-molecule bonds predominate in warm conditions.Publicationtype: Journal ArticleScopus© Citations 19 182 865 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Training deep neural networks to reconstruct nanoporous structures from FIB tomography images using synthetic training data(Frontiers Media S.A., 2022-02-28); ; ; ; Focused ion beam (FIB) tomography is a destructive technique used to collect three-dimensional (3D) structural information at a resolution of a few nanometers. For FIB tomography, a material sample is degraded by layer-wise milling. After each layer, the current surface is imaged by a scanning electron microscope (SEM), providing a consecutive series of cross-sections of the three-dimensional material sample. Especially for nanoporous materials, the reconstruction of the 3D microstructure of the material, from the information collected during FIB tomography, is impaired by the so-called shine-through effect. This effect prevents a unique mapping between voxel intensity values and material phase (e.g., solid or void). It often substantially reduces the accuracy of conventional methods for image segmentation. Here we demonstrate how machine learning can be used to tackle this problem. A bottleneck in doing so is the availability of sufficient training data. To overcome this problem, we present a novel approach to generate synthetic training data in the form of FIB-SEM images generated by Monte Carlo simulations. Based on this approach, we compare the performance of different machine learning architectures for segmenting FIB tomography data of nanoporous materials. We demonstrate that two-dimensional (2D) convolutional neural network (CNN) architectures processing a group of adjacent slices as input data as well as 3D CNN perform best and can enhance the segmentation performance significantly.Publicationtype: Journal ArticleScopus© Citations 5 156 895 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Quantifying the high-temperature separation behavior of lamellar interfaces in γ-titanium aluminide under tensile loading by molecular dynamicsγ-titanium aluminide (TiAl) alloys with fully lamellar microstructure possess excellent properties for high-temperature applications. Such fully lamellar microstructure has interfaces at different length scales. The separation behavior of the lamellae at these interfaces is crucial for the mechanical properties of the whole material. Unfortunately, quantifying it by experiments is difficult. Therefore, we use molecular dynamics (MD) simulations to this end. Specifically, we study the high-temperature separation behavior under tensile loading of the four different kinds of lamellar interfaces appearing in TiAl, namely, the (Formula presented.), (Formula presented.), (Formula presented.), and (Formula presented.) interfaces. In our simulations, we use two different atomistic interface models, a defect-free (Type-1) model and a model with preexisting voids (Type-2). Clearly, the latter is more physical but studying the former also helps to understand the role of defects. Our simulation results show that among the four interfaces studied, the (Formula presented.) interface possesses the highest yield strength, followed by the (Formula presented.), (Formula presented.), and (Formula presented.) interfaces. For Type-1 models, our simulations reveal failure at the interface for all γ/γ interfaces but not for the (Formula presented.) interface. By contrast, for Type-2 models, we observe for all the four interfaces failure at the interface. Our atomistic simulations provide important data to define the parameters of traction–separation laws and cohesive zone models, which can be used in the framework of continuum mechanical modeling of TiAl. Temperature-dependent model parameters were identified, and the complete traction–separation behavior was established, in which interface elasticity, interface plasticity, and interface damage could be distinguished. By carefully eliminating the contribution of bulk deformation from the interface behavior, we were able to quantify the contribution of interface plasticity and interface damage, which can also be related to the dislocation evolution and void nucleation in the atomistic simulations.Publicationtype: Journal ArticleScopus© Citations 5 162 303 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Machine learning and data mining in materials science : editorial(Frontiers Media S.A., 2020-02-28); ; ; Publicationtype: Editorial39 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Computation of thickness and mechanical properties of interconnected structures: accuracy, deviations, and approaches for correctionIdentifying local thickness information of fibrous or highly porous structures is challenging. The analysis of tomography data calls for computationally fast, robust, and accurate algorithms. This work systematically investigates systematic errors in the thickness computation and the impact of observed deviations on the predicted mechanical properties using a set of 16 model structures with varying ligament shape and solid fraction. Strongly concave, cylindrical, and convex shaped ligaments organized in a diamond structure are analyzed. The predicted macroscopic mechanical properties represent a highly sensitive measure for systematic errors in the computed geometry. Therefore, the quality of proposed correction methods is assessed via FEM beam models that can be automatically generated from the measured data and allow an efficient prediction of the mechanical properties. The results show that low voxel resolutions can lead to an overprediction of up to 30% in the Young's modulus. A model scanned with a resolution of 200 voxels per unit cell edge (8M voxels) reaches an accuracy of a few percent. Analyzing models of this resolution with the Euclidean distance transformation showed an underprediction of up to 20% for highly concave shapes whereas cylindrical and slightly convex shapes are determined at high accuracy. For the Thickness algorithm, the Young's modulus and yield strength are overpredicted by up to 100% for highly concave ligament shapes. A proposed Smallest Ellipse approach corrects the Thickness data and reduces this error to 20%. It can be used as input for a further robust correction of the Thickness data using an artificial neural network. This approach is highly accurate with remnant errors in the predicted mechanical properties of only a few percent. Furthermore, the data from the FEM beam models are compared to results from FEM solid models