Numerical approximation of the fractional Cahn-Hilliard equation by operator splitting method (Q2189413)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Numerical approximation of the fractional Cahn-Hilliard equation by operator splitting method |
scientific article |
Statements
Numerical approximation of the fractional Cahn-Hilliard equation by operator splitting method (English)
0 references
15 June 2020
0 references
This paper studies the following fractional Cahn-Hilliard (CH) equation \[ \begin{split} \displaystyle \dfrac{\partial u}{\partial t}=\Delta\left( (-\Delta)^{\frac{\alpha}{2}}u+\dfrac{1}{\varepsilon^2}f(u) \right) \,\, (x,t) \in \Omega\times (0,T], \end{split} \] subject to the free boundary conditions: \[ \Delta u\cdot n_{\Omega}=0, \,\,\,\nabla\left( (-\Delta)^{\frac{\alpha}{2}}u+\dfrac{1}{\varepsilon^2}f(u) \right) \cdot n_{\Omega}=0 , \,\,\, x\in \partial\Omega, \] and the initial condition: \[ u(x,0)=u_0(x) , \,\,\, x\in \Omega, \] where \(\Omega\) is a bounded domain in \(\mathbb{R}^d\) \((d = 1, 2, 3)\). The operator \((-\Delta)^{\frac{\alpha}{2}}\) with \(\alpha \in (1, 2]\) is the fractional Laplacian operator. The function \(f(u) = \frac{dF(u)}{du}\) with \(F(u) = \frac{1}{4}(u^2-1)^2\) being a given Helmholtz free energy density. The small parameter \(\varepsilon\) is known as the interaction length. Here, the authors propose a second-order operator splitting method based on a pseudospectral method and the strong stability preserving Runge-Kutta (SSP-RK) method to speed up the simulation for the above CH equation. The original problem is split into linear and nonlinear subproblems. For the linear part \(u_t=\Delta(-\Delta)^{\frac{\alpha}{2}}u\), the pseudo-spectral method is adopted, and thus an ordinary differential equation is obtained. For the nonlinear part \(u_t=\frac{1}{\varepsilon^2}\Delta f(u)\), a second-order SSP-RK method together with the pseudo-spectral method is used. The stability and convergence of the proposed method in the \(L^2\)-norm are studied. The authors establish a recursive relationship of the errors in the time direction and then, applying the discrete Gronwall inequality, obtain the global errors for the operator splitting method. Some numerical results are obtained to demonstrate the accuracy and efficiency of the numerical algorithm by using computational simulation of the fractional Cahn-Hilliard equation for a variety of choices of fractional order \(\alpha\).
0 references
fractional-in-space Cahn-Hilliard equation
0 references
operator splitting method
0 references
pseudo-spectral method
0 references
SSP-RK method
0 references
stability
0 references
convergence
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references