Commun. Comput. Phys., 2 (2007), pp. 806-826.


Comparison of Finite Difference and Mixed Finite Element Methods for Perfectly Matched Layer Models

V. A. Bokil 1*, M. W. Buksas 2

1 Department of Mathematics, Oregon State University, Corvallis, OR 97331, USA.
2 CCS-4, Mail Stop D409, Los Alamos National Laboratory, NM 87544, USA.

Received 29 April 2006; Accepted (in revised version) 5 January 2007
Available online 15 February 2007

Abstract

We consider the anisotropic uniaxial formulation of the perfectly matched layer (UPML) model for Maxwell's equations in the time domain. We present and analyze a mixed finite element method for the discretization of the UPML in the time domain to simulate wave propagation on unbounded domains in two dimensions. On rectangles the spatial discretization uses bilinear finite elements for the electric field and the lowest order Raviart-Thomas divergence conforming elements for the magnetic field. We use a centered finite difference method for the time discretization. We compare the finite element technique presented to the finite difference time domain method (FDTD) via a numerical reflection coefficient analysis. We derive the numerical reflection coefficient for the case of a semi-infinite PML layer to show consistency between the numerical and continuous models, and in the case of a finite PML to study the effects of terminating the absorbing layer. Finally, we demonstrate the effectiveness of the mixed finite element scheme for the UPML by a numerical example and provide comparisons with the split field PML discretized by the FDTD method. In conclusion, we observe that the mixed finite element scheme for the UPML model has absorbing properties that are comparable to the FDTD method.


Notice: Undefined variable: ams in /var/www/html/issue/abstract/readabs.php on line 163
PACS: 02.70Bf, 02.70.Dh, 41.20.Jb
Key words: Perfectly matched layers, mixed finite element methods, FDTD, Maxwell's equations.

*Corresponding author.
Email: bokilv@math.oregonstate.edu (V. A. Bokil), mwbuksas@lanl.gov (M. W. Buksas)
 

The Global Science Journal