Table of Contents
Fetching ...

An initial-corrected splitting approach for convection-diffusion-reaction problems

Thi Tam Dang, Lukas Einkemmer, Alexander Ostermann

TL;DR

The paper tackles order reduction observed in Strang splitting for convection-diffusion-reaction problems with Dirichlet boundaries. It introduces an initial-corrected splitting that subtracts the step’s initial data to enforce homogeneous Dirichlet conditions, then applies a Strang-type splitting to the transformed system; for time-dependent boundaries, a first-order correction $z_n(t)$ is used. The authors prove a global second-order convergence bound $∥u_n-u(t_n)∥ ≤ C τ^2 (1 + |log τ|)$ under standard analytic-semigroup and Lipschitz assumptions, and validate the theory with numerical experiments in 1D and 2D showing robust second-order accuracy across boundary condition types. Overall, the method provides a simple, efficient, and reliable way to avoid order reduction in practical convection-diffusion-reaction simulations with Dirichlet BCs.

Abstract

Splitting methods constitute a well-established class of numerical schemes for solving convection-diffusion-reaction problems. They have been shown to be effective in solving problems with periodic boundary conditions. However, in the case of Dirichlet boundary conditions, order reduction has been observed even with homogeneous boundary conditions. In this paper, we propose a novel splitting approach, the so-called `initial-corrected splitting method', which succeeds in overcoming order reduction. A convergence analysis is performed to demonstrate second-order convergence of this modified Strang splitting method. Furthermore, we conduct numerical experiments to illustrate the performance of the newly developed splitting approach.

An initial-corrected splitting approach for convection-diffusion-reaction problems

TL;DR

The paper tackles order reduction observed in Strang splitting for convection-diffusion-reaction problems with Dirichlet boundaries. It introduces an initial-corrected splitting that subtracts the step’s initial data to enforce homogeneous Dirichlet conditions, then applies a Strang-type splitting to the transformed system; for time-dependent boundaries, a first-order correction is used. The authors prove a global second-order convergence bound under standard analytic-semigroup and Lipschitz assumptions, and validate the theory with numerical experiments in 1D and 2D showing robust second-order accuracy across boundary condition types. Overall, the method provides a simple, efficient, and reliable way to avoid order reduction in practical convection-diffusion-reaction simulations with Dirichlet BCs.

Abstract

Splitting methods constitute a well-established class of numerical schemes for solving convection-diffusion-reaction problems. They have been shown to be effective in solving problems with periodic boundary conditions. However, in the case of Dirichlet boundary conditions, order reduction has been observed even with homogeneous boundary conditions. In this paper, we propose a novel splitting approach, the so-called `initial-corrected splitting method', which succeeds in overcoming order reduction. A convergence analysis is performed to demonstrate second-order convergence of this modified Strang splitting method. Furthermore, we conduct numerical experiments to illustrate the performance of the newly developed splitting approach.

Paper Structure

This paper contains 9 sections, 1 theorem, 63 equations, 2 figures, 2 algorithms.

Key Result

Theorem 3.1

Let the Assumptions Bas1-Bas3 be satisfied. Then there exists a constant $\tau_0>0$ such that for all step sizes $0<\tau \le \tau_0$ and $t_{n}=n\tau$ we have that the modified Strang splitting applied to B1.1 satisfies the global error bound where the constant $C$ depends on $T$ but is independent of $\tau$ and $n$.

Figures (2)

  • Figure 1: The absolute error in the discrete infinity norm is computed at $t=1$ by comparing the numerical solution to a reference solution.
  • Figure 2: The absolute error in the discrete infinity norm is computed at $t=1$ by comparing the numerical solution to a reference solution. Slope $2$ is displayed by a dash-dotted line.

Theorems & Definitions (4)

  • Theorem 3.1
  • Example 4.1
  • Example 4.2
  • Example 4.3