Abstract. The local discontinuous Galerkin method for the numerical approximation of the time-harmonic Maxwell equations in a low-frequency regime is introduced and analyzed. Topologically nontrivial domains and heterogeneous media are considered, containing both conducting and insulating materials. The presented method involves discontinuous Galerkin discretizations of the curl-curl and grad-div operators, derived by introducing suitable auxiliary variables and so-called numerical fluxes. An hp-analysis is carried out and error estimates that are optimal in the meshsize h and slightly suboptimal in the approximation degree p are obtained.
IntroductionIn this paper, we propose and analyze an hp-local discontinuous Galerkin (LDG) method for the low-frequency time-harmonic Maxwell equations in heterogeneous media, containing both conducting and insulating materials: Find the complex field E that satisfiesin Ω 0 ⊂ Ω, (1.2) together with suitable boundary conditions (see Alonso and Valli [2] and Alonso [1]). Here, the field E is related to the electric field E by the identity E(x, t) = Re(E(x)e iωt ), where ω = 0 is a given frequency. The parameter µ = µ(x) is the magnetic permeability, ε = ε(x) the electric permittivity, J s is the phasor associated with a given current density, and σ = σ(x) is the electric conductivity, which is zero in the subdomain Ω 0 occupied by insulating materials. We remark that the electric field-based formulation in (1.1)-(1.2) is only one of several field-and potential-based formulations proposed in the literature for the solution of eddy current problems The main motivation for using discontinuous Galerkin (DG) methods for the numerical approximation of the above problem is that these methods, being based on discontinuous finite element spaces, can easily handle meshes with hanging nodes, elements of general shape, and local spaces of different types. Thus, they are ideally suited for hp-adaptivity and multiphysics or multimaterial problems. This flexibility in the mesh design is not shared in a straightforward way by standard edge or face elements commonly used in computational electromagnetics. Indeed, these elements are designed to enforce the continuity of either the tangential or the normal components of the fields across interelement boundaries (see, e.g., Nédéléc [37, 38], Bossavit [10, 11], and Monk [36]). This makes the handling of nonmatching grids and high-order approximations rather inconvenient from an implementational point of view. Nevertheless, efficient hp-adaptive edge element methods have been developed recently by Demkowicz and Vardapetyan [29,45].There are several possibilities for dealing with the divergence-free constraint (1.2) in the subregion Ω 0 . One way of imposing this constraint is to use mixed formulations, where new unknowns are introduced as Lagrange multipliers (see, e.g., Chen, Du and Zou [21], Demkowicz and Vardapetyan [29,45] and the references therein). Other approaches consist in regularizing the formulation in Ω 0 by adding suitable terms contai...