betaFeedback

WorksMathematics

An Active Flux Method for Maxwell's equations

Abstract

We present a novel Active Flux method and a divergence-free method using exact evolution operators for solving Maxwell's equations in three space dimensions on Cartesian grids. The proposed methods are fully discrete, third-order accurate and rely on a truly multidimensional evolution operator based on Kirchhoff's formula for the electromagnetic wave equation. The first method is a classical Active Flux method with globally continuous reconstruction while the second method is an adaptation of the Active Flux method to a locally divergence-free reconstruction by Balsara. The applied reconstruction is built from facial degrees of freedom and is no longer globally continuous. The classical method keeps the divergence error bounded. The divergence-free method with face averages initialized by a potential keeps the constraints at round-off. We compare the accuracy of both methods for smooth and discontinuous test problems.

Introduction

In this publication, we present the first fully discrete Active Flux method for Maxwell’s equations in three dimensions. In 2011, Roe and his former student Eymann introduced Active Flux methods for advection and acoustics in one and two dimensions [1, 2]. For the simplest case of one-dimensional advection, the method agrees with van Leer’s method V from [3]. Its defining feature is a continuous reconstruction, which removes the need for Riemann solvers and allows the computation of numerical fluxes via Simpson’s rule. Roe later argued that third order represents a suitable candidate, being an odd-order method [4] with not too much complexity [5]. In one and two spatial dimensions, higher-order approximations have been considered [6, 7]. In this publication we restrict ourselves to a third-order accurate method. Besides the cell-average values, Active Flux uses point values as additional degrees of freedom. These define the reconstruction and enter the numerical fluxes. The point values allow for a compact stencil in space and time while retaining third-order accuracy. This makes Active Flux methods attractive to parallelization and capable of achieving high accuracy on coarse grids owing to their low numerical dissipation. Interest in this method grew, with extensions to Cartesian grids [8, 9, 10]. With appropriate evolution operators for the point values, Active Flux presents a finite volume approach that tries to incorporate the true physics and achieves a numerical domain of dependence in close agreement with the physical one. Recent publications continue to address this matter by developing evolution operators for the Euler equations [11, 12, 13, 14]. Although Active Flux was initially formulated as a fully discrete method, researchers have since developed semi-discrete methods. Among these are the method of Barsukow, Klingenberg and coworkers and their approach to the generalized Active Flux methods [7, 15, 16], which have also been applied to the ideal MHD equations [17], as well as Abgrall et al. who developed the PAMPA schemes that use the same underlying degrees of freedom [6, 18, 19]. Recently, works have been published that presented Active Flux methods for the Vlasov–Poisson and Vlasov–Maxwell systems [20, 21, 22]. In kinetic plasma models such as the Vlasov–Maxwell system, Maxwell’s equations describe the propagation of electromagnetic waves coupled to the Vlasov equation, which describes the evolution of a plasma [23]. In [22] the newly developed Active Flux methods for the Vlasov equation were combined with a classical Yee scheme to provide a new method for the Vlasov–Maxwell system. The authors showed that Active Flux methods for the Vlasov equation have significantly lower dissipation than the conservative PFC method [24]. This motivates our work, where we propose an Active Flux method for Maxwell’s equations in three dimensions. The main difficulty with such methods is constructing a suitable evolution operator. This can be resolved exactly using a multidimensional operator based on Kirchhoff’s formula. Combined with Active Flux’s low numerical dissipation, this promises high accuracy on coarse grids without the need for Riemann solvers or a separate time integration method.

The development of numerical methods for Maxwell’s equations has a long history. The most prominent method is the Yee scheme [25, 26], a second-order finite difference time-domain method employing a staggered grid in space and time. Many subsequent methods mimic the structure of the de Rham complex, starting from edge elements [27]. They evolve differential forms in function spaces chosen to respect the exterior calculus, following Arnold, Falk and Winther on finite element exterior calculus [28]. These structure-preserving discretizations include both finite difference [29] and finite element approaches [30], but are typically semi-discrete in time, requiring a suitable time integration method. Alternatively, the divergence constraints can be controlled by divergence cleaning [31] or enforced locally by divergence-free basis functions in discontinuous Galerkin methods [32].