providing deeper insights toward further developments on nodal corrections for FEM beam models. As expected, the FEM beam models show an increasing overprediction of the compliance with increasing solid fraction. As an unexpected result, the mechanical strength can however be underpredicted or overpredicted, depending on the ligament shape. Therefore, a nodal correction is needed that solves contradicting tasks in terms of stiffness and strength.Publicationtype: Journal ArticleScopus© Citations 15 309 203 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , A review of the application of machine learning and data mining approaches in continuum materials mechanicsMachine learning tools represent key enablers for empowering material scientists and engineers to accelerate the development of novel materials, processes and techniques. One of the aims of using such approaches in the field of materials science is to achieve high-throughput identification and quantification of essential features along the process-structure-property-performance chain. In this contribution, machine learning and statistical learning approaches are reviewed in terms of their successful application to specific problems in the field of continuum materials mechanics. They are categorized with respect to their type of task designated to be either descriptive, predictive or prescriptive; thus to ultimately achieve identification, prediction or even optimization of essential characteristics. The respective choice of the most appropriate machine learning approach highly depends on the specific use-case, type of material, kind of data involved, spatial and temporal scales, formats, and desired knowledge gain as well as affordable computational costs. Different examples are reviewed involving case-by-case dependent application of different types of artificial neural networks and other data-driven approaches such as support vector machines, decision trees and random forests as well as Bayesian learning, and model order reduction procedures such as principal component analysis, among others. These techniques are applied to accelerate the identification of material parameters or salient features for materials characterization, to support rapid design and optimization of novel materials or manufacturing methods, to improve and correct complex measurement devices, or to better understand and predict fatigue behavior, among other examples. Besides experimentally obtained datasets, numerous studies draw required information from simulation-based data mining. Altogether, it is shown that experiment- and simulation-based data mining in combination with machine leaning tools provide exceptional opportunities to enable highly reliant identification of fundamental interrelations within materials for characterization and optimization in a scale-bridging manner. Potentials of further utilizing applied machine learning in materials science and empowering significant acceleration of knowledge output are pointed out.Publicationtype: Journal ArticleScopus© Citations 240 432 709 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , General multi-fidelity framework for training artificial neural networks with computational models(2019-04-17); ; Training of artificial neural networks (ANNs) relies on the availability of training data. If ANNs have to be trained to predict or control the behavior of complex physical systems, often not enough real-word training data are available, for example, because experiments or measurements are too expensive, time-consuming or dangerous. In this case, generating training data by way of realistic computational simulations is a viable and often the only promising alternative. Doing so can, however, be associated with a significant and often even prohibitive computational cost, which forms a serious bottleneck for the application of machine learning to complex physical systems. To overcome this problem, we propose in this paper a both systematic and general approach. It uses cheap low-fidelity computational models to start the training of the ANN and gradually switches to higher-fidelity training data as the training of the ANN progresses. We demonstrate the benefits of this strategy using examples from structural and materials mechanics. We demonstrate that in these examples the multi-fidelity strategy introduced herein can reduce the total computational cost–compared to simple brute-force training of ANNs–by a half up to one order of magnitude. This multi-fidelity strategy can thus be hoped to become a powerful and versatile tool for the future combination of computational simulations and artificial intelligence, in particular in areas such as structural and materials mechanics.Publicationtype: Journal ArticleScopus© Citations 25 472 422 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Data science based mg corrosion engineering(2019-03-08); ; ; ; Magnesium exhibits a high potential for a variety of applications in areas such as transport, energy and medicine. However, untreated magnesium alloys are prone to corrosion, restricting their practical application. Therefore, it is necessary to develop new approaches that can prevent or control corrosion and degradation processes in order to adapt to the specific needs of the application. One potential solution is using corrosion inhibitors which are capable of drastically reducing the degradation rate as a result of interactions with the metal surface or components of the corrosive medium. As the sheer number of potential dissolution modulators makes it impossible to obtain a detailed atomistic understanding of the inhibition mechanisms for each additive, other measures for inhibition prediction are required. For this purpose, a concept is presented that combines corrosion experiments, machine learning, data mining, density functional theory calculations and molecular dynamics to estimate corrosion inhibition properties of still untested molecules. Concomitantly, this approach will provide a deeper understanding of the fundamental mechanisms behind the prevention of corrosion events in magnesium-based materials and enables more accurate continuum corrosion simulations. The presented concept facilitates the search for molecules with a positive or negative effect on the inhibition efficiency and could thus significantly contribute to the better control of magnesium / electrolyte interface properties. © 2019 Würger, Feiler, Musil, Feldbauer, Höche, Lamaka, Zheludkevich and Meißner.Publicationtype: Journal ArticleScopus© Citations 36 650 392