Work

Numerical solution of the phase-field equation with minimized discretization error

Year: 2012
Type: conference-paper
Abstract: Phase-field models are gaining importance as application-oriented tools for alloy and process optimization. In this context, numerical performance plays an essential role, but is often difficult to reconcile with accuracy. The numerical solution of the phase-field equation based on finite differences inherently implies a discretization error which may strongly bias the simulation results. A direct correlation between the accuracy of the finite-difference solution and the number of numerical cells resolving the diffuse interface profile has been observed. This presents a problem, because high numbers of interface cells are unfavorable for computational performance. To overcome the problem, an improved numerical solution of the phase-field equation is proposed, which allows obtaining highly accurate results with three or four interface cells only. The solution takes advantage of the a priori knowledge of the phase-field profile, determined by the choice of the potential function in the free energy functional. This knowledge enables an almost exact quantification and compensation of the error evoked by the discrete nature of the grid spacing and interface width. Benchmark simulations demonstrate the significant gain in both accuracy and numerical efficiency
Author: Janin Eiken
Language: English
Open Access status: bronze
OpenAlex ID: W2015504928
FWCI: 5.684
Cites: 5
Cited by: 34
Related to: 10