Residual-Based a posteriori error estimation for hp-adaptive finite element methods for the Stokes equations

Arezou Ghesmati,Wolfgang Bangerth,Bruno Turcksin

DOI: https://doi.org/10.1515/jnma-2018-0047

2018-05-08

Abstract:We derive a residual-based a posteriori error estimator for the conforming hp-Adaptive Finite Element Method (hp-AFEM) for the steady state Stokes problem describing the slow motion of an incompressible fluid. This error estimator is obtained by extending the idea of a posteriori error estimation for the classical $h$-version of AFEM. We also establish the reliability and efficiency of the error estimator. The proofs are based on the well-known Clement-type interpolation operator introduced in 2005 in the context of the hp-AFEM. Numerical experiments show the performance of an adaptive hp-FEM algorithm using the proposed a posteriori error estimator.

Numerical Analysis

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to provide a residual - based posteriori error estimation method for Stokes equations within the framework of hp - adaptive finite element method (hp - AFEM). Specifically, the researchers aim to:

1. **Derive posteriori error estimators applicable to the hp - adaptive finite element method**: Different from the traditional h - adaptive method, the hp - adaptive method adjusts both the mesh size and the polynomial order simultaneously, thus requiring more complex error estimation techniques.

2. **Verify the reliability and efficiency of the error estimators**: Ensure that the proposed error estimators can accurately reflect the true error and are efficient in terms of computational cost. This involves proving the upper and lower bounds of the error estimators, namely reliability and efficiency.

3. **Propose an adaptive strategy based on the error estimators**: Based on the results of the error estimators, decide whether to reduce the error by refining the mesh (h - refinement) or increasing the polynomial order (p - refinement) to achieve optimal error reduction.

### Main contributions

- **New error estimators**: The authors introduce a family of parameterized error estimators \(\eta_\alpha\), where \(\alpha\in[0, 1]\), and verify their validity and reliability through numerical experiments.

- **Theoretical analysis**: Through detailed analysis of residuals and jump terms, establish the theoretical basis for the reliability and efficiency of the error estimators.

- **Adaptive algorithm**: Propose an adaptive algorithm based on the error estimators to guide mesh refinement or increase of the polynomial order to achieve an exponential convergence rate.

### Key formulas

- **Stokes equations**:
\[
\begin{cases}
-\nu\Delta\mathbf{u}+\nabla p = \mathbf{f}&\text{in }\Omega,\\
-\nabla\cdot\mathbf{u}=0&\text{in }\Omega,\\
\mathbf{u} = 0&\text{on }\Gamma.
\end{cases}
\]

- **Error estimators**:
\[
\eta^2_{\alpha;K}=\eta^2_{\alpha;K;R}+\eta^2_{\alpha;K;B},
\]
where,
\[
\eta^2_{\alpha;K;R}=\frac{h_K^2}{p_K^2}\left\|\left(I_K^{p_K}f+\nu\Delta\mathbf{u}_{FE}-\nabla p_{FE}\right)\Phi_K^{\alpha/2}\right\|^2_{L^2(K)}+\left\|(\nabla\cdot\mathbf{u}_{FE})\Phi_K^{\alpha/2}\right\|^2_{L^2(K)},
\]
\[
\eta^2_{\alpha;K;B}=\sum_{f\in F(K)}\frac{h_f}{p_f}\left\|\left[\nu\frac{\partial\mathbf{u}_{FE}}{\partial n}\right]\Phi_{\omega_f}^{\alpha/2}\right\|^2_{L^2(f)}.
\]

These formulas and theoretical analysis provide key support for understanding the core content of the paper.