Computation of Maxwell s equations on Manifold using DEC

Transcription

1 arxiv: v6 [math.na] 9 Dec 009 Computation of Maxwell s equations on Manifold using DEC Zheng Xie Yujie Ma. Center of Mathematical Sciences, Zhejiang University (3007),China. Key Laboratory of Mathematics Mechanization, Chinese Academy of Sciences, (00090), China Abstract In this paper, the method of discrete exterior calculus for numerically solving Maxwell s equations in space manifold and the time is discussed, which is a kind of lattice gauge theory. The analysis of its stable condition and error is also accomplished. This algorithm has been implemented on C++ plateform for simulating TE/M waves in vacuum. Keywords: Discrete exterior calculus, Discrete variation, Maxwell s equations, Lattice gauge theory. PACS(00): 4.0.Jb, 0.30.Jr, 0.40.Sf, 0.60.Cb. Introduction Computational electromagnetism is concerned with the numerical study of Maxwell s equations. The Yee scheme is known as finite difference time domain and is one of the most successful numerical methods, particularly in the area of microwave problems []. It preserves important structural lenozhengxie@yahoo.com.cn yjma@mmrc.iss.ac.cn This work is partially supported by CPSFFP (No ), NKBRPC (No. 004CB38000), and NNSFC (No )

2 features of Maxwell s equations [ 6]. Bossavit et al present the Yee-like scheme and extend Yee scheme to unstructured grids. This scheme combines the best attributes of the finite element method (unstructured grids) and Yee scheme (preserving geometric structure) [7, 8]. Stern et al [9] generalize the Yee scheme to unstructured grids not just in space, but in 4-dimensional spacetime by discrete exterior calculus(dec) [0 0]. This relaxes the need to take uniform time steps. In this paper, we generalize the Yee scheme to the discrete space manifold and the time. The spacetime manifold used here is split as a product of D time and D or 3D space manifold. The space manifold can be approximated by triangular and tetrahedrons depending on dimension, and the time by segments. So the spacetime manifold is approximated by prism lattice, on which the discrete Lorentz metric can be defined.. With the technique of discrete exterior calculus, the R value discrete connection, curvature and Bianchi identity are defined on prim lattice. With discrete variation of an inner product of discrete forms and their dual forms, the discrete source equation and continuity equation are derived.. Those equations compose the discrete Maxwell s equations in vacuum case, which just need the local information of triangulated manifold such as length and area. The discrete Maxwell s equations here can be re-grouped into two sets of explicit iterative schemes for TE and TM waves, respectively. Those schemes can directly use acute triangular, rectangular, regular polygon and their combination, which has been implemented on C++ plateform to simulate the electromagnetic waves propagation and interference on manifold. DEC for Maxwell s equations Maxwell s equations can be simply expressed once the language of exterior differential forms is used. The electric and magnetic fields are jointly described by a curvature form F in a 4-D spacetime manifold. The Maxwell s equations reduce to the Bianchi identity and the source equation df = 0 d F = J ()

3 where d denotes the exterior differential operator, denotes the Hodge star operator, and -form J is called the electric current form satisfying the continuity equation d J = 0. As the exterior derivative is defined on any manifold, the differential form version of the Bianchi identity makes sense for any 3D or 4D spacetime manifold, whereas the source equation is defined if the manifold is oriented and has a Lorentz metric. Now, we introduce the discrete counterpart of those differential geometric objects to derive the numerical computational schemes for Maxwell s equations. Discrete Lorentz metric The spacetime manifold used here is split as a product of D time and D or 3D space manifold. The D or 3D space manifold can be approximated by triangular or tetrahedrons, and the time by segments. The length of edge and area of triangular and volume of tetrahedrons gives the discrete Riemann metric on space grids. The metric on time grid is the minus of length square. The spacetime manifold is approximated by prism lattice, on which the discrete Lorentz metric can be defined as the product of discrete metric on space and time. Discrete exterior calculus A discrete differential k-form, k Z, is the evaluation of the differential k- form on all k-simplices. Dual forms, i.e., forms that we evaluate on the dual cell. Suppose each simplex contains its circumcenter. The circumcentric dual cell D(σ 0 ) of simplex σ 0 is D(σ 0 ) := Int(c(σ 0 )c(σ ) c(σ r )), σ 0 σ σ r where σ i is all the simplices which contains σ 0,..., σ i, and c(σ i ) is the circumcenter of σ i. The two operators in Eqs.() can be discretized as follows:. Discrete exterior differential operator d, this operator is the transpose of the incidence matrix of k-cells on k +-cells. 3

