Optimal preconditioners for a Nitsche stabilized fictitious domain finite element method

TL;DR

This paper develops and analyzes a new class of preconditioners for Nitsche stabilized fictitious domain finite element methods applied to the Poisson equation, demonstrating their stability and efficiency through theoretical proofs and numerical experiments.

Contribution

A novel subspace decomposition-based preconditioner for fictitious domain FEM with Nitsche stabilization, proven to be stable and uniformly effective.

Findings

Preconditioner is stable and uniform in discretization and boundary location.

Discretization in one subspace yields a well-conditioned matrix.

Discretization in the other subspace is equivalent to a standard FEM for Poisson.

Abstract

In this paper we consider a class of fictitious domain finite element methods known from the literature. These methods use standard finite element spaces on a fixed unfitted triangulation combined with the Nitsche technique and a ghost penalty stabilization. As a model problem we consider the application of such a method to the Poisson equation. We introduce and analyze a new class of preconditioners that is based on a subspace decomposition approach. The finite element space is split into two subspaces, where one subspace is spanned by all nodal basis functions corresponding to nodes on the boundary of the fictitious domain and the other space is spanned by all remaining nodal basis functions. We will show that this splitting is stable, uniformly in the discretization parameter and in the location of the problem boundary in the triangulation. We also prove that the Galerkin discretization in the first subspace leads to a uniformly well-conditioned matrix and that the Galerkin discretization in the second subspace is uniformly equivalent to a standard finite element discretization of a Poisson equation on the fictitious domain with homogeneous Dirichlet boundary conditions. Results of numerical experiments that illustrate optimality of such a preconditioner are included.