Abstract:In this paper, we propose and analyze the least squares finite element methods for the linear elasticity interface problem in the stress-displacement system on unfitted meshes. We consider the cases that the interface is $C^2$ or polygonal, and the exact solution $(\sigma,u)$ belongs to $H^s(div; \Omega_0 \cup \Omega_1) \times $H^{1+s}(\Omega_0 \cup \Omega_1)$ with $s > 1/2$. Two types of least squares functionals are defined to seek the numerical solution. The first is defined by simply applying the $L^2$ norm least squares principle, and requires the condition $s \geq 1$. The second is defined with a discrete minus norm, which is related to the inner product in $H^{-1/2}(\Gamma)$. The use of this discrete minus norm results in a method of optimal convergence rates and allows the exact solution has the regularity of any $s > 1/2$. The stability near the interface for both methods is guaranteed by the ghost penalty bilinear forms and we can derive the robust condition number estimates. The convergence rates under $L^2$ norm and the energy norm are derived for both methods. We illustrate the accuracy and the robustness of the proposed methods by a series of numerical experiments for test problems in two and three dimensions.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper aims to solve the numerical solution problem of the linear elasticity interface problem on unfitted meshes. Specifically, the paper proposes the Least Squares Finite Element Methods (LSFEMs) to solve the linear elasticity interface problem in the stress - displacement system.
#### Main problems and challenges
1. **Discontinuous material properties**: The linear elasticity interface problem usually involves structures with different or even singular material properties, which makes the coefficients of the governing equations discontinuous on the interface and includes non - homogeneous jump conditions.
2. **Handling of complex geometries**: Due to the discontinuities near the interface and the irregularity of the interface geometry, it is still challenging to design efficient numerical methods to solve such equations.
3. **Handling of solutions with low regularity**: Many existing methods require that the exact solution has piecewise \(H^2\) regularity, but for solutions with only piecewise \(H^{1 + s}\) (\(s>\frac{1}{2}\)) regularity, these techniques may not be applicable or may lead to sub - optimal convergence rates.
#### Main contributions of the paper
1. **Two least - squares functionals**:
- The first functional is defined by simply applying the \(L^2\) - norm least - squares principle and is suitable for the case of \(s\geq1\).
- The second functional uses the discrete minus inner product, which is related to the inner product in the \(H^{-\frac{1}{2}}(\Gamma)\) space, allows the exact solution to have any regularity of \(s >\frac{1}{2}\), and gives the optimal convergence rate.
2. **Stability and convergence analysis**: By introducing ghost penalty bilinear forms, the stability near the interface is ensured, and the convergence rates in the \(L^2\) - norm and energy norm are derived.
3. **Numerical experiment verification**: Through a series of two - dimensional and three - dimensional numerical experiments, the accuracy and robustness of the proposed method are demonstrated.
#### Summary of mathematical formulas
- **Least - squares functional**:
\[
J(\tau, v; f, a, b)=J(\tau, v; f)+B_\sigma(\tau; a)+B_u(v; b)
\]
where
\[
J(\tau, v; f)=\|A\tau-\varepsilon(v)\|^2_{L^2(\Omega_0\cup\Omega_1)}+\|\nabla\cdot\tau + f\|^2_{L^2(\Omega_0\cup\Omega_1)}
\]
\[
B_\sigma(\tau; a)=\|[ [\tau] ]_N - a\|^2_{H^{-\frac{1}{2}}(\Gamma)}, \quad B_u(v; b)=\|[ [v] ] - b\|^2_{H^{\frac{1}{2}}(\Gamma)}
\]
- **Regularity assumption**:
\[
(\sigma, u)\in H^s(\text{div}; \Omega_0\cup\Omega_1)\times H^{1 + s}(\Omega_0\cup\Omega_1), \quad s >\frac{1}{2}
\]
- **Ghost penalty bilinear form**:
\[
S^m_h(\tau_h, \rho_h)=s^m_{h,0}(\pi_0\tau_h, \pi_0\rho_h)+s^m_{h,1}(\pi_1\tau_h,