On energy stable discontinuous Galerkin spectral element approximations of the perfectly matched layer for the wave equation

• Published in: arXiv.org
• Main Authors: Duru, Kenneth, Alice-Agnes Gabriel, Kreiss, Gunilla
• Format: Paper Journal Article
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 18.02.2018
• Subjects: Acoustic waves; Acoustics; Boundary conditions; Boundary element method; Convergence; Damping; Differential equations; Energy; Estimates; Fluxes; Galerkin method; Mathematics - Numerical Analysis; Numerical methods; Numerical stability; Perfectly matched layers; Spectral element method; Wave equations; Well posed problems
• ISSN: 2331-8422
• DOI: 10.48550/arxiv.1802.06388

We develop a provably energy stable discontinuous Galerkin spectral element method (DGSEM) approximation of the perfectly matched layer (PML) for the three and two space dimensional (3D and 2D) linear acoustic wave equations, in first order form, subject to well-posed linear boundary conditions. First, using the well-known complex coordinate stretching, we derive an efficient un-split modal PML for the 3D acoustic wave equation. Second, we prove asymptotic stability of the continuous PML by deriving energy estimates in the Laplace space, for the 3D PML in a heterogeneous acoustic medium, assuming piece-wise constant PML damping. Third, we develop a DGSEM for the wave equation using physically motivated numerical flux, with penalty weights, which are compatible with all well-posed, internal and external, boundary conditions. When the PML damping vanishes, by construction, our choice of penalty parameters yield an upwind scheme and a discrete energy estimate analogous to the continuous energy estimate. Fourth, to ensure numerical stability when PML damping is present, it is necessary to systematically extend the numerical numerical fluxes, and the inter-element and boundary procedures, to the PML auxiliary differential equations. This is critical for deriving discrete energy estimates analogous to the continuous energy estimates. Finally, we propose a procedure to compute PML damping coefficients such that the PML error converges to zero, at the optimal convergence rate of the underlying numerical method. Numerical experiments are presented in 2D and 3D corroborating the theoretical results.