Abstract
In spinodal decomposition phenomena, the initial perturbations evolve at a faster time scale while there is a slower growth at a later time. Therefore, using uniform small time-steps for tracking the fast dynamics is computationally expensive. On the other hand, uniform large time-steps may overlook the rapid changes. In this article, we propose an adaptive time-stepping scheme for faster time scale simulations of spinodal decomposition by solving the Cahn-Hilliard equation numerically. We consider the double well potential having a polynomial of 6th-order along with 4th-order polynomial for the variable mobility. A diagonally implicit fractional step θ-scheme(DIFSTS) for temporal discretization while conforming finite element method for spatial discretization is used. Accuracy and efficiency of the method are given and simulations of 2D spinodal decomposition are illustrated graphically.
| Original language | English |
|---|---|
| Article number | e201800246 |
| Journal | ZAMM ‐ Journal of Applied Mathematics and Mechanics / Zeitschrift für Angewandte Mathematik und Mechanik |
| Volume | 99 |
| Issue number | 7 |
| DOIs | |
| State | Published - Jul 2019 |
| Externally published | Yes |
Bibliographical note
Publisher Copyright:© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Keywords
- 65M60
- DUNE-PDELab
- adaptive time-stepping
- cahn-hilliard equation
- diagonally implicit fractional step θ-scheme
- spinodal decomposition
ASJC Scopus subject areas
- Computational Mechanics
- Applied Mathematics