Abstract:In this article, we discuss a couple of nonlinear Galerkin method (NLG) in finite element set up for viscoelastic fluid flow, mainly equations of motion arising in the flow of 2D Oldroyd model. We obtain improved error estimate in $L^{\infty}(\bL^2)$ norm, which is optimal in nature, for linear finite element approximation, in view of the error estimate available in literature, in $L^2(\bH^1)$ norm.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to improve the accuracy of error estimation in the nonlinear Galerkin finite - element method (NLG) when solving viscoelastic fluid flow problems, especially for the first - order equations of the Oldroyd model. Specifically, the author aims to obtain the optimal error estimate in the \(L^\infty(L^2)\) norm and tries to improve the existing error estimation results.
### Overview of the Main Problem
1. **Background and Motivation**:
- When dealing with dissipative evolution partial differential equations with long - time integration, the traditional Galerkin method ignores the influence of some orthogonal spaces, which may lead to large errors.
- The nonlinear Galerkin method (NLG) improves the accuracy and stability of numerical solutions by considering the long - term influence of these ignored terms.
2. **Research Objectives**:
- For the first - order equations of the Oldroyd model, obtain the optimal error estimate in the \(L^\infty(L^2)\) norm.
- Improve the error estimation results in the existing literature, especially for the error estimation of Navier - Stokes equations and semilinear parabolic problems.
3. **Specific Problems**:
- How to handle nonlinear terms and time - dependent terms in the nonlinear Galerkin method to improve the order of convergence.
- Explore the influence of different types of nonlinear Galerkin methods (such as NLG I and NLG II) on error estimation.
### Formula Summary
- **Equations of Motion for the Oldroyd Model**:
\[
\frac{\partial u}{\partial t}+u\cdot\nabla u-\mu\Delta u-\int_{0}^{t}\beta(t - \tau)\Delta u(x,\tau)d\tau+\nabla p = f(x,t),\quad x\in\Omega,\quad t > 0
\]
\[
\nabla\cdot u = 0,\quad x\in\Omega,\quad t > 0
\]
\[
u(x,0)=u_{0}\text{ in }\Omega,\quad u = 0\text{ on }\partial\Omega,\quad t\geq0
\]
- **Error Estimation Formulas**:
- For the classical Galerkin method:
\[
\|(u - u_{h})(t)\|+h\|\nabla(u - u_{h})(t)\|\leq C e^{Ct}h^{2}t^{- 1/2}
\]
- For the nonlinear Galerkin method NLG I:
\[
\|(u_{h}-u_{h})(t)\|\leq K(t)H^{4}
\]
- For the nonlinear Galerkin method NLG II:
\[
\|(u_{h}-u_{h})(t)\|\leq K(t)H^{3}
\]
### Conclusion
By introducing new projection techniques and an improved nonlinear Galerkin method, the paper successfully improves the accuracy of error estimation for viscoelastic fluid flow problems, especially obtaining the optimal error estimate in the \(L^\infty(L^2)\) norm. This result is applicable not only to the Oldroyd model but also to Navier - Stokes equations.