In the present work we propose two fully discrete methods for Maxwell’s equations without source terms in three spatial dimensions. The first uses a tensor-product construction from lower dimensions, resulting in 27 degrees of freedom per cell for the second-degree polynomial reconstruction required for third-order accuracy. The time step is restricted by a CFL condition of 0.5 to maintain the compact stencil, and point values are evolved using a truly multidimensional exact evolution operator for the homogeneous Maxwell system. The evolution operator is derived from the wave equation and is based on Kirchhoff’s formula. This method preserves the divergence constraints only up to the truncation error of its reconstruction. The resulting divergence error stays bounded but does not vanish. This method is an extension of Active Flux methods to Maxwell’s equations in three dimensions. It relies on the previous ideas of evolution operators for wave dominated systems [11, 12, 13, 33, 34] and extends the ideas of Active Flux methods for Maxwell’s equations in two dimensions [11] to three dimensions. In 2D, Maxwell’s equations reduce to two sets of acoustic equations, where a discrete divergence is preserved by an exact evolution operator.

As the second method, we propose one that relies on the same evolution operator but has a reconstruction that is locally divergence-free. The reconstruction is based on the one derived by Balsara in [35, 36, 37] for MHD and applied to Maxwell’s equations in [38, 39]. In that setting, the reconstruction is designed to be divergence-free for the magnetic field. In our proposed method, we adapt this reconstruction to both the electric and magnetic field, to obtain a divergence-free method for Maxwell’s equations. In [37] the reconstruction is built from facial moments, which are reconstructed from the face averages of neighboring faces. In our approach, we compute these remaining degrees of freedom from point values on the cell faces, which are evolved using the same evolution operator as in the classical Active Flux method. Since the reconstruction is not continuous, the curl in the evolution operator is understood in the distributional sense. We use appropriate finite difference methods to compute the facial moments from the face point values.

We update the face averages of B\mathbf{B} and E\mathbf{E} via Stokes’ theorem to maintain the divergence at machine precision. The arising edge integrals are approximated using appropriate quadrature rules from the evolved point values. The reconstruction is not continuous and does not interpolate point values on the boundaries. Consequently, point values on the boundary of a cell are not uniquely defined at time tnt_n. Therefore, we replace Simpson’s rule as the quadrature rule in time by a right Radau rule. This gives up the continuity of the reconstruction for the divergence-free property.

The article is structured as follows: § 2 introduces the Active Flux framework. Section 3 derives the evolution operators for Maxwell’s equations without source terms. § 6 presents the divergence-free method. Numerical experiments are presented in § 7, and § 8 concludes this article.

Fully discrete Active Flux methods in three dimensions

To illustrate the formulation of Active Flux, we will derive the general form of the fully discrete method for a general three-dimensional conservation law on a domain Ω⊂R3\Omega\subset\mathbb{R}^{3}. Consider the vector of conserved quantities q:Ω×R+→Rm\mathbf{q} : \Omega\times\mathbb{R}_{+} \to\mathbb{R}^{m} with a flux function f\mathbf{f} defined by f(q):=(f(q),g(q),h(q)):Rm→R3×m\mathbf{f}(\mathbf{q}) := (f(\mathbf{q}),g(\mathbf{q}),h(\mathbf{q})) : \mathbb{R}^{m} \to\mathbb{R}^{3 \times m}. Then the general conservation law is given by the following partial differential equation

