A structure-preserving parametric finite element method for area-conserved generalized mean curvature flow

TL;DR

This paper introduces a novel structure-preserving parametric finite element method for simulating 2D area-conserved generalized mean curvature flow, ensuring key geometric properties are maintained during numerical evolution.

Contribution

The paper develops a new finite element scheme that rigorously preserves area and perimeter decrease in simulating curvature flows, with proven theoretical properties and validated numerical experiments.

Findings

The scheme conserves enclosed area during evolution.

The scheme guarantees perimeter decrease over time.

Numerical experiments confirm convergence and stability.

Abstract

We propose and analyze a structure-preserving parametric finite element method (SP-PFEM) to simulate the motion of closed curves governed by area-conserved generalized mean curvature flow in two dimensions (2D). We first present a variational formulation and rigorously prove that it preserves two fundamental geometric structures of the flows, i.e., (a) the conservation of the area enclosed by the closed curve; (b) the decrease of the perimeter of the curve. Then the variational formulation is approximated by using piecewise linear parametric finite elements in space to develop the semi-discrete scheme. With the help of the discrete Cauchy's inequality and discrete power mean inequality, the area conservation and perimeter decrease properties of the semi-discrete scheme are shown. On this basis, by combining the backward Euler method in time and a proper approximation of the unit normal vector, a structure-preserving fully discrete scheme is constructed successfully, which can preserve the two essential geometric structures simultaneously at the discrete level. Finally, numerical experiments test the convergence rate, area conservation, perimeter decrease and mesh quality, and depict the evolution of curves. Numerical results indicate that the proposed SP-PFEM provides a reliable and powerful tool for the simulation of area-conserved generalized mean curvature flow in 2D.