A uniformly well-conditioned, unfitted Nitsche method for interface problems (Q369416)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A uniformly well-conditioned, unfitted Nitsche method for interface problems |
scientific article |
Statements
A uniformly well-conditioned, unfitted Nitsche method for interface problems (English)
0 references
24 September 2013
0 references
The following boundary value problem is considered in the domain \(\Omega =\Omega_{1}\cup\Omega_{2}\) and \(\Omega_{1},\Omega_{2} \subset \mathbb{R}^{2}\): \[ -\nabla (\mu_{1}\nabla u_{1})=f_{1}\qquad \text{in}\quad \Omega_{1},\quad -\nabla (\mu_{2}\nabla u_{2})=f_{2}\qquad \text{in}\quad \Omega_{2}, \] \[ u_{1}- u_{2}=b \quad \text{on} \quad \Gamma=\partial\Omega_{1}, \quad \mu_{1}\partial_{n}u_{1}-\mu_{2}\partial_{n}u_{2}=g \quad \text{on} \quad \Gamma, \quad u_{2}=0 \quad \text{on} \quad \partial\Omega. \] The proposed method represents some development of the known approach for interface problems. The main distinctive feature of the method is the fact that the interface can be located arbitrary relative to the fixed background mesh. Usually, the discontinuities in the solutions of the partial differential equations are approximated by globally continuous finite element functions. This approximation is possible only in the case when the interface coincides with the mesh lines. This qualitative feature of the method seems to be of particular importance for the simulation of multi-phase-flow, where the interface changes significantly over the time. Another important feature of the proposed method is the fact that the dimension of the problem (the number of unknowns) is independent of the location of the interface. Reviewer's remark: Regarding the presented numerical results, I would like to say the following. We have here two examples in two space dimensions with different types of singularities and the corresponding comparisons with known results obtained by another approach. It is very a important comparison, of course. But both examples are considered in a very similar and almost regular domain (a circle in a square). In my opinion it would be useful to demonstrate the possibilities of the method in the case when the interface is a curve line, for example.
0 references
interface problem
0 references
elliptic equation
0 references
finite element method
0 references
stability
0 references
multi-phase flow
0 references
numerical results
0 references
0 references
0 references
0 references
0 references
0 references