An unfitted interface penalty method for the numerical approximation of contrast problems (Q719434)

From MaRDI portal
Revision as of 21:37, 4 July 2023 by Importer (talk | contribs) (‎Created a new Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
An unfitted interface penalty method for the numerical approximation of contrast problems
scientific article

    Statements

    An unfitted interface penalty method for the numerical approximation of contrast problems (English)
    0 references
    0 references
    0 references
    0 references
    10 October 2011
    0 references
    The authors focus on the developement of a fictitious domain method for interface problems governed by elliptic equations with large contrast between the diffusion coefficients \[ -\nabla\cdot(\varepsilon_{i}\nabla u_{i}) +\mu_{i}u_{i} =f_{i}\text{ in }\Omega_{i},\;i=1,2, \] \[ u_{i}=0\text{ on }\partial\Omega\cap\partial\Omega_{i},\;u_{1}=u_{2}\text{ on }\Gamma ,\;\epsilon_{1}\partial_{n}u_{1}=\epsilon_{2}\partial_{n}u_{2}\text{ on }\Gamma. \] They develop a robust discretization scheme based on extended finite elements on meshes that do not fit the interface. In order to improve the unsatisfactory behaviour of Lagrangian elements, they utilize an enriched approximation space, which involves elements cut by the interface. They analyze the condition number of mass and stiffness matrices in presence of small sub-elements and conclude that their optimal condition number is uniformly independent on how the interface cuts the mesh. The numerical scheme may be ill posed for some configurations of the interface. The authors propose a stabilization stategy, based on a scaling technique, which restores the standard properties of a Lagrangian finite element space. The robustness of the discretization scheme is gained thanks to a suitable choice of the parameters that define the penalty technique combined with the application of a diagonal preconditioner to the solution of the discrete system of equations. Some concluding numerical experiments confirm the efficiency of the scheme.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    large contrast problems
    0 references
    fictitious domain methods
    0 references
    extended finite elements
    0 references
    Nitsche's method
    0 references
    interface problems
    0 references
    preconditioning
    0 references
    condition number
    0 references
    stabilization
    0 references
    scaling
    0 references
    numerical experiments
    0 references