Settings

Now showing1 - 5 of 5
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Publication, Access status: with files ,
    Development and application of machine-learned interatomic potentials for the study of high-temperature TiAlNb alloys
    (2026)
    Chandran, Anju
    ;
    Cyron, Christian J.
    The lack of accurate and efficient interatomic potentials for technologically vital TiAlNb alloys has hindered their computational design and optimization for aerospace and energy applications. The thesis addresses the gap through a hierarchical investigation that begins by establishing the limitations of a conventional classical interatomic potential for the system. To address these limitations, the work then details the development and application of two state of the art machine learning interatomic potentials (ML-IAPs), the neural network-based deep potential molecular dynamics (DeePMD) and the linear regression-based moment tensor potentials (MTP), to investigate the influence of Niobium on the thermomechanical properties of TiAlNb alloys. Furthermore, the efficiency of the interatomic potential development process itself is systematically evaluated by contrasting a traditional passive learning approach with a data-efficient active learning strategy. The results demonstrate that the developed ML-IAPs achieve near-ab initio accuracy, revealing a clear trade-off between the higher fidelity of DeePMD and the greater computational and data efficiency of MTP. Critically, the active learning strategy is shown to reduce the required number of ab initio calculations by an order of magnitude without sacrificing model accuracy. Application of these potentials confirms that Nb addition enhances ductility while decreasing the strength of TiAl phases, providing a clear mechanistic link through simulated tension tests and the analysis of stacking fault energies. This research thus makes three principal contributions: it delivers a quantitative and mechanistic understanding of Niobium's influence on the thermomechanical properties of TiAl-based alloys; it provides a set of rigorously validated, high-fidelity machine learning (ML) potentials for the TiAlNb system; and it presents a detailed methodological guide for the efficient construction of such potentials via active learning. The benchmark datasets are made publicly available to accelerate the data-driven design of high-performance materials.
    Publicationtype: Doctoral Thesis
      63  27
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Publication, Access status: with files ,
    Multiscale investigations of mechanobiology: in vitro observations and computational modeling
    (2025)
    Paukner, Daniel
    ;
    Cyron, Christian J.
    Mechanobiology is an inherently multiscale field of study that explores the relationship between mechanical forces and biological processes at various levels, from molecular and cellular to tissue and organ levels. This work presents different aspects of this multiscale system, ranging from cell migration to mechanical testing of tissue equivalents to computational modeling of tissue adaptation. The results help identify key macroscopic quantities and (sub)cellular mechanisms essential in regulating biological processes on different length and time scales in response to mechanical perturbations.
    Publicationtype: Doctoral Thesis
      77  62
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Publication, Access status: with files ,
    3D reconstruction of FIB tomography data using machine learning
    (2024)
    Sardhara, Trushal
    ;
    Cyron, Christian J.
    This thesis introduces methods for accurate 3D reconstruction of FIB tomography data using machine learning. It addresses challenges in obtaining large datasets and ground truth values, proposing a method for virtual FIB tomography data generation and an isotropy-based validation method. ML-based segmentation methods tackle issues in BSE images, like the shine-through effect and image intensity ambiguities. Integrating the slice repositioning method with image inpainting resolves inconsistencies in slice thicknesses in FIB tomography data. Additionally, a multimodal machine learning approach using multivoltage images further enhances 3D reconstruction accuracy.
    Publicationtype: Doctoral Thesis
      150  261
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Publication, Access status: with files ,
    Peridynamic models in multi-physics applications on bounded and unbounded domains
    (2024)
    Hermann, Alexander
    ;
    Cyron, Christian J.
    The following dissertation concerns with elevating the role of nonlocal peridynamic continuum models in real-world engineering and material applications by establishing efficient and robust numerical techniques and providing comprehensive model descriptions that aim to accurately represent the underlying physical properties. A number of versatile numerical discretization schemes are thoroughly evaluated for both bond-based peridynamic models in brittle fracture modeling as well as nonlocal diffusion-type models for up to three-dimensional problems. These schemes can adaptively refine the computational grid around areas of interest, are asymptotically compatible, computationally efficient, and easy to implement. Efficient solution schemes are further provided for unbounded domain problems, based on residual-free exponential basis functions, which may be used in a collocation scheme to obtain Dirichlet-type boundary conditions to approximate the far-field domain solution over a truncating boundary. The feasibility and improved capability of the proposed numerical schemes are demonstrated on the complex multi-physics simulation of the biodegradation of magnesium-based bone implant screws under physiological conditions on (nearly) unbounded three-dimensional domains, utilizing the standard bi-material peridynamic corrosion model with an efficient constitutive model of the corrosion kinetics that directly links to the experimentally easy-to-measure macroscopic volume loss. In order to augment the capabilities of standard peridynamic models towards the description of multi-ionic mass transport of electromigration, a generalized peridynamic framework of nonlocal Nernst-Planck-Poisson equations is proposed, which effectively models the diffusion-electromigration-reaction and moving boundary problems of corrosion models with arbitrary constitutive corrosion kinetics.
    Publicationtype: Doctoral Thesis
      133  111
  • Some of the metrics are blocked by your 
    consent settings
    Item type:Publication, Access status: with files ,
    Computational micromechanics of matrix-inclusion composites
    (2019)
    Schneider, Konrad
    ;
    Cyron, Christian J.
    This thesis is concerned with aspects of the computational retrievement of macroscopic material responses and investigations of the microscopic deformation behavior of random matrix-inclusion composites. To this end, the focus lies on the representative volume element (RVE) approach in the scope of finite element analyses and its applicability and integration into engineering practice. The artificial random generation of the microstructural geometry, in a periodic and non-periodic manner, is established by studying and enhancing existing methods such as the random sequential adsorption method or collective rearrangement methods for cubic RVEs. Based on distance computations between two inclusions, the non-overlap requirement, which is characteristic for matrix-inclusion composites, is ensured for ellipsoidal, capsular, cylindrical and convex polyhedral inclusion shapes. To comply with the state-of-the-art approach that utilizes fully periodic RVEs featuring a periodic topology, a periodic discretization and periodic boundary conditions (PBC), a new algorithm that automatically generates periodic meshes is developed. The algorithm systematically combines various meshing tools in an efficient and robust way. Special emphasis is placed on the discretization procedure to maintain a high quality mesh with as few elements as possible, thus, manageable for computer simulations applicable to high volume concentrations, high number of inclusions and complex inclusion geometries. Examples elucidate the ability of the proposed approach to efficiently generate large RVEs with a high number of anisotropic inclusions but still maintaining a high quality mesh and a low number of elements. Since the generation of such a state-of-the-art RVE can significantly reduce the overall efficiency in multiscale finite element simulations, a benchmark study that investigates possible relaxations to this cumbersome task is performed. In particular, the RVE size, periodic and non-periodic RVE topologies, different discretization variants and various types of boundary conditions that either require periodicity or do not require periodicity of the underlying discretization are benchmarked. Approximate periodic boundary conditions are discussed in detail. The benchmark study proves that a fully periodic topology and mesh discretization with periodic boundary conditions is not necessary in order to identify effective macroscopic material parameters for technologically relevant composites. Eventually, the gathered methodologies are applied to model a modern high-tech polymer nanocomposite. By utilizing a sophisticated material model for the polymeric matrix, the entropic and energetic deformation regime can be represented in a physically meaningful way. Furthermore, the explicit consideration of the matrix-inclusion interface failure via a cohesive zone model accounts for the non-negligible surface effects. Comparisons with experiments underline the predictive character of the modeling approach and allow for investigations of local deformation mechanisms.
    Publicationtype: Doctoral Thesis
      666  358