Spelling suggestions: "subject:"asynchronous c.method"" "subject:"asynchronous 20method""
1 |
Méthode numérique asynchrone pour la modélisation de phénomènes multi-échelles / Asynchrononous numerical scheme for modeling multi-scale phenomenaToumi, Asma 26 September 2016 (has links)
La simulation numérique est devenue un outil central dans la modélisation de nombreux systèmes physiques tels que la dynamique des fluides, les plasmas, l'électromagnétisme, etc. L'existence de phénomènes multi-échelles rend l'intégration numérique de ces modèles très difficile du point de vue de la précision et du temps de calcul. En effet, dans les méthodes classiques d'intégration temporelle, le pas de temps est limité par la taille des plus petites mailles au travers d'une relation de type CFL. De plus, la forte disparité entre le pas de temps effectif et la condition CFL favorise les phénomènes de diffusion numérique. Dans la littérature, des nombreux algorithmes à pas de temps locaux (LTS) ont été développés. Pour la plupart des algorithmes LTS, les pas de temps locaux doivent être choisis parmi les fractions du pas de temps global. Nous présentons dans cette thèse une méthode asynchrone pour l'intégration explicite des équations différentielles multi-échelles. Cette méthode repose sur l'utilisation de critères de stabilité locaux, critères déterminés non pas globalement mais à partir de conditions CFL locales. De plus, contrairement aux schémas LTS, l'algorithme asynchrone permet la sélection de pas de temps indépendants pour chaque cellule de maillage. Cette thèse comporte plusieurs volets. Le premier concerne l'étude mathématique des propriétés du schéma asynchrone. Le deuxième a pour objectif d'étudier la montée en ordre, à la fois temporelle et spatiale, des méthodes asynchrones. De nombreux développements dans le cadre des méthodes de haute précision en temps ou en espace, telles que les méthodes de type Galerkin Discontinu, peuvent offrir un cadre naturel pour l'amélioration de la précision des méthodes asynchrones. Toutefois, les estimations garantissant l'ordre de précision de ces méthodes peuvent ne pas être directement compatibles avec l'aspect asynchrone. L'objectif de cette thèse est donc de développer un schéma numérique asynchrone d'ordre élevé mais qui permet également de limiter la quantité de calculs à effectuer. Le troisième volet de cette thèse se focalise sur l'application numérique puisqu'il concerne la mise en oeuvre de la méthode asynchrone dans la simulation des cas-tests représentatifs de problèmes multi-échelles. / Numerical simulation has become a central tool for the modeling of many physical systems (Fluid dynamics, plasmas, electromagnetism, etc). Multi-scale phenomena make the integration of these physical systems difficult in terms of accuracy and computational time. Numerical time-stepping integration techniques used for modeling such problems generally fall into two categories : explicit and implicit schemes. In the explicit schemes, all unknown variables are computed at the current time level from quantities already available. The time step used is then limited by the most restrictive CFL condition over the whole computation domain. In the implicit method the time step is no longer limited by the CFL conditions. However the scheme is generally not suitable for strongly coupled problems. To solve such problems, a number of local time-stepping (LTS) approaches have been developed. These methods are restricted by a local CFL condition rather than the traditional global CFL condition. For most of these LTS algorithms, local time steps are usually selected to be fractions of the global time step so that regular meeting points in time exist, and only little work is available on LTS methods with independent time steps. We present in this thesis an asynchronous method for the explicit integration of multi-scale partial differential equations. This method is restricted by a local CFL condition rather than the traditional global CFL condition. Moreover, contrary to other LTS methods, the asynchronous algorithm permits the selection of independent time steps in each mesh element. Our work consists of several components. The first one concerns the mathematical study of the properties of the asynchronous method. the objective of the second part is to study the improvement of the convergence rate for asynchronous methods. Many approaches in the context of high precision methods in time or in space, such as the Discontinuous Galerkin methods, may offer a natural setting to improve the precision of the asynchronous methods. However, the estimates ensuring the order of the accuracy of the method may not be directly compatible with the asynchronous aspect. Then, the objective is to develop a high order asynchronous numerical scheme which also preserves the computational time reduction. Finally, the third part is focused on the implementation of the asynchronous method and illustrate the advantages of the method on test-cases representative of multiscale problems.
|
2 |
Multiple interval methods for ODEs with an optimization constraintYu, Xinli January 2020 (has links)
We are interested in numerical methods for the optimization constrained second order ordinary differential equations arising in biofilm modelling. This class of problems is challenging for several reasons. One of the reasons is that the underlying solution has a steep slope, making it difficult to resolve. We propose a new numerical method with techniques such as domain decomposition and asynchronous iterations for solving certain types of ordinary differential equations more efficiently. In fact, for our class of problems after applying the techniques of domain decomposition with overlap we are able to solve the ordinary differential equations with a steep slope on a larger domain than previously possible. After applying asynchronous iteration techniques, we are able to solve the problem with less time.~We provide theoretical conditions for the convergence of each of the techniques. The other reason is that the second order ordinary differential equations are coupled with an optimization problem, which can be viewed as the constraints. We propose a numerical method for solving the coupled problem and show that it converges under certain conditions. An application of the proposed methods on biofilm modeling is discussed. The numerical method proposed is adopted to solve the biofilm problem, and we are able to solve the problem with larger thickness of the biofilm than possible before as is shown in the numerical experiments. / Mathematics
|
Page generated in 0.032 seconds