Abstract:We consider bivariate piecewise polynomial finite element spaces for curved domains bounded by piecewise conics satisfying homogeneous boundary conditions, construct stable local bases for them using Bernstein-Bézier techniques, prove error bounds and develop optimal assembly algorithms for the finite element system matrices. Numerical experiments confirm the effectiveness of the method.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to construct a bivariate piecewise polynomial finite - element space for solving second - order elliptic problems with homogeneous boundary conditions on curved - edged regions bounded by piecewise quadratic algebraic curves (i.e., conic curves). Specifically, the authors focus on the following points:
1. **Constructing a stable local basis**: Use the Bernstein - Bézier technique to construct a stable local basis for these piecewise polynomial spaces.
2. **Error estimation**: Prove the error bounds of these finite - element methods to ensure their effectiveness in numerical implementation.
3. **Optimal assembly algorithm**: Develop an optimal assembly algorithm for the finite - element system matrix to improve computational efficiency.
### Detailed Explanation
#### 1. Constructing a stable local basis
To deal with elliptic problems with homogeneous Dirichlet boundary conditions, it is necessary to describe a suitable local basis. To this end, the authors use the Bernstein - Bézier technique and construct a stable minimal determining set (MDS) for the piecewise polynomial space \( S_d^0(\Delta) \). The main idea is to factorize the polynomial on the wedge - shaped triangle, and in this way, the polynomial on the curved - edged region can be processed more effectively.
#### 2. Error estimation
The authors analyze the approximation ability of these piecewise polynomial spaces and provide error estimates. Specifically, for \( u\in H^m(\Omega)\cap H_0^1(\Omega) \), the error bound is:
\[
\inf_{s\in S_d^0(\Delta)}\|u - s\|_{H^k(\Omega)}\leq C_4h^{m - k}\|u\|_{H^m(\Omega)}, \quad k = 0,\ldots,m - 1
\]
where \( h \) is the maximum diameter of the triangles in the triangulation, and \( C_4 \) is a constant depending only on \( d \) and \( R \).
#### 3. Optimal assembly algorithm
To improve computational efficiency, the authors develop an optimal assembly algorithm for the finite - element system matrix. These algorithms can significantly reduce the computational complexity, making numerical experiments more efficient and feasible.
### Numerical experiment verification
Finally, the authors verify the effectiveness of this method through several numerical experiments, including Poisson problems on different curved - edged regions and circular membrane eigenvalue problems. The results show that this method exhibits good performance in both \( h \)-refinement and \( p \)-refinement settings.
### Summary
This paper proposes a new finite - element method that can effectively solve second - order elliptic problems with homogeneous boundary conditions on curved - edged regions bounded by piecewise quadratic algebraic curves. By constructing a stable local basis, providing strict error estimates, and developing an efficient assembly algorithm, this method is not only theoretically guaranteed but also performs well in practical applications.