### Browsing by Subject "Computational and Applied Mechanics"

Now showing 1 - 3 of 3

###### Results Per Page

###### Sort Options

- ItemOpen AccessComplex fluid dynamical computations via the Finite Volume Method(2018) Rahantamialisoa, Faniry Nadia Zazaravaka; Chinyoka, TiriNumerical simulations of the complex flows of viscoelastic fluids are investigated. The viscoelastic fluids are modelled, primarily, via the Johnson-Segalman constitutive model. Our Numerical approach is based on finite volume method, based on the Johnson-Segalman constitutive model and implemented on the OpenFOAM® platform. The Johnson-Segalman model also easily reduces to the Oldroyd-B model under certain conditions of the material parameters. Since computations using the Oldroyd-B model have been extensively documented in the literature, we take advantage of the mathematical modelling connection between the Johnson-Segalman and Oldroyd-B models to validate the accuracy of our Johnson-Segalman solver via reduction to the Oldroyd-B model. Numerical validation of our results is conducted via the most commonly used benchmark problems. The final aim of our work is to assess the viability and efficiency of our numerical solver via an investigation into the complex fluid dynamical processes associated with shear banding.
- ItemOpen AccessDevelopment of a Micromorphic (Multiscale) Material Model aimed at Cardiac Tissue Mechanics(2020) Dollery, Devin; Skatulla, SebastianComputational cardiac mechanics has historically relied on classical continuum models; however, classical models amalgamate the behaviour of a material's micro-constituents, and thus only approximate the macroscopically observable material behaviour as a purely averaged response that originated on micro-structural levels. As such, classical models do not directly and independently address the response of the cardiac tissue (myocardium) components, such as the muscle fibres (myocytes) or the hierarchically organized cytoskeleton. Multiscale continuum models have developed over time to account for some of the micro-architecture of a material, and allow for additional degrees of freedom in the continuum over classical models. The micromorphic continuum [15] is a multiscale model that contains additional degrees of freedom which lend themselves to the description of fibres, referred to as micro-directors. The micromorphic model has great potential to replicate certain characteristics of the myocardium in more detail. Specifically, the micromorphic micro-directors can represent the myocytes, thus allowing for non-affine relative deformations of the myocytes and the extracellular matrix (ECM) of tissue constraining the myocytes, which is not directly possible with classical models. A generalized micromorphic approach of Sansour [73, 74, 75] is explored in this study. Firstly, numerical examples are investigated and several novel proofs are devised to understand the behaviour of the micromorphic model with regards to numerical instabilities, micro-director displacements, and macro-traction vector contributions. An alternative micromorphic model is developed by the author for comparison against Sansour's model regarding the handling of micro-boundary conditions and other numerical artifacts. Secondly, Sansour's model is applied to cardiac modelling, whereby a macro-scale strain measure represents the deformation of the ECM of the tissue, a micro-scale strain measure represents the muscle fibres, and a third strain measure describes of the interaction of both constituents. Separate constitutive equations are developed to give unique stiffness responses to both the ECM and the myocytes. The micromorphic model is calibrated for cardiac tissue, first using triaxial shear experiments [80], and subsequently, to a pressure-volume relationship. The contribution of the micromorphic additional degrees of freedom to the various triaxial shear modes is quantified, and an analytical explanation is provided for differences in contributions. The passive filling phase of the heart cycle is investigated using a patient-specific left ventricle geometry supplied by the Cape Universities Body Imaging Centre (CUBIC) [38].
- ItemOpen AccessTime integration schemes for piecewise linear plasticity(1991) Rencontre, LJ; Martin, JBThe formulation of a generalized trapezoidal rule for the integration of the constitutive equations for a convex elastic-plastic solid is presented. This rule, which is based on an internal variable description, is consistent with a generalized trapezoidal rule for creep. It is shown that by suitable linear extrapolation, the standard backward difference algorithm can lead to this generalized trapezoidal rule or to a generalized midpoint rule. In either case, the generalized rules retain the symmetry of the consistent tangent modulus. It is also shown that the generalized trapezoidal and midpoint rules are fully equivalent in the sense that they lead to the establishment of the same minimum principle for the increment. The generalized trapezoidal rule thus inherits the notion of B-stability and both rules offer the opportunity to exploit the second order rate of convergence for a = ½. However, in the generalized trapezoidal rule, the equilibrium. and constitutive equations are fully satisfied at the end of the time increment. This may be more convenient than the generalized midpoint rule, in which equilibrium and plastic consistency are satisfied at the generalized midpoint. A backward difference return algorithm for piecewise linear yield surfaces is then formulated, with attention restricted to an associated flow rule and isotropic material behavior. Both the Tresca and Mohr-Coulomb yield surfaces with perfectly plastic and linear hardening rules are considered in detail. The algorithm has the advantage of being fully linked to the governing principles and avoids the inherent problems associated with corners on the yield surface. It is fully consistent in that no heuristic assumptions are made. The algorithm is extended to include the generalized trapezoidal rule in such a way that the general structure of the backward difference algorithm is maintained. This allows both for the computational advantages of the generalized trapezoidal rule to be utilized, and for a basis for comparison between this algorithm and existing backward difference algorithms to be established. Using this fully consistent algorithm, the return paths in stress space for the Tresca and Mohr-Coulomb yield surfaces with perfectly plastic and linear hardening rules are identified. These return paths thus provide a basis against which heuristically developed algorithms can be compared.