Abstract:We present a local construction of H(curl)-conforming piecewise polynomials satisfying a prescribed curl constraint. We start from a piecewise polynomial not contained in the H(curl) space but satisfying a suitable orthogonality property. The procedure employs minimizations in vertex patches and the outcome is, up to a generic constant independent of the underlying polynomial degree, as accurate as the best-approximations over the entire local versions of H(curl). This allows to design guaranteed, fully computable, constant-free, and polynomial-degree-robust a posteriori error estimates of Prager-Synge type for Nédélec finite element approximations of the curl-curl problem. A divergence-free decomposition of a divergence-free H(div)-conforming piecewise polynomial, relying on over-constrained minimizations in Raviart-Thomas spaces, is the key ingredient. Numerical results illustrate the theoretical developments.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to design a balanced flux reconstruction method for the curl - curl problem. This method is implemented based on local minimization in the H(curl) space and is robust with respect to the polynomial degree p. Specifically, the research aims to:
1. **Construct piecewise polynomials conforming to H(curl)**: These polynomials satisfy the given curl constraints. The author starts from a piecewise polynomial not in the H(curl) space, but this polynomial satisfies appropriate orthogonality properties.
2. **Ensure the accuracy of the reconstructed flux**: Through minimization operations on vertex patches, the resulting accuracy, up to a universal constant (independent of the underlying polynomial degree), is comparable to the best approximation in the entire local version of H(curl).
3. **Design a posteriori error estimates**: This method allows the design of guaranteed, fully computable, constant - free, and polynomial - degree - p - robust Prager - Synge - type a posteriori error estimates. In particular, when Nédélec finite elements are used to approximate the curl - curl problem, it can provide reliable error estimates.
4. **Handle divergence - free decomposition**: By performing over - constrained minimization in the Raviart - Thomas space, the decomposition of divergence - free piecewise polynomials is achieved, which is one of the key steps.
### Summary of Mathematical Formulas
- **Definitions of H(curl) and H(div) spaces**:
\[
H(\text{curl}, \omega):=\left\{v \in L^{2}(\omega)^{3} ; \nabla \times v \in L^{2}(\omega)^{3}\right\}
\]
\[
H(\text{div}, \omega):=\left\{v \in L^{2}(\omega)^{3} ; \nabla \cdot v \in L^{2}(\omega)\right\}
\]
- **Discretization assumptions**:
\[
j \in H_{0}^{N}(\text{div}, \Omega), \quad \nabla \cdot j = 0, \quad(j, \phi)=0 \quad \forall \phi \in H(\Omega, \Gamma_{D})
\]
- **curl - curl problem**:
\[
\nabla \times(\nabla \times A)=j, \quad \nabla \cdot A = 0 \quad \text{in } \Omega
\]
\[
A \times n_{\Omega}=0 \quad \text{on } \Gamma_{D}, \quad(\nabla \times A) \times n_{\Omega}=0, \quad A \cdot n_{\Omega}=0 \quad \text{on } \Gamma_{N}
\]
- **Assumptions on the discrete magnetic vector potential \(A_{h}\)**:
\[
A_{h} \in N_{p}(T_{h}) \cap H_{0}^{D}(\text{curl}, \Omega), \quad(\nabla \times A_{h}, \nabla \times v_{h})=(j, v_{h}) \quad \forall v_{h} \in N_{0}(T_{h}) \cap H_{0}^{D}(\text{curl}, \Omega)
\]
- **Data oscillation term**:
\[
\tilde{\eta}_{a, \text{osc}, j}:=\left(\sum_{K \in \tilde{T}_{a}}\left(\frac{h_{K}}{\pi}\|j - \Pi_{p'}(j)\|_{K}\right)^{2}\right)^{1 / 2}
\]
### Conclusion