∂∂tq(x,t)+∇⋅f(q(x,t))=0,(1)\frac{\partial}{\partial t}\mathbf{q}(\mathbf{x},t)+\nabla\cdot\mathbf{f}(\mathbf{q}(\mathbf{x},t))=0, \tag*{(1)}
q(x,0)=q0(x),(2)\mathbf{q}(\mathbf{x},0)=\mathbf{q}_{0}(\mathbf{x}), \tag*{(2)}

with initial data q0(x)\mathbf{q}_{0}(\mathbf{x}) and x∈Ω\mathbf{x} \in\Omega. We restrict our formulation of Active Flux to a Cartesian grid of Ω\Omega with cells Ci,j,k=[xi−12,xi+12]×[yj−12,yj+12]×[zk−12,zk+12]C_{i,j,k} = [x_{i-\frac{1}{2}},x_{i+\frac{1}{2}}] \times[y_{j-\frac{1}{2}},y_{j+\frac{1}{2}}] \times[z_{k-\frac{1}{2}},z_{k+\frac{1}{2}}], i=1,…,Nxi = 1,\ldots,N_{x}, j=1,…,Nyj = 1,\ldots,N_{y}, k=1,…,Nzk = 1,\ldots,N_{z} and constant mesh widths Δx=xi+12−xi−12\Delta x = x_{i+\frac{1}{2}} - x_{i-\frac{1}{2}}, Δy=yj+12−yj−12\Delta y = y_{j+\frac{1}{2}} - y_{j-\frac{1}{2}} and Δz=zk+12−zk−12\Delta z = z_{k+\frac{1}{2}} - z_{k-\frac{1}{2}}. Integrating (1) over one cell Ci,j,kC_{i,j,k}, and a timestep [tn,tn+1][t_{n},t_{n+1}] with Δt=tn+1−tn\Delta t = t_{n+1} - t_{n}, one obtains

0=∫tntn+1∫Ci,j,k∂∂tq(x,t)+∇⋅f(q(x,t)) dx dt=∫Ci,j,kq(x,tn+1) dx−∫Ci,j,kq(x,tn) dx+∫tntn+1∫Ci,j,k∇⋅f(q(x,t)) dx dt.(3)\begin{aligned} 0 &= \int_{t_n}^{t_{n+1}}\int_{C_{i,j,k}}\frac{\partial}{\partial t}\mathbf{q}(\mathbf{x},t)+\nabla\cdot\mathbf{f}(\mathbf{q}(\mathbf{x},t))\,\mathrm{d}x\,\mathrm{d}t\\ &= \int_{C_{i,j,k}}\mathbf{q}(\mathbf{x},t_{n+1})\,\mathrm{d}x-\int_{C_{i,j,k}}\mathbf{q}(\mathbf{x},t_n)\,\mathrm{d}x\\ &\quad+\int_{t_n}^{t_{n+1}}\int_{C_{i,j,k}}\nabla\cdot\mathbf{f}(\mathbf{q}(\mathbf{x},t))\,\mathrm{d}x\,\mathrm{d}t. \tag*{(3)} \end{aligned}

Applying the divergence theorem with outer normal vector ν\boldsymbol{\nu} in the latter integral one obtains

