Abstract
We propose to use a hybridizable discontinuous Galerkin (HDG) method combined with the continuous Galerkin (CG) method to approximate Maxwell's equations. We make two contributions in this paper. First, even though there are many papers using HDG methods to approximate Maxwell's equations, to our knowledge they all assume that the coefficients are smooth (or constant). Here, we derive optimal convergence estimates for our HGD-CG approximation when the electromagnetic coefficients are piecewise smooth. This requires new techniques of analysis. Second, we use CG elements to approximate the Lagrange multiplier used to enforce the divergence condition and we obtain a discrete system in which we can decouple the discrete the Lagrange multiplier. Because we are using a continuous Lagrange multiplier space, the number of degrees of freedom devoted to this are less than for other HDG methods. We present numerical experiments to confirm our theoretical results.
Introduction
Maxwell’s equations are fundamental to electromagnetics, underpinning technologies from radar systems to medical imaging. However, solving the "indefinite" time-harmonic version of these equations numerically is notoriously difficult. Standard Finite Element Methods (FEM) often suffer from "spurious modes" or require excessive computational resources at high frequencies. While Hybridizable Discontinuous Galerkin (HDG) methods have emerged as a powerful alternative due to their efficiency and scalability, existing mathematical proofs for their accuracy typically rely on the assumption that material properties (permittivity and permeability) are perfectly smooth or constant throughout the domain.
In real-world applications, materials are rarely uniform; they possess boundaries, layers, and sharp jumps in properties. This paper bridges the gap between theoretical analysis and heterogeneous materials by proposing a hybrid HDG-CG scheme that maintains optimal convergence even under minimal regularity.
The Proposed HDG-CG Method
The core of the approach involves a specific formulation of the time-harmonic Maxwell equations, incorporating a Lagrange multiplier $p$ to enforce the divergence constraint. This ensures the physical validity of the solution in the presence of source terms.
The authors propose a hybrid scheme with two distinct components:
- HDG for the Electric Field ($u$): The electric field is approximated using HDG, which allows for "static condensation." This technique eliminates most variables locally within each element, leaving only a global system for the "traces" on element boundaries, drastically reducing computational costs.
- CG for the Lagrange Multiplier ($p$): Unlike traditional HDG methods that use discontinuous spaces for all variables, this method employs a Continuous Galerkin (CG) space for the Lagrange multiplier.
This specific combination allows the Lagrange multiplier to be decoupled from the main system, resulting in a Symmetric Positive Definite (SPD) system for $p$ that can be solved efficiently using standard solvers like Multigrid (MG).
Key Innovations and Theoretical Breakthroughs
Breaking the Smoothness Barrier
Most HDG proofs fail when material coefficients jump because the associated "dual problem" lacks sufficient regularity. The authors bypassed this limitation by utilizing the "M-decomposition" framework and constructing new stable, commuting quasi-interpolation operators. This allows for the derivation of optimal convergence estimates—$O(h^{k+1})$ for the electric field and $O(h^k)$ for its curl—even when coefficients are only piecewise smooth.
Efficiency and Stability
By using a continuous space for the Lagrange multiplier, the method significantly reduces the total number of degrees of freedom (DoF) compared to "pure" HDG methods. Furthermore, the method remains stable even as the wave number $\kappa$ approaches zero, avoiding the "low-frequency breakdown" that plagues many other electromagnetic solvers.
Numerical Validation
The method was evaluated using three rigorous 3D benchmarks on a unit cube:
- Example 1: Constant coefficients with varying wave numbers.
- Example 2: Piecewise smooth coefficients with a smooth exact solution.
- Example 3: Both coefficients and the exact solution are piecewise smooth, simulating the jump between two different materials.
In all cases, the numerical tests confirmed the theoretical optimal convergence rates. The method performed robustly regardless of whether the material was uniform or had sharp jumps in properties, and the decoupling of the Lagrange multiplier proved effective in reducing total solve times.
Conclusion and Future Work
This HDG-CG approach provides a mathematically rigorous and computationally efficient framework for solving Maxwell's equations in complex, heterogeneous media. Future research will focus on extending this work to include Robin-type (impedance) boundary conditions for radiating waves and addressing the "pollution effect" at very high frequencies. Additionally, extending certain superconvergence properties from 2D to 3D remains an important open challenge in the field.
Summary updated