A second-order accurate numerical method with unconditional energy stability for the Lifshitz-Petrich equation on curved surfaces (Q7023893)

This paper contributes to the development of numerical techniques for partial differential equations posed on complex geometries and surfaces. In particular, the authors present a new numerical algorithm for solving the Lifshitz-Petrich equation on closed surfaces, with the aim of modeling crystal and quasicrystal formation.\N\NThe authors propose a spatial discretization based on a triangular mesh of the surface, which allows for an explicit construction of a discrete Laplace-Beltrami operator using the neighborhood relationships between mesh elements. For temporal discretization, the Backward Differentiation Formula (BDF) scheme is combined with the Scalar Auxiliary Variable (SAV) approach. The resulting method is second-order accurate in both space and time.\N\NIn this framework, the nonlinear system that arises at each time step is reformulated into a linear one by introducing an extrapolation of the unknown approximation at time \(t_{n+1}\) within the coefficient matrix, which significantly simplifies the computational procedure. The resulting linear system is solved using the biconjugate gradient stabilized method with incomplete LU decomposition as a preconditioner.\N\NAs is characteristic of SAV-based methods, the scheme achieves energy stability by satisfying a dissipation law associated with a modified discrete energy rather than the original physical energy.\N\NSeveral numerical experiments are performed to validate the method, including discrete dissipation tests for the modified energy, comparisons between torus and flat surfaces, and simulations of crystal interface evolution. The results demonstrate the effectiveness of the proposed numerical method in simulating crystallization processes on curved surfaces.