Adaptive 2D IGA boundary element methods

Michael Feischl,Gregor Gantner,Alexander Haberl,Dirk Praetorius

DOI: https://doi.org/10.1016/j.enganabound.2015.10.003

2015-04-23

Abstract:We derive and discuss a posteriori error estimators for Galerkin and collocation IGA boundary element methods for weakly-singular integral equations of the first-kind in 2D. While recent own work considered the Faermann residual error estimator for Galerkin IGA boundary element methods, the present work focuses more on collocation and weighted- residual error estimators, which provide reliable upper bounds for the energy error. Our analysis allows piecewise smooth parametrizations of the boundary, local mesh-refinement, and related standard piecewise polynomials as well as NURBS. We formulate an adaptive algorithm which steers the local mesh-refinement and the multiplicity of the knots. Numerical experiments show that the proposed adaptive strategy leads to optimal convergence, and related IGA boundary element methods are superior to standard boundary element methods with piecewise polynomials.

Numerical Analysis

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper is mainly devoted to solving the problem of how to provide reliable posteriori error estimators for weakly - singular integral equations in two - dimensional Isogeometric Analysis (IGA) Boundary Element Method (BEM). Specifically:

1. **Posteriori error estimation**:
- The paper derives and discusses posteriori error estimators for Galerkin and collocation IGA boundary element methods. These estimators can provide reliable upper bounds for energy errors.
- In particular, the paper focuses on the collocation method and weighted residual error estimators and proves their effectiveness in controlling energy errors.

2. **Adaptive algorithm**:
- An adaptive algorithm is proposed, which can guide local mesh refinement and the increase of node multiplicity according to the results of error estimators.
- The goal of the adaptive strategy is to achieve the optimal convergence rate by optimizing the mesh and node distribution.

3. **Numerical experiments**:
- The superiority of the proposed adaptive IGA boundary element method is verified through numerical experiments. The experiments show that the adaptive strategy not only improves the convergence speed but also performs better when dealing with problems with singularities or jump solutions.

4. **Comparative analysis**:
- The adaptive IGA boundary element method is compared with the standard boundary element method, and the results show that the former has obvious advantages in dealing with complex geometries and irregular solutions.

### Formula summary

- **Weakly - singular integral equation**:
\[
V\varphi(x) := -\frac{1}{2\pi} \int_{\Gamma} \log|x - y| \varphi(y) dy = f(x) \quad \text{on } \Gamma
\]

- **Galerkin BEM discretization**:
\[
\int_{\Gamma} V\varphi_h \psi_h dx = \int_{\Gamma} f \psi_h dx \quad \forall \psi_h \in X_h
\]

- **Collocation BEM discretization**:
\[
V\varphi_h(x_j) = f(x_j) \quad \forall x_j \in \{x_1, \ldots, x_{N_{\text{col}}}\}
\]

- **Posteriori error estimation**:
- Galerkin IGA - BEM:
\[
C_{\text{rel}}^{-1} \| \varphi - \varphi_h \| \leq \eta_h := \left( \sum_{z \in N_h} \eta_h(z)^2 \right)^{1/2} \leq C_{\text{eff}} \| \varphi - \varphi_h \|
\]
where,
\[
\eta_h(z)^2 := \iint_{\omega_h(z) \times \omega_h(z)} \frac{|r_h(x) - r_h(y)|^2}{|x - y|^2} dydx
\]
- Collocation IGA - BEM:
\[
C_{\text{rel}}^{-1} \| \varphi - \varphi_h \| \leq \mu_h := \left( \sum_{z \in N_h} \mu_h(z)^2 \right)^{1/2}
\]
where,
\[
\mu_h(z)^2 := | \omega_h(z) | \int_{\omega_h(z)} |r'_h(x)|^2 dx
\]

Through these formulas, the paper shows how in the IGA boundary element method.