Controlling Long-Time Error Growth in Residual Correction Discretizations

  • Worku, Zelalem Arega (University of Waterloo)
  • Del Rey Fernandez, David (University of Waterloo)

Please login to view abstract download link

Hadamard-form entropy-stable discretizations have enabled robust implementation of high-order methods; however, they can be computationally expensive. In contrast, the residual correction method offers a more efficient, discretization-agnostic route to enforce discrete energy or entropy stability [1, 2]. While attractive in cost and generality, the method can be less robust than split-form entropy-stable schemes in some settings [3], and its impact on long-time accuracy has received limited attention. We study error growth for residual correction semidiscretizations in a simplified setting: the 1D linear advection equation with periodic boundary conditions. Although the advection equation is linear, the semidiscretization is nonlinear due to the correction term. For both discontinuous summation-by-parts (D-SBP) and continuous SBP (C-SBP) semidiscretizations, we recover the standard energy estimate and an a priori error bound with at most linear-in-time growth proportional to the truncation error, tau. We then analyze the residual correction formulation with C-SBP-type operators that do not satisfy the SBP property to machine precision. In this case, the correction introduces a truncation-error contribution that can lead to faster than linear-in-time error growth. Under mild smoothness assumptions, we prove an a priori bound of the form ||e(t)||_H <= ∫_0^t exp(c(t-s)) ||tau(s)||_H ds, revealing a potentially exponential-in-time factor controlled by a constant c that depends on the SBP defect. We derive explicit bounds linking c to the SBP property, quadrature weight ratios, and the spectral properties of the correction coefficient. These results motivate practical design principles, near-SBP operator construction, and tailored choices of the correction coefficients to reduce error growth while retaining stability. References [1] R. Abgrall, "A general framework to construct schemes satisfying additional conservation relations. Application to entropy conservative and entropy dissipative schemes," J. Comput. Phys., 372:640–666, 2018. [2] R. Abgrall, P. Öffner, and H. Ranocha, "Reinterpretation and extension of entropy correction terms for residual distribution and discontinuous Galerkin schemes: application to structure preserving discretization," J. Comput. Phys., 453:110955, 2022. [3] J. Markert and G. J. Gassner, "Comparison of different entropy stabilization techniques for discontinuous Galerkin spectral element methods," in Proc. ECCOMAS 2022, 2022.