References

  1. [1]Eymann, T., Roe, P.: Active Flux schemes. In: 49th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition (2011). https://doi.org/10.2514/6.2011-382 . AIAA Paper 2011-382
  2. [2]Eymann, T., Roe, P.: Active Flux schemes for systems. In: 20th AIAA Computational Fluid Dynamics Conference (2011). https://doi.org/10.2514/6.2011-3840 . AIAA Paper 2011-3840
  3. [3]van Leer, B.: Towards the ultimate conservative difference scheme. IV. A new approach to numerical convection. Journal of Computational Physics 23(3), 276–299 (1977) https://doi.org/10.1016/0021-9991(77)90095-X
  4. [4]Roe, P.: Musings of a computational philosopher. In: Tyacke, J.C., Vadlamani, N.R. (eds.) Proceedings of the Cambridge Unsteady Flow Symposium 2024, pp. 1–35. Springer, Cham (2025). https://doi.org/10.1007/978-3-031-69035-8_1
  5. [5]Roe, P.: Is discontinuous reconstruction really a good idea? Journal of Scientific Computing 73(2), 1094–1114 (2017) https://doi.org/10.1007/s10915-017-0555-z
  6. [6]Abgrall, R., Barsukow, W.: Extensions of Active Flux to arbitrary order of accuracy. ESAIM: Mathematical Modelling and Numerical Analysis 57(2), 991–1027 (2023) https://doi.org/10.1051/m2an/2023004
  7. [7]Barsukow, W., Chandrashekar, P., Klingenberg, C., Lechner, L.: A generalized Active Flux method of arbitrarily high order in two dimensions. Computers & Fluids 307, 106886 (2026) https://doi.org/10.1016/j.compfluid.2025.106886
  8. [8]Helzel, C., Kerkmann, D., Scandurra, L.: A new ADER method inspired by the Active Flux method. Journal of Scientific Computing 80(3), 1463–1497 (2019) https://doi.org/10.1007/s10915-019-00898-1
  9. [9]Barsukow, W., Hohm, J., Klingenberg, C., Roe, P.L.: The Active Flux scheme on Cartesian grids and its low Mach number limit. Journal of Scientific Computing 81(1), 594–622 (2019) https://doi.org/10.1007/s10915-019-01031-z
  10. [10]Chudzik, E., Helzel, C., Kerkmann, D.: The Cartesian grid Active Flux method: Linear stability and bound preserving limiting. Applied Mathematics and Computation 393, 125501 (2021) https://doi.org/10.1016/j.amc.2020.125501
  11. [11]Chudzik, E., Helzel, C., Lukáčová-Medviďová, M.: Active Flux methods for hyperbolic systems using the method of bicharacteristics. Journal of Scientific Computing 99(1), 16 (2024) https://doi.org/10.1007/s10915-024-02462-z
  12. [12]Chudzik, E., Helzel, C., Porfıetye, A.: A fully discrete truly multidimensional Active Flux method for the two-dimensional Euler equations. SIAM Journal on Scientific Computing 48(5), 2572–2598 (2026) https://doi.org/10.1137/25M1770928
  13. [13]Barsukow, W.: An Active Flux method for the Euler equations based on the exact acoustic solution operator. Journal of Scientific Computing 108(3), 87 (2026) https://doi.org/10.1007/s10915-026-03405-6
  14. [14]Helzel, C., Porfıetye, A.: A fully discrete Active Flux method for the Euler equations comparing different truly multi-dimensional evolution operators (2026). https://arxiv.org/abs/2606.21303
  15. [15]Duan, J., Barsukow, W., Klingenberg, C.: Active Flux methods for hyperbolic conservation laws—flux vector splitting and bound-preservation. SIAM Journal on Scientific Computing 47(2), 811–837 (2025) https://doi.org/10.1137/24M1658887
  16. [16]Abgrall, R., Barsukow, W., Klingenberg, C.: A semi-discrete Active Flux method for the Euler equations on Cartesian grids. Journal of Scientific Computing 102(2), 36 (2025) https://doi.org/10.1007/s10915-024-02749-1
  17. [17]Duan, J., Chandrashekar, P., Klingenberg, C.: Active flux for ideal magnetohydrodynamics: A positivity-preserving scheme with the Godunov-Powell source term (2025). https://arxiv.org/abs/2506.04857
