• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 2
  • Tagged with
  • 3
  • 3
  • 3
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Efficient "black-box" multigrid solvers for convection-dominated problems

Rees, Glyn Owen January 2011 (has links)
The main objective of this project is to develop a "black-box" multigrid preconditioner for the iterative solution of finite element discretisations of the convection-diffusion equation with dominant convection. This equation can be considered a stand alone scalar problem or as part of a more complex system of partial differential equations, such as the Navier-Stokes equations. The project will focus on the stand alone scalar problem. Multigrid is considered an optimal preconditioner for scalar elliptic problems. This strategy can also be used for convection-diffusion problems, however an appropriate robust smoother needs to be developed to achieve mesh-independent convergence. The focus of the thesis is on the development of such a smoother. In this context a novel smoother is developed referred to as truncated incomplete factorisation (tILU) smoother. In terms of computational complexity and memory requirements, the smoother is considerably less expensive than the standard ILU(0) smoother. At the same time, it exhibits the same robustness as ILU(0) with respect to the problem and discretisation parameters. The new smoother significantly outperforms the standard damped Jacobi smoother and is a competitor to the Gauss-Seidel smoother (and in a number of important cases tILU outperforms the Gauss-Seidel smoother). The new smoother depends on a single parameter (the truncation ratio). The project obtains a default value for this parameter and demonstrated the robust performance of the smoother on a broad range of problems. Therefore, the new smoothing method can be regarded as "black-box". Furthermore, the new smoother does not require any particular ordering of the nodes, which is a prerequisite for many robust smoothers developed for convection-dominated convection-diffusion problems. To test the effectiveness of the preconditioning methodology, we consider a number of model problems (in both 2D and 3D) including uniform and complex (recirculating) convection fields discretised by uniform, stretched and adaptively refined grids. The new multigrid preconditioner within block preconditioning of the Navier-Stokes equations was also tested. The numerical results gained during the investigation confirm that tILU is a scalable, robust smoother for both geometric and algebraic multigrid. Also, comprehensive tests show that the tILU smoother is a competitive method.
2

Modélisation mathématique et numérique de la migration cellulaire / Mathematical and numerical modelling of cell migration