4 . Discrete Hodge Star, the operator scales the cells by the volumes of the corresponding dual and primal cells. Discrete connection and curvature Discrete connection form or gauge field A assigns to each element in the set of edges E an element of the gauge group R: A : E R. Discrete curvature form is the discrete exterior derivative of the discrete connection form F = da : P R. The value of F on each element in the set of triangular P is the coefficient of Holonomy group of this face. The form F automatically satisfies the discrete Bianchi identity df = 0. () Note that since the gauge group R used here is abelian, we need not pick a starting vertex for the loop. We may traverse the edges in any order, so long as taking orientations into account. Discrete Maxwell s equations For source case, we need discrete current form J. Let A = E Lagrangian functional be A i and the L(A,J) = da,da + A,J, where da,da := (A) E (d) E F ( ) F F (d) T F E (A)T E A,J := (A) E ( ) E E (J T ) E Supposing that there is a variation of A i, vanishing on the boundary, we 4

5 have Ai L(A,J) = Ai ( da,da + A,J ) = Ai ( (A) E (d) E F ( ) F F (d) T F E (A)T E +(A) E ( ) E E (J T ) E ) = (0,..., }{{},...,0) E (d) E F ( ) F F (d) T F E (A)T E i (A) E (d) E F ( ) F F (d) T F E (0,..., }{{},...,0) T E i +(0,..., }{{},...,0) E ( ) E E (J T ) E i = (0,..., }{{},...,0) E (d) E F ( ) F F (d) T F E (A)T E i +(0,..., }{{},...,0) E ( ) E E (J T ) E i The Hamilton s principle of stationary action states that this variation must equal zero for any such vary of A i, implying the Euler-Lagrange equations (d) E F ( ) F F (d) T F E (A)T E +( ) E E (J T ) E = 0, which is the discrete source equation δf = J, (3) where δ = d T. Since (d T ) = 0, the discrete continuity equation can express as: d T J = 0. (4) The equations of discrete Bianchi identity (), source equation (3), and continuity equation (4) are called discrete Maxwell s equations. Discrete Gauge transformations Discrete gauge transformations are maps A A+df for any 0 form or scalar function f on vertex. Since the discrete exterior derivative maps F F +d f = F, 5

6 the discrete Maxwell s equations(-4) are invariant under discrete gauge transformations. Since the discrete continuity equation (4) ensures so we have df,j = (f) V (d T ) V E ( ) E E (J T ) E = 0, L(A+df,J) = d(a+df),d(a+df) + A+df,J = L(A,J). That is to say the Lagrangian function is also invariant under discrete gauge transformations. 3 Explicit schemes Schemes for TE wave The discrete current form, discrete curvature form and its dual can be written as J = ( ρ e dt,j e ) F n+ = E n+ dt+b n F n = H n dt D n, where n and n+ denote the coordinate of the time, E = E i e i (electric E field) is discrete form on space, B = B i P i (magnetic field) is discrete P form on space, H = H i P i (magnetizing field) is the dual of B on P space, D = D i e i (electric displacement field) is the dual of E on space, E ρ e dt (charge density) is the discrete form on time, J e = J ei e i (electric E current density) is the discrete form on space. The discrete Maxwell s equations can be rewritten as d s B n = 0 d s E n+ dt = d t B n d T sd n = (ρ e dt) n d T sh n dt = d T t D n + Je, n where d s, d T s are the restriction of d, dt on space, and d t B n := Bn+ B n dt d T t D n := Dn+ D n 6 dt. (5)

