Second-order accurate, maximum principle-preserving, and convergent schemes for the phase-field shape transformation model
摘要
We present linear maximum principle-preserving numerical schemes with second-order accuracy for a new phase-field model of shape transformation. The leap-frog strategy is adopted to update the solution in time and the finite difference method is used to implement the spatial discretization. The fully discretized schemes in two- and three-dimensional spaces are described in detail. For each scheme, we analytically estimate the discrete maximum principle-preserving property and derive different stability conditions. Moreover, the associated convergence estimations are established. In the numerical experiments, we first validate the accuracy and maximum principle-preserving property via the simulations of different shape transformations. The shape transformation of dendritic crystal is performed to show the merit of present model. The image transformations are conducted to show the capability of proposed method in restoring original images. Finally, several three-dimensional simulations are implemented to further display the effectiveness in actual applications.