Abstract:Optimally convergent (with respect to the regularity) quadratic finite element method for two dimensional obstacle problem on simplicial meshes is studied in (Brezzi, Hager, Raviart, Numer. Math, 28:431--443, 1977). There was no analogue of a quadratic finite element method on tetrahedron meshes for three dimensional obstacle problem. In this article, a quadratic finite element enriched with element-wise bubble functions is proposed for the three dimensional elliptic obstacle problem. A priori error estimates are derived to show the optimal convergence of the method with respect to the regularity. Further a posteriori error estimates are derived to design an adaptive mesh refinement algorithm. Numerical experiment illustrating the theoretical result on {\em a priori} error estimate is presented.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the lack of quadratic finite element methods applicable to tetrahedral meshes in the 3D - elliptic obstacle problem. Specifically, previous research has proposed using quadratic finite element methods to solve obstacle problems in two - dimensional cases, and these methods exhibit optimal convergence on triangular meshes. However, in three - dimensional space, there are no similar quadratic finite element methods for obstacle problems on tetrahedral meshes.
To solve this problem, the author proposes a new method: by adding cell - level bubble functions in the standard P2 Lagrange finite element space, a new quadratic finite element space is constructed. This method can not only maintain the original quadratic polynomial approximation ability, but also enhance the local approximation performance by introducing bubble functions, so as to better handle the free - boundary phenomenon in the 3D obstacle problem.
In addition, the paper also derives a priori error estimates and a posteriori error estimates to prove that the proposed numerical method has optimal convergence, and designs an adaptive mesh refinement algorithm. Finally, the effectiveness of the theoretical results is verified through numerical experiments.
### Key Formulas
1. **Definition of Bubble Function**:
\[
b_T = 256 \lambda_{T,1} \lambda_{T,2} \lambda_{T,3} \lambda_{T,4}
\]
where \(\lambda_{T,i}\) is the barycentric coordinate of tetrahedron \(T\).
2. **Discrete Problem**:
Seek \(u_h\in K_h\) such that:
\[
a(u_h, v_h - u_h)\geq (f, v_h - u_h)\quad \forall v_h\in K_h
\]
where \(K_h = \{v_h\in V_h : A_h(v_h)\geq A_h(\chi)\}\), and \(A_h(v_h)\) represents the average value of \(v_h\) on each cell.
3. **A Priori Error Estimate**:
\[
\|\nabla (u - u_h)\|\leq Ch^{3/2 - \epsilon}\left( \|u\|_{H^{5/2 - \epsilon}(\Omega)}+ \|f\|_{H^1(\Omega)}+ \|\chi\|_{H^3(\Omega)}+ \|u\|_{H^3(\Omega_N)} \right)
\]
4. **A Posteriori Error Estimate**:
\[
\|\nabla (u - u_h)\|^2+\|\sigma - \sigma_h\|^2_{H^{-1}(\Omega)}\leq C\left( \eta_1^2+\eta_2^2+\|\nabla (\chi - u_h)^+\|^2 - \sum_{T\in C_h}\int_T\sigma_h (\chi - u_h)^- dx \right)
\]
Through these formulas and methods, the paper successfully solves the numerical solution problem of the 3D elliptic obstacle problem on tetrahedral meshes.