7 If allowing for the possibility of magnetic charges and current discrete 3 form J = (ρ m, J m dt), the symmetric scheme can be written as where d s B n = ρ n m d s E n+ dt = d t B n J n+ m dt d T sd n = (ρ e dt) n d T sh n dt = d T t D n + Je, n ρ m = Tetρ mi T i (magnetic charges) is discrete 3-form on space, (6) J m = P J mi P i (current) is discrete form on space. The compact form of Eqs.(6) can be written as df = J d T F = J, with discrete continuity equations or integrability conditions d J = 0 d T J = 0. Proposition 3. If the initial condition satisfies the first and third equations in Eqs.(6), the solution of the second and fourth equations in Eqs.(6) automatically satisfy Eqs.(6). Proof. Because the dimension of spacetime is 3 +, therefore d T s (ρ e dt) = 0 d T t J e = 0 d s ρ m = 0 d t J m dt = 0, and the continuity equations can be reduced to So we have d T t ( ρ edt n ) d T s J n e = 0 d sj n+ m dt+d t ρ n m = 0. d T t dt s Dn d T t (ρ e dt) n = d T t (ρ e dt) n d T s (dt s Hn dt J n e ) = 0 d t d s B n d t ρ n m = d t ρ n m +d s(d s E n+ dt+j n+ m = 0. 7 dt)

8 Nowwe show thescheme (5) ontheproduct ofddiscrete spacemanifold and time. The second and fourth equations in Eqs.(5) based on Fig. are D n+ D n Bn+ B n +J n e = Hn Hn e = En+ e +E n+ e +E n+ 3 e 3. P where denotes the measure of forms and dual. The summation on the right is orient, that is to say, inverse the orientation of e i, then multiply with Ēi. Eqs.(7) can be implemented on D discrete manifold directly(see Fig.). Eq.(7) on rectangular gird is just the Yee scheme. (7) Figure : edge and face with direction In the absence of magnetic or dielectric materials, the relations are simple: D i = ε 0 E i B i = µ 0 H i, (8) where ε 0 and µ 0 are two universal constants, called the permittivity of free space and permeability of free space, respectively. With relations(8), Eqs.(7) can be rewritten into an explicit iterative scheme for TE wave. E n+ E n ǫ 0 µ 0 H n+ H n +J n e = Hn Hn e = En+ e +E n+ e +E n+ 3 e 3 P TE (9) 8

9 The symmetric TE wave scheme induced from Eqs.(6) can be written as follows. E n+ E n ǫ 0 +Je n = Hn Hn e µ 0 H n+ H n m = En+ e +E n+ e +E n+ 3 e 3 P +J n+ Schemes for TM wave If writing F n+ = H n+ dt D n F n = E n dt B n J = ( ρ e,j e dt) J = ( ρ m dt,j m ), TE (0) where H = H i e i is the discrete form on space, D = H i P i is the E P discrete form on space, E = D i P i is the dual of D on space, B = P B i e i is the dual of H on space, ρ e = ρ ei T i is the discrete 3 form E Tet on space, J e = J ei P i is the discrete form on space, ρ m dt is the discrete P form on time, J m = J mi E i is the discrete form on space, the discrete E Maxwell s equations can be rewritten as d s D n = ρ n e d s H n+ dt = d t D n +J n+ e dt d T s Bn = (ρ m dt) n d T s En dt = d T t Bn Jm n. () Proposition 3. If the initial condition satisfies the first and third equations in Eqs.(), the solution of the second and fourth equations in Eqs.() automatically satisfy Eqs.(). Proof.Because the dimension of spacetime is 3+ or +, therefore d T s (ρ m dt) = 0 d T t J m = 0 d s ρ e = 0 d t J e dt = 0, 9

