Virtual Element Methods for hyperbolic problems on polygonal meshes

Giuseppe Vacca
DOI: https://doi.org/10.48550/arXiv.1602.05781
2016-02-18
Abstract:In the present paper we develop the Virtual Element Method for hyperbolic problems on polygonal meshes, considering the linear wave equations as our model problem. After presenting the semi-discrete scheme, we derive the convergence estimates in H^1 semi-norm and L^2 norm. Moreover we develop a theoretical analysis on the stability for the fully discrete problem by comparing the Newmark method and the Bathe method. Finally we show the practical behaviour of the proposed method through a large array of numerical tests.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop a Virtual Element Method (VEM) for solving hyperbolic problems on polygonal meshes, and take the linear wave equation as a model problem. Specifically, the main objectives of the paper include: 1. **Propose a semi - discrete scheme**: The author first proposes a semi - discrete scheme to deal with the wave equation and derives the convergence estimates in the \(H^1\) semi - norm and the \(L^2\) norm. 2. **Analyze the stability of the fully - discrete scheme**: The author compares the Newmark method and the Bathe method and conducts a theoretical stability analysis of the fully - discrete scheme. 3. **Verify the effectiveness of the method**: Through a series of extensive numerical experiments, the actual performance of the proposed method is demonstrated, and its effectiveness and accuracy are verified. ### Main problem description The specific mathematical problem studied in the paper is the time - dependent wave equation, which is a typical hyperbolic partial differential equation. Consider the wave equation on a two - dimensional region \(\Omega\subset\mathbb{R}^2\): \[ \begin{cases} u_{tt}-\Delta u = f&\text{in }\Omega\times(0, T),\\ u = 0&\text{on }\partial\Omega\times(0, T),\\ u(\cdot, 0)=u_0,\quad u_t(\cdot, 0)=z_0&\text{in }\Omega \end{cases} \] where: - \(u\) is the unknown variable, - \(u_t\) and \(u_{tt}\) represent the first - order and second - order time derivatives of \(u\), respectively, - \(f\in L^2(\Omega\times(0, T))\) is the external force term, - the initial conditions \(u_0, z_0\in H^1_0(\Omega)\). ### Variational form The standard variational form of the above problem is: \[ \begin{cases} \text{Seek }u\in C^0(0, T; H^1_0(\Omega))\cap C^1(0, T; L^2(\Omega)),\text{ such that}\\ (u_{tt}(t), v)+a(u(t), v)=\langle f(t), v\rangle&\forall v\in H^1_0(\Omega),\text{ almost everywhere in }(0, T),\\ u(0)=u_0,\quad u_t(0)=z_0 \end{cases} \] where: - \((\cdot,\cdot)\) represents the standard \(L^2\) inner product on \(\Omega\), - \(a(\cdot,\cdot)\) represents the gradient form \(a(u, v)=(\nabla u,\nabla v)\), - \(\langle f(t),\cdot\rangle\) represents the dual product on \(H^1_0(\Omega)\). ### Main contributions of the paper 1. **Virtual element space and bilinear form**: Define the enhanced virtual element space and introduce an approximate discrete bilinear form to approximate the continuous gradient form and the \(L^2\) inner product. 2. **Error analysis**: Conduct an error analysis of the semi - discrete problem and obtain the convergence estimates in the \(H^1\) semi - norm and the \(L^2\) norm. 3. **Stability and convergence of the fully - discrete scheme**: Analyze in detail the stability and convergence of the Newmark method and the Bathe method in the fully - discrete scheme. 4. **Numerical experiments**: Verify the effectiveness and accuracy of the method through a large number of numerical tests, especially the performance under different meshes and time steps. In short, this paper aims to develop a...