Accelerating a phase field method by linearization for eigenfrequency topology optimization

Topology optimization of eigenfrequencies has significant applications in science, engineering, and industry. Eigenvalue problems as constraints of optimization with partial differential equations are solved repeatedly during optimization and design process. The nonlinearity of the eigenvalue problem leads to expensive numerical solvers and thus requires huge computational costs for the whole optimization process. In this paper, we propose a simple yet efficient linearization approach and use a phase field method for topology optimization of eigenvalue problems with applications in two models: vibrating structures and photonic crystals. More specifically, the eigenvalue problem is replaced by a linear source problem every few optimization steps for saving computational costs. Numerical evidence suggests first-order accuracy of approximate eigenvalues and eigenfunctions with respect to the time step and mesh size. Numerical examples are presented to illustrate the effectiveness and efficiency of the algorithms.