10 and the continuity equations can be reduced to So we have d T t (ρ m dt) n d T s ( J n m ) = 0 d sj n+ e dt d t ρ n e = 0. d t d s D n d t ρ n e = d t ρ n e d s(d s H n+ dt J n+ e dt) = 0 d T t d T sb n d T t (ρ m dt) n = d T t (ρ m dt) n +d T s(d T se n+ dt+ Jm) n = 0. Now we show the scheme () on the product of D discrete space manifold and time. The second and fourth equations in Eqs.() based on Fig. are B n+ B n D n+ D n +J n m = En En e +J n+ e = Hn+ e +H n+ e +H n+ 3 e 3. P () With relations (8), Eqs.() can be rewritten into an explicit iterative scheme for TM wave. E n+ E n ǫ 0 +J n+ e = Hn+ e +H n+ e +H n+ 3 e 3 P H n+ H n µ 0 General schemes +J n m = En En e TM (3) For real world materials, the constitutive relations are not simple proportionalities, except approximately. The relations can usually still be written: D = εe B = µh, but ε and µ are not, in general, simple constants, but rather functions. With Ohm s law E = σ J, J m = σ m H, 0

11 where σ is the electrical conductivity and σ m is magnetic conductivity. The DEC schemes can be written as ǫ En+ E n +σ En+ +E n = Hn Hn e µ Hn+ H n ǫ En+ E n µ Hn+ H n +σ m H n+ +H n +σ En+ +E n H n+ +H n +σ m = En+ e +E n+ e +E n+ 3 e 3, P = Hn+ e +H n+ e +H n+ 3 e 3 P = En En e 4 Stability, convergence and accuracy Stability. TE TM The Courant-Friedrichs-Lewy condition is a necessary condition for convergence while solving certain partial differential equations numerically. Now, we find this condition for scheme (0). Condition for scheme (3) can be induced in the same way. First, we decompose DEC algorithm into temporal and spacial eigenvalue problems. The temporal eigenvalue problem: H n 0 t = ΛH n 0 It can approximated by difference equation Supposing H n+ 0 H n 0 +Hn 0 () = ΛH n 0. (4) H n+ 0 = H n 0 cos(n ) H n 0 = H n 0 cos(n ) and H n+ 0 = H n 0 sin(n ) H n 0 = H n 0 sin(n ),

12 and substituting those into Eq.(4), we obtain therefore cos(n )+cos(n ) () = Λ, sin(n )+sin(n ) () = Λ, 4 () Λ 0. This is the stabile condition for the temporal eigenvalue problem. The spacial eigenvalue problem: c H = ΛH It can be approximated by difference equation (5) based on Fig.. P 3 c ΛH 0 = l 3 l A0 (H A H 0 )+ l 34 l B0 (H B H 0 )+ l 45 l C0 (H C H 0 ) (5) Let Figure : Face and dual face H i = H 0 cos(cl 0i ) or H i = H 0 sin(cl 0i ), and substitute into Eq.(5) to obtain P 3 c Λ = l 3 l A0 (cos(cl 0A ) )+ l 34 l B0 (cos(cl 0B ) )+ l 45 l C0 (cos(cl 0C ) ) P 3 c Λ = l 3 l A0 (sin(cl 0A ) )+ l 34 l B0 (sin(cl 0B ) )+ l 45 l C0 (sin(cl 0C ) ).

13 So we have ( c l3 + l 34 + l ) 45 Λ 0. P 3 l A0 l B0 l C0 In order to keep the stability of scheme (3), we need ( () c l3 + l 34 + l ) 45, (6) P 3 l A0 l B0 l C0 or Convergence P Mim P3 P ( 3 c l3 + l 34 + l ) 45. l A0 l B0 l C0 By the definition of truncation error, the exact solution of Maxwell s equations satisfy the same relation as DEC scheme except for an additional term O(() + e ). This expresses the consistency, and so convergence for DEC scheme by Lax equivalence theorem (consistency + stability = convergence). Accuracy The derivative of Maxwell s equations is approximated by first order difference in schemes (0) and (3). Equivalently, H and E are approximated by linear interpolation functions. Consulting the definition about accuracy of finite volume method, we can also say that schemes (0) and (3) have first order spacial and temporal accuracy, and have second order spacial and temporal accuracy on rectangular grid with equivalent space and time steps. 3

