Robust Algebraic Multigrid Methods in Magnetic Shielding Problems
نویسنده
چکیده
The aim of this diplom thesis is to provide a robust and e cient solver for large sparse and poor conditioned linear systems arising from the FEmethod for elliptic scalar PDEs of second order. For a counter example the problem of magnetic shielding is used. Therefore the Maxwell's equations for stationary objects are reduced to a scalar PDE of second order with appropriate boundary conditions. In order to solve the equation by means of FEM, a discretization for micro scales is introduced. Especially long thin elements are suggested to keep the number of unknowns small in areas of micro structures. Constructively a nite element analysis is carried out where also a convergence result of the FE-solution in H1 is presented. To achieve an e cient and robust solution strategy the algebraic multigrid method of Ruge and St uben is introduced. Additionally three di erent areas of application are presented for this AMG method, i.e. preconditioner, coarse grid solver for a full multigrid method, and black box solver. Because this AMG method normally works well for M-matrices, a technique is presented to attain M-matrices, if the underlying linear system arises from an FE-discretization. The method to achieve the M-matrix property is based on the element matrices. The algorithm was implemented as black box solver in the nite element package FEPP. Therein AMG was applied as preconditioner for the conjugate gradient method. Some numerical experiments are presented, where long thin quadrilaterals are used with ratio of the longest and shortest side of 1 to 10 3. Additionally parameter jumps of order 10 6 to 10+6 are considered. Concluding AMG has been proven, at least in a numerical way, to be an e cient and robust solver for magnetic shielding problems, if it is used as a preconditioner for the CG-method. If long thin quadrilaterals are used for discretization the modi ed preconditioner also behaves very robust. 3 Zusammenfassung Das Ziel dieser Diplomarbeit ist die Entwicklung einer robusten und e zienten Losungsstrategie f ur gro e, d unngesetzte und schlecht konditionierte lineare Gleichungssysteme, welche bei der Diskretisierung von elliptischen Di erentialgleichungen 2.Ordnung entstehen. Als Beispiele dienen Abschirmprobleme in 2D, wof ur die Maxwell'schen Gleichungen als Grundlage f ur die Modellierung ben utzt werden. Diese Gleichungen reduzieren sich zu einer skalaren partiellen Di erentialgleichung 2.Ordnung mit geeigneten Randbedingungen. Um diese Problemklasse numerisch mit Hilfe der Methode der Finiten Elemente zu losen, wird eine geeignete Diskretisierung f ur Mikroskalen eingef uhrt. Zu diesem Zweck werden lange d unne Elemente vorgeschlagen, um die Anzahl der Unbekannten gering zu halten. Weiters wird eine Konvergenzanalysis f ur die FEM-Losung in H1 bereitgestellt. Als Grundlage f ur einen e zienten und robusten Loser dient die algebraische Multigrid Methode von Ruge und St uben. Aufbauend auf dieser Methode werden drei Anwendungsgebiete beschrieben: Vorkonditionierer, Grobgitterloser f ur die Full Multigrid Methode und Black Box Loser. Da die vorgeschlagene Technik im allgemeinen nur f ur M-Matrizen zufriedenstellend funktioniert wird eine Methode bereitgestellt, welche spektralaquivalente M-Matrizen generiert. Diese Technik basiert auf den Elementmatrizen. Der Algorithmus wurde im Finite Elemente Programm FEPP implementiert, in dem AMG als Vorkonditionierer f ur das konjugierte Gradientenverfahren verwendet wird. Abschlie end werden numerische Experimente durchgef uhrt, in denen lange, d unne Vierecke mit einem Seitenverhaltnis von ca. 1 : 10 3 verwendet werden. Zusatzlich werden Parameterspr unge in der Gro enordnung von 10 6 bis 10+6 angenommen. AMG ist ein, im zumindest numerischen Sinn, e zienter und robuster Loser f ur Abschirmprobleme, falls es als Vorkonditionierer f ur das CG Verfahren verwendet wird. Werden lange d unne Vierecke zur Diskretisierung verwendet, so ist der modi zierte Vorkonditionierer ebenfalls sehr robust. 4
منابع مشابه
Spectral element agglomerate algebraic multigrid methods for elliptic problems with high-contrast coefficients
We apply a recently proposed ([3]) robust overlapping Schwarz method with a certain spectral construction of the coarse space in the setting of element agglomeration algebraic multigrid methods (or agglomeration AMGe) for elliptic problems with high-contrast coefficients. Our goal is to design multilevel iterative methods that converge independent of the contrast in the coefficients. We present...
متن کاملNumerical study of multigrid methods with various smoothersfor the elliptic grid generation
A robust solver for the elliptic grid generation equations is sought via a numerical study. The system of PDEs is discretized with nite diierences and multigrid methods are applied to the resulting nonlinear algebraic equations. Multigrid iterations are compared with respect to the robustness and eeciency. Diierent smoothers are tried to improve the convergence of iterations. The methods are ap...
متن کاملScaling Structured Multigrid to 500K+ Cores through Coarse-Grid Redistribution
The efficient solution of sparse, linear systems resulting from the discretization of partial differential equations is crucial to the performance of many physics-based simulations. The algorithmic optimality of multilevel approaches for common discretizations makes them a good candidate for an efficient parallel solver. Yet, modern architectures for high-performance computing systems continue ...
متن کاملAlgebraic multigrid methods for constrained linear systems with applications to contact problems in solid mechanics
This article develops a general framework for applying algebraic multigrid techniques to constrained systems of linear algebraic equations that arise in applications with discretized PDEs. We discuss constraint coarsening strategies for constructing multigrid coarse grid spaces and several classes of multigrid smoothers for these systems. The potential of these techniques is investigated with t...
متن کاملA Boundary Condition-Capturing Multigrid Approach to Irregular Boundary Problems
We propose a geometric multigrid method for solving linear systems arising from irregular boundary problems involving multiple interfaces in two and three dimensions. In this method, we adopt a matrix-free approach; i.e., we do not form the fine grid matrix explicitly and we never form nor store the coarse grid matrices, as many other robust multigrid methods do. The main idea is to construct a...
متن کامل