Etchegaray, Christèle 29 November 2016 (has links)
Les déplacements cellulaires, collectifs ou individuels, sont essentiels pour assurer des fonctions fondamentales de l'organisme (réponse immunitaire, morphogenèse), mais jouent également un rôle crucial dans le développement de certaines pathologies (invasion métastatique).Les processus cellulaires à l'origine du déplacement forment une activité complexe, auto-organisée et fortement multi-échelle en temps mais aussi en espace. Mettre en évidence des principes généraux de la migration est donc un enjeu majeur. Dans cette thèse, nous nous intéressons à la construction de modèles de migration individuelle qui prennent en compte ce caractère multi-échelle de manière minimale.Dans une première partie, nous nous intéressons à des modèles particulaires. Nous décrivons des processus intracellulaires clés de la migration de manière discrète au moyen de processus de population. Puis, par une renormalisation en grand nombre d'individus, taille infinitésimale et dynamique accélérée, nous obtenons des équations de dynamique continue et stochastique, permettant de faire le lien entre la dynamique intracellulaire et le déplacement macroscopique.Nous nous confrontons d'abord à la situation d'un leucocyte se déplaçant dans une artère, et développant des liaisons de différentes natures avec les molécules de la paroi, jusqu'à éventuellement s'arrêter. La dynamique de formation de liaisons est décrite par un processus stochastique de type Naissance et Mort avec Immigration. Ces liaisons correspondent à des forces de résistance au mouvement. Nous obtenons explicitement le temps d'arrêt moyen de la cellule.Puis, nous nous intéressons à la reptation cellulaire, qui se produit grâce à la formation d'excroissances au bord de la cellule, appelées protrusions, qui avancent sur le substrat et exercent des forces de traction. Nous modélisons cette dynamique au moyen d'un processus de population structurée par l'orientation de la protrusion. Le modèle continu limite obtenu peut être étudié pour la migration 1D, et donne lieu à une équation de Fokker-Planck sur la distribution de probabilité de la population de protrusion. L'étude d'une configuration stationnaire permet de mettre en avant une dichotomie entre un état non motile et un état de déplacement directionnel.Dans une seconde partie, nous construisons un modèle déterministe minimal de migration dans un domaine discoïdal non déformable. Nous nous basons sur l'idée selon laquelle les structures responsables de la migration renforcent la polarisation de la cellule, ce qui favorise en retour un déplacement directionnel. Cette boucle positive passe par le transport d'un marqueur moléculaire dont la répartition inhomogène caractérise un état polarisé.Le modèle comporte un problème de convection-diffusion sur la concentration en marqueur, où le champs d'advection correspond à la vitesse d'un fluide de Darcy modélisant le cytosquelette. Son caractère actif est porté par des termes de bord, ce qui fait l'originalité du modèle.Du point de vue analytique, le modèle 1D présente une dichotomie face à une masse critique. Dans les cas sous-critique et critique, il est possible de montrer l'existence globale de solutions faibles, ainsi que la convergence à taux explicite vers l'unique état stationnaire correspondant à un état non polarisé. Au delà de la masse critique et pour des masses intermédiaires, nous mettons en évidence deux états stationnaires supplémentaires correspondant à des profils polarisés. De plus, pour des conditions initiales assez asymétrique, nous démontrons l'apparition d'un blow-up en temps fini.Du point de vue numérique, des tests numériques en 2D sont effectués en volumes finis (Matlab) et éléments finis (FreeFem++). Ils permettent de mettre en évidence à nouveau des états motiles et non motiles. L'effet de perturbations stochastiques est étudié, permettant d'aborder des cas de réponse à des signaux extérieurs chimique (chimiotactisme) ou mécanique (obstacle). / Collective or individual cell displacements are essential in fundamental physiological processes (immune response, embryogenesis) as well as in pathological developments (tumor metastasis). The intracellular processes responsible for cell motion have a complex self-organized activity spanning different time and space scales. Highlighting general principles of migration is therefore a challenging task.In a first part, we build stochastic particular models of migration. To do so, we describe key intracellular processes as discrete in space by using stochastic population models. Then, by a renormalization in large population, infinitesimal size and accelerated dynamics, we obtain continuous stochastic equations for the dynamics of interest, allowing a relation between the intracellular dynamics and the macroscopic displacement.First, we study the case of a leukocyte carried by the blood flow and developing adhesive bonds with the artery wall, until an eventual stop. The binding dynamics is described by a stochastic Birth and Death with Immigration process. These bonds correspond to resistive forces to the motion. We obtain explicitly the mean stopping time of the cell.Then, we study the case of cell crawling, that happens by the formation of protrusions on the cell edge, that grow on the substrate and exert traction forces. We describe this dynamics by a structured population process, where the structure comes from the protrusions' orientations. The limiting continuous model can be analytically studied in the 1D migration case, and gives rise to a Fokker-Planck equation on the probability distribution for the protrusion density. For a stationary profile, we can show the existence of a dichotomy between a non motile state and a directional displacement state.In a second part, we build a deterministic minimal migration model in a discoïdal cell domain. We base our work on the idea such that the structures responsible for migration also reinforce cell polarisation, which favors in return a directional displacement. This positive feedback loop involves the convection of a molecular marker, whose inhomogeneous spatial repartition is characteristic of a polarised state.The model writes as a convection-diffusion problem for the marker's concentration, where the advection field is the velocity field of the Darcy fluid that describes the cytoskeleton. Its active character is carried by boundary terms, which makes the originality of the model.From the analytical point of vue, the 1D model shows a dichotomy depending on a critical mass for the marker. In the subcritical and critical cases, it is possible to show global existence of weak solutions, as well as a rate-explicit convergence of the solution towards the unique stationary profile, corresponding to a non-motile state. Above the critical mass, for intermediate values, we show the existence of two additional stationary solutions corresponding to polarised motile profiles. Moreover, for asymmetric enough initial profiles, we show the finite time apparition of a blowup.Studying a more complex model involving activation of the marker at the cell membrane permits to get rid of this singularity.From the numerical point of vue, numerical experiments are led in 2D either in finite volumes (Matlab) or finite elements (FreeFem++) discretizations. They allow to show both motile and non motile profiles. The effect of stochastic fluctuations in time and space are studied, leading to numerical simulations of cases of responses to an external signal, either chemical (chemotaxis) or mechanical (obstacles).
3

Interakce stlačitelného proudění a struktur / Fluid-structure interaction of compressible flow

Hasnedlová, Jaroslava January 2012 (has links)
Title: Fluid-structure interaction of compressible flow Author: RNDr. Jaroslava Hasnedlová Department: Department of Numerical Mathematics, Institute of Applied Mathematics Supervisors: Prof. RNDr. Miloslav Feistauer, DrSc., Dr. h. c., Prof. Dr. Dr. h. c. Rolf Rannacher Supervisors' e-mail addresses: feist@karlin.mff.cuni.cz, rannacher@iwr.uni-heidelberg.de Abstract: The presented work is split into two parts. The first part is devoted to the theory of the discontinuous Galerkin finite element (DGFE) method for the space-time discretization of a nonstationary convection-diffusion initial-boundary value problem with nonlinear convection and linear diffusion. The DGFE method is applied sep- arately in space and time using, in general, different space grids on different time levels and different polynomial degrees p and q in space and time discretization. The main result is the proof of error estimates in L2 (L2 )-norm and in DG-norm formed by the L2 (H1 )-seminorm and penalty terms. The second part of the thesis deals with the realization of fluid-structure interaction problem of the compressible viscous flow with the elastic structure. The time-dependence of the domain occupied by the fluid is treated by the ALE (Arbitrary Lagrangian-Eulerian) method, when the compress- ible Navier-Stokes equations are formulated in...

Page generated in 0.1244 seconds