Bishnu P. Lamichhane
Abstract:We present two simple finite element methods for the discretization of Reissner--Mindlin plate equations with {\em clamped} boundary condition. These finite element methods are based on discrete Lagrange multiplier spaces from mortar finite element techniques. We prove optimal a priori error estimates for both methods.
What problem does this paper attempt to address?
This paper attempts to solve the locking problem that occurs when the Reissner - Mindlin plate equations are discretized by the finite element method under clamped boundary conditions. Specifically, when the plate thickness becomes very small, the traditional finite - element method will have problems of numerical instability or inaccuracy. To address this challenge, the authors propose two simple and efficient finite - element methods, which are based on the discrete Lagrange multiplier space and combined with the mortar finite - element technique.
### Main problems and solutions
1. **Problem description**:
- When the plate thickness \( t \) becomes very small, the standard finite - element discretization method will have a locking phenomenon, resulting in inaccurate or unstable numerical solutions.
- The clamped boundary condition is one of the most difficult boundary conditions in plate theory and thus requires special treatment.
2. **Solutions**:
- The authors propose two finite - element methods, which are based on the discrete Lagrange multiplier space and combined with the mortar finite - element technique.
- The first method uses the standard Lagrange multiplier space to obtain continuous Lagrange multipliers.
- The second method uses the dual Lagrange multiplier space to obtain discontinuous Lagrange multipliers. This method allows for the effective elimination of Lagrange multipliers through static condensation, thereby obtaining a positive - definite system and improving computational efficiency.
### Mathematical formula representation
- **Mixed variational form**:
Given the mixed variational form of the Reissner - Mindlin plate equations as:
\[
a(\varphi, u; \psi, v)+b(\psi, v; \zeta)=\ell(v), \quad (\psi, v) \in H^1_0(\Omega) \times H^1_0(\Omega),
\]
\[
b(\varphi, u; \eta)-\frac{t^2}{\lambda(1 - t^2)}(\zeta, \eta)=0, \quad \eta \in L^2(\Omega),
\]
where:
- \( a(\varphi, u; \psi, v)=\int_{\Omega} C \varepsilon(\varphi): \varepsilon(\psi) \, dx+\lambda \int_{\Omega}(\varphi - \nabla u)\cdot(\psi - \nabla v) \, dx \)
- \( b(\psi, v; \eta)=\int_{\Omega}(\psi - \nabla v)\cdot \eta \, dx \)
- \( \ell(v)=\int_{\Omega} g v \, dx \)
- **Discretization**:
Using the finite - element spaces \( V_h \subset H^1_0(\Omega) \), \( W_h \subset H^1_0(\Omega) \), and \( M_h \subset L^2(\Omega) \) for discretization, we obtain the discrete saddle - point problem:
\[
a(\varphi_h, u_h; \psi_h, v_h)+b(\psi_h, v_h; \zeta_h)=(g, v_h), \quad (\psi_h, v_h) \in V_h \times S_h,
\]
\[
b(\varphi_h, u_h; \eta_h)-\frac{t^2}{\lambda(1 - t^2)}(\zeta_h, \eta_h)=0, \quad \eta_h