  18. [18]Abgrall, R., Jiao, M., Liu, Y., Wu, K.: A novel and simple invariant-domain-preserving framework for PAMPA scheme: 1D case. SIAM Journal on Scientific Computing 47(6), 3536–3565 (2025) https://doi.org/10.1137/24M1715015
  19. [19][19*] Abgrall, R., Boscheri, W., Liu, Y.: Virtual finite element and hyperbolic problems: The PAMPA algorithm. Journal of Computational Physics 546, 114521 (2026) https://doi.org/10.1016/j.jcp.2025.114521
  20. [20]Hensel, L., Grünwald, G., Kormann, K., Grauer, R.: A split-step Active Flux method for the Vlasov-Poisson system. Journal of Computational Physics 540, 114294 (2025) https://doi.org/10.1016/j.jcp.2025.114294
  21. [21]Kiechle, Y., Chudzik, E., Helzel, C.: A positivity-preserving Active Flux method for the Vlasov-Poisson system. Journal of Computational Physics 524, 113693 (2025) https://doi.org/10.1016/j.jcp.2024.113693
  22. [22]Grünwald, G., Hensel, L., Deisenhofer, M., Lautenbach, S., Kormann, K., Grauer, R.: Solving the six-dimensional Vlasov–Maxwell system with Active Flux and splitting methods. Journal of Computational Physics 568, 115398 (2027) https://doi.org/10.1016/j.jcp.2026.115398
  23. [23]Kraus, M., Kormann, K., Morrison, P.J., Sonnendrücker, E.: GEMPIC: geometric electromagnetic particle-in-cell methods. Journal of Plasma Physics 83(4), 905830401 (2017) https://doi.org/10.1017/S002237781700040X
  24. [24]Filbet, F., Sonnendrücker, E., Bertrand, P.: Conservative numerical schemes for the vlasov equation. Journal of Computational Physics 172(1), 166–187 (2001) https://doi.org/10.1006/jcph.2001.6818
  25. [25]Yee, K.S.: Numerical solution of initial boundary value problems involving Maxwell’s equations in isotropic media. IEEE Transactions on Antennas and Propagation 14(3), 302–307 (1966) https://doi.org/10.1109/TAP.1966.1138693
  26. [26]Taflove, A., Hagness, S.C., Piket-May, M.: 9 - computational electromagnetics: The finite-difference time-domain method. In: CHEN, W.-K. (ed.) The Electrical Engineering Handbook, pp. 629–670. Academic Press, Burlington (2005). https://doi.org/10.1016/B978-012170960-0/50046-3
  27. [27]Nédélec, J.C.: Mixed finite elements in ℝ³. Numerische Mathematik 35(3), 315–341 (1980) https://doi.org/10.1007/BF01396415
  28. [28]Arnold, D.N., Falk, R.S., Winther, R.: Finite element exterior calculus, homological techniques, and applications. Acta Numerica 15, 1–155 (2006) https://doi.org/10.1017/S09624929062010018
  29. [29]Bach, D., Rueda-Ramírez, A.M., Sonnendrücker, E., Del Rey Fernández, D.C., Gassner, G.J.: SBP-FDEC: Summation-by-parts finite difference exterior calculus. Journal of Computational Physics, 115509 (2026) https://doi.org/10.1016/j.jcp.2026.115509
  30. [30]Kormann, K., Nazarov, M., Wen, J.: A structure-preserving finite element framework for the Vlasov–Maxwell system. Computer Methods in Applied Mechanics and Engineering 446, 118290 (2025) https://doi.org/10.1016/j.cma.2025.118290
  31. [31]Munz, C.-D., Omnes, P., Schneider, R., Sonnendrücker, E., Voß, U.: Divergence correction techniques for Maxwell solvers based on a hyperbolic model. Journal of Computational Physics 161(2), 484–511 (2000) https://doi.org/10.1006/jcph.2000.6507
  32. [32]Cockburn, B., Li, F., Shu, C.-W.: Locally divergence-free discontinuous galerkin methods for the maxwell equations. Journal of Computational Physics 194(2), 588–610 (2004) https://doi.org/10.1016/j.jcp.2003.09.007
  33. [33]Eymann, T.A., Roe, P.L.: Multidimensional active flux schemes. In: 21st AIAA Computational Fluid Dynamics Conference. American Institute of Aeronautics and Astronautics. https://doi.org/10.2514/6.2013-2940 . https://arc.aiaa.org/doi/10.2514/6.2013-2940 Accessed 2024-12-13
