Adaptive time-step control for a monolithic multirate scheme coupling the heat and wave equation

TL;DR

This paper develops an adaptive monolithic multirate time-stepping scheme for coupled heat and wave equations, incorporating decoupling methods and an a posteriori error estimator for efficient and accurate simulations.

Contribution

It introduces a novel monolithic multirate scheme with adaptive time-stepping and decoupling strategies for coupled parabolic-hyperbolic problems.

Findings

Efficient solution of large algebraic systems via decoupling methods.

Validated accuracy of the a posteriori error estimator.

Demonstrated effectiveness through numerical experiments.

Abstract

We consider the dynamics of a parabolic and a hyperbolic equation coupled on a common interface and develop time-stepping schemes that can use different time-step sizes for each of the subproblems. The problem is formulated in a strongly coupled (monolithic) space-time framework. Coupling two different step sizes monolithically gives rise to large algebraic systems of equations where multiple states of the subproblems must be solved at once. For efficiently solving these algebraic systems, we inherit ideas from the partitioned regime and present two decoupling methods, namely a partitioned relaxation scheme and a shooting method. Furthermore, we develop an a posteriori error estimator serving as a mean for an adaptive time-stepping procedure. The goal is to optimally balance the time step sizes of the two subproblems. The error estimator is based on the dual weighted residual method and relies on the space-time Galerkin formulation of the coupled problem. As an example, we take a linear set-up with the heat equation coupled to the wave equation. We formulate the problem in a monolithic manner using the space-time framework. In numerical test cases, we demonstrate the efficiency of the solution process and we also validate the accuracy of the a posteriori error estimator and its use for controlling the time step sizes.