Refine
Year of publication
- 2011 (3)
Language
- English (3)
Keywords
- Helmholtz equation (1)
- a priori error estimate (1)
- convergence (1)
- discrete network (1)
- finite element method (1)
- finite elements (1)
- heat equation (1)
- high contrast (1)
- micro-structure (1)
- multigrid (1)
Project
- C (3) (remove)
Application Area
- C (3)
This note constructs a local generalized finite element basis for elliptic problems with heterogeneous and highly varying diffusion tensor. The basis functions are solutions of local problems on vertex patches. The error of the corresponding generalized finite element method decays exponentially w.r.t. the number of element layers in the patches. Hence, on a uniform mesh of size H, patches of diameter log(1/H) are sufficient to preserve the convergence rates of the classical P1-FEM for the Poisson problem.
The analysis does not rely on regularity of the solution or scale separation in the coefficient.
The result justifies the use of the class of variational multiscale methods, introduced in [Comput. Methods Appl. Mech. Engrg., 196:2313--2324, 2007].
A new finite element method computes conductivity in some unstructured particle-reinforced composite material. The 2-phase material under consideration is composed of a poorly conducting matrix material filled by highly conducting circular inclusions which are randomly dispersed. The mathematical model is a Poisson-type problem with discontinuous coefficients. The discontinuities are huge in contrast and quantity.
The proposed method generalizes classical continuous piecewise affine finite elements to special computational meshes which encode the particles in a network structure. Important geometric parameters such as the volume fraction are preserved exactly.
The computational complexity of the resulting method is proportional to the number of inclusions and this is optimal in comparison to data complexity.
The discretization error is proportional to the particle distance and does not depend on the conductivity contrast in the medium.
Time-stepping procedures for the solution of evolution equations can be performed on parallel architecture by parallelizing the space computation at each time step. This, however, requires heavy communication between processors and becomes inefficient when many time-steps are to be computed and many processors are available. In such
cases parallelization in time is advantageous.
In this paper we present a method for parallelization in time of linear multistep discretizations of linear evolution problems; we consider a model parabolic and a model hyperbolic problem, and their, respectively, A(theta)-stable and A-stable linear multistep discretizations. The method consists of a discrete decoupling procedure, whereby N+1 decoupled Helmholtz problems with complex frequencies are obtained; N being the number of time steps computed in parallel. The usefulness of the method rests on our ability to solve these Helmholtz problems efficiently. We discuss the theory and give numerical examples for multigrid preconditioned iterative solvers of relevant
complex frequency Helmholtz problems. The parallel
approach can easily be combined with a time-stepping procedure, thereby obtaining a block time-stepping method where each block of steps is computed in parallel. In this way we are able to optimize the algorithm with respect to the number of processors available, the difficulty of solving the Helmholtz problems, and the possibility of both time and space adaptivity. Extensions to other linear evolution problems and to Runge-Kutta time discretization
are briefly mentioned.