14 5 Implementation The DEC algorithm of Maxwell s equations was implemented in C++ platform. The Fig.3 shows the flowchart of DEC schemes for Maxwell s equations. Figure 3 In the common practice, not every simulation step needs to be visualized, especially when the time step size is too small. Fig.4 exhibit Gaussian pluses 4

15 waveforms simulated by DEC. Figure 4: Simulation of Gaussian pluse on Stanford bunny by DEC Fig.5 exhibits two sources Gaussian pluses interference simulated by DEC. Our algorithm can simulate more complex situation on surface and 3D-space manifold. Figure 5: Simulation of Gaussian pluse interference on sphere by DEC References [] K.S. Yee, Numerical solution of inital boundary value problems involving Maxwell s equations in isotropic media. IEEE Trans. Ant. Prop. 4(3), (966). [] A. Bondeson, T. Rylander, P. Ingelstrom, Computational electromagnetics, Texts in Applied Mathematics, vol. 5. Springer, New York (005). [3] M. Clemens, T. Weiland, Magnetic field simulation using conformal FIT formulations. IEEE Trans. Magn. 38(), (00). 5

16 [4] P.W. Gross, P.R. Kotiuga, Electromagnetic theory and computation: a topological approach, Mathematical Sciences Research Institute Publications, vol. 48. Cambridge University Press, Cambridge, (004). [5] E. Hairer, C. Lubich, G. Wanner, Geometric numerical integration, Springer Series in Computational Mathematics, vol. 3. Springer-Verlag, Berlin (00). [6] A. Nicolet, F. Zolla, Y.O. Agha and S. Guenneau, Geometrical transformations and equivalent materials in computational electromagnetism, International Journal for Computation and Mathematics in Electrical and Electronic Engineering COMPEL, vol. 7, , (008). [7] A. Bossavit, L. Kettunen, : Yee-like schemes on a tetrahedral mesh, with diagonal lumping. Int. J. Numer. Modell. (-), 9 4 (999). [8] A. Bossavit, L. Kettunen, Yee-like schemes on staggered cellular grids: A synthesis between FIT and FEM approaches. IEEE Trans. Magn. 36(4), (000). [9] A. Stern, Computational Electromagnetism with Variational Integrators and Discrete Differential Forms. arxiv: [0] H. Whitney, Geometric integration theory. Princeton University Press, Princeton, (957). [] D.N. Arnold, R.S. Falk, R. Winther, Finite element exterior calculus, homological techniques, and applications. Acta Numer. 5, -55, (006). [] S. P. Novikov, Discrete connection on the triangulated manifold and difference linear equations arxiv: math-ph/ [3] B. Auchmann, S. Kurz, A geometrically defined discrete hodge operator on simplicial cells. IEEE Trans. Magn. 4(4), , (006). [4] A. Dimakis, F. Muller-Hoissen, Discrete Riemannian geometry. J Math Phys, 999, 40(3): [5] M. Desbrun, A.N. Hirani, M. Leok, J. E. Marsden, Discrete exterior calculus arxiv: math.dg/

17 [6] M. Meyer, M. Desbrun, P. Schröder, A.H. Barr, Discrete differential geometry operators for triangulated -manifolds. In InternationalWorkshop on Visualization and Mathematics, VisMath, (00). [7] J. M. Hyman, M. Shashkov, Natural discretizations for the divergence, gradient, and curl on logically rectangular grids. Comput. Math. Appl., 33(4):8-04, (997). [8] R. Hiptmair, Discrete Hodge operators, Numer. Math., 90():65-89, (00). [9] D.K. Wise, p-form electromagnetism on discrete spacetimes. Classical Quantum Gravity 3(7), (006). [0] M. Leok, Foundations of computational geometric mechanics. Ph.D. thesis, California Institute of Technology (004). 7