A simple and accurate discontinuous Galerkin scheme for modeling scalar-wave propagation in media with curved interfaces
Name
Zhang-2015-A simple and accurat.pdf
Size
580.82 KB
Format
Adobe PDF
Checksum (MD5)
98714b644df34063e9fd068ea2ca818a
Author(s)
Zhang, Xiangxiong
Tan, Sirui
Date Issued
January 2015
Journal
Geophysics
Publisher
Society of Exploration Geophysicists
Citation
Zhang, Xiangxiong, and Sirui Tan. “A Simple and Accurate Discontinuous Galerkin Scheme for Modeling Scalar-Wave Propagation in Media with Curved Interfaces.” GEOPHYSICS 80, no. 2 (January 14, 2015): T83–T89. © 2015 Society of Exploration Geophysicists
Version
Final published version
Abstract
Conventional high-order discontinuous Galerkin (DG) schemes suffer from interface errors caused by the misalignment between straight-sided elements and curved material interfaces. We have developed a novel DG scheme to reduce those errors. Our new scheme uses the correct normal vectors to the curved interfaces, whereas the conventional scheme uses the normal vectors to the element edge. We modify the numerical fluxes to account for the curved interface. Our numerical modeling examples demonstrate that our new discontinuous Galerkin scheme gives errors with much smaller magnitudes compared with the conventional scheme, although both schemes have second-order convergence. Moreover, our method significantly suppresses the spurious diffractions seen in the results obtained using the conventional scheme. The computational cost of our scheme is similar to that of the conventional scheme. The new DG scheme we developed is, thus, particularly useful for large-scale scalar-wave modeling involving complex subsurface structures.
MIT Department
Massachusetts Institute of Technology. Department of Mathematics
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1190/geo2014-0164.1
