Parallel FETI algorithms for mortars

Authors:
Dan Stefanica
Affiliations:
Baruch College, City University of New York, New York, NY and Massachusetts Institute of Technology, Cambridge, MA
Venue:
Applied Numerical Mathematics - 6th IMACS International symposium on iterative methods in scientific computing
Year:
2005

Citing 9
Cited 0

A Scalable Substructuring Method by Lagrange Multipliers for Plate Bending Problems

SIAM Journal on Numerical Analysis
A Numerical Study of FETI Algorithms for Mortar Finite Element Methods

SIAM Journal on Scientific Computing
A FETI Preconditioner for Two Dimensional Edge Element Approximations of Maxwell's Equations on Nonmatching Grids

SIAM Journal on Scientific Computing
A Domain Decomposition Method with Lagrange Multipliers and Inexact Solvers for Linear Elasticity

SIAM Journal on Scientific Computing
A Mortar Finite Element Method Using Dual Spaces for the Lagrange Multiplier

SIAM Journal on Numerical Analysis
A FETI Domain Decomposition Method for Edge Element Approximations in Two Dimensions with Discontinuous Coefficients

SIAM Journal on Numerical Analysis
Dual-Primal FETI Methods for Three-Dimensional Elliptic Problems with Heterogeneous Coefficients

SIAM Journal on Numerical Analysis
FETI and FETI-DP Methods for Spectral and Mortar Spectral Elements: A Performance Comparison

Journal of Scientific Computing
Domain Decomposition Methods for Mortar Finite Elements

Domain Decomposition Methods for Mortar Finite Elements

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose a new version of the FETI algorithm for mortar discretizations based on generalized coupling conditions across the interface. The new algorithm preserves the parallelization properties of, and has the same scalability properties as, the FETI method for continuous finite elements. We compare the numerical performance of the new algorithm with that of other FETI methods for mortars. When mortar conditions are required only on a small part of the interface, while continuity is required elsewhere, we show that the extra storage and extra computational effort required by the parallel algorithm is minimal.