  34. [34]Lukáčová-Medvidová, M., Saibertová, J., Warnecke, G., Zahaykah, Y.: On evolution Galerkin methods for the Maxwell and the linearized Euler equations. Applications of Mathematics 49(5), 415–439 (2004) https://doi.org/10.1023/B:APOM.0000048121.68355.2a
  35. [35]Balsara, D.S.: Divergence-free adaptive mesh refinement for magnetohydrodynamics. Journal of Computational Physics 174(2), 614–648 (2001) https://doi.org/10.1006/jcph.2001.6917
  36. [36]Balsara, D.S.: Second-order-accurate schemes for magnetohydrodynamics with divergence-free reconstruction. The Astrophysical Journal Supplement Series 151(1), 149–184 (2004) https://doi.org/10.1086/381377
  37. [37]Balsara, D.S.: Divergence-free reconstruction of magnetic fields and WENO schemes for magnetohydrodynamics. Journal of Computational Physics 228(14), 5040–5056 (2009) https://doi.org/10.1016/j.jcp.2009.03.038
  38. [38]Hazra, A., Chandrashekar, P., Balsara, D.S.: Globally constraint-preserving FR/DG scheme for Maxwell’s equations at all orders. Journal of Computational Physics 394, 298–328 (2019) https://doi.org/10.1016/j.jcp.2019.06.003
  39. [39]Balsara, D.S., Käppeli, R.: von Neumann stability analysis of globally constraint-preserving DGTD and PNPM schemes for the Maxwell equations using multidimensional Riemann solvers. Journal of Computational Physics 376, 1108–1137 (2019) https://doi.org/10.1016/j.jcp.2018.09.051
  40. [40]Liu, X.-D., Osher, S., Chan, T.: Weighted essentially non-oscillatory schemes. Journal of Computational Physics 115(1), 200–212 (1994) https://doi.org/10.1006/jcph.1994.1187
  41. [41]Jiang, G.-S., Shu, C.-W.: Efficient implementation of weighted ENO schemes. Journal of Computational Physics 126(1), 202–228 (1996) https://doi.org/10.1006/jcph.1996.0130
  42. [42]LeVeque, R.J.: Wave propagation algorithms for multidimensional hyperbolic systems. Journal of Computational Physics 131(2), 327–353 (1997) https://doi.org/10.1006/jcph.1996.5603
  43. [43]Helzel, C.: A third-order accurate wave propagation algorithm for hyperbolic partial differential equations. Communications on Applied Mathematics and Computation 2(3), 403–427 (2020) https://doi.org/10.1007/s42967-019-00056-3
  44. [44]Hesthaven, J.S., Warburton, T.: Nodal Discontinuous Galerkin Methods: Algorithms, Analysis, and Applications. Springer, New York (2008). https://doi.org/10.1007/978-0-387-72067-8
  45. [45]Fan, D., Roe, P.L.: Investigations of a new scheme for wave propagation. In: 22nd AIAA Computational Fluid Dynamics Conference (2015). https://doi.org/10.2514/6.2015-2449 . AIAA Paper 2015-2449
  46. [46]Evans, L.C.: Partial Differential Equations, Second edition edn. Graduate studies in mathematics. American Mathematical Society, Providence, RI (2010)
  47. [47]Fambri, F., Sonnendrücker, E.: Structure preserving hybrid finite volume finite element method for compressible MHD. Journal of Computational Physics 523, 113691 (2025) https://doi.org/10.1016/j.jcp.2024.113691
  48. [48]Kanwal, R.P.: Generalized Functions: Theory and Applications, 3rd edn. Birkhäuser, Boston (2004). https://doi.org/10.1007/978-0-8176-8174-6
  49. [49]Helzel, C., Rossmanith, J.A., Taetz, B.: An unstaggered constrained transport method for the 3D ideal magnetohydrodynamic equations. Journal of Computational Physics 230(10), 3803–3829 (2011) https://doi.org/10.1016/j.jcp.2011.02.009

Paper details

vlnc:W_wm8vew34
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY-SA 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents