Approximation of the phase-field transition system via fractional steps method

Theoretical approximation in context of PDEs (35A35) Probabilistic models, generic numerical methods in probability and statistics (65C20) Numerical aspects of computer graphics, image analysis, and computational geometry (65D18) Numerical computation of solutions to systems of equations (65H10) Numerical solutions to equations with nonlinear operators (65J15) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30)

Review text

The author studies the phase-field transition system in a bounded domain \(\Omega\subset \mathbb{R}^2\), i.e. \[ u_t+{l\over 2} \varphi_t- k\Delta u= f,\quad \tau\varphi_t- \xi^2\Delta\varphi= {1\over 2a} (\varphi- \varphi^3)+ 2u+ g\quad\text{in }\Omega\times (0,T), \] \[ {\partial u\over\partial\nu}+ hu= 0,\;\varphi= 0\quad\text{on }\partial\Omega\times (0,T),\quad u(.,0)= u_0,\;\varphi(.,0)= \varphi_0\quad\text{in }\Omega. \] Here, \(u\) denotes the reduced temperature, and \(\varphi\) is the order parameter. The above problem is approximated by a sequence of linear systems on the subintervals \((k\varepsilon,(k+ 1)\varepsilon)\subset (0,T)\), \(k= 0,\dots, N_\varepsilon\) as follows: \[ u^\varepsilon_t+{l\over 2} \varphi^\varepsilon_t- k\Delta u^\varepsilon= f,\quad\tau\varphi^\varepsilon_t- \xi^2\Delta\varphi^\varepsilon= {1\over 2a} \varphi^\varepsilon+ 2u^\varepsilon+ g\quad\text{in }\Omega\times (k\varepsilon,(k+ 1)\varepsilon), \] \[ {\partial u^\varepsilon\over\partial\nu}+ hu^\varepsilon= 0,\;\varphi^\varepsilon= 0\quad\text{on }\partial\Omega\times (k\varepsilon,(k+ 1)\varepsilon), \] \[ \varphi^\varepsilon_+(.,k\varepsilon)= z(\varepsilon, \varphi^\varepsilon_-(., k\varepsilon)),\;u^\varepsilon_+(., k\varepsilon)= u^\varepsilon_-(., k\varepsilon)\quad\text{in }\Omega, \] and \(z(t,x)\) is the solution of \(z'+ (1/2a)z^3= 0\), \(z(0,x)= x\). Here \(f_+\) and \(f_-\) denote upper and lower limits of \(f\). The main result of the paper says that \(\lim_{\varepsilon\to 0}(u^\varepsilon(t), \varphi^\varepsilon(t))= (u(t),\varphi(t))\) in \(L^2(\Omega)\times L^2(\Omega)\) for all \(t\in [0,T]\). Based on this convergence result, the author introduces a finite element method and presents numerical results.

Recommendations

Cites work

Cited in

(25)

This page was built for publication: Approximation of the phase-field transition system via fractional steps method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4385298)