Parallel algorithm with modulus structure for simulation of seismic wave propagation in 3D multiscale multiphysics media. / Kostin, Victor; Lisitsa, Vadim; Reshetova, Galina et al.
Parallel Computing Technologies - 14th International Conference, PaCT 2017, Proceedings. ed. / Victor Malyshkin. Springer-Verlag GmbH and Co. KG, 2017. p. 42-57 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10421 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
}
TY - GEN
T1 - Parallel algorithm with modulus structure for simulation of seismic wave propagation in 3D multiscale multiphysics media
AU - Kostin, Victor
AU - Lisitsa, Vadim
AU - Reshetova, Galina
AU - Tcheverda, Vladimir
PY - 2017
Y1 - 2017
N2 - This paper presents a problem-oriented approach, designed for the numerical simulation of seismic wave propagation in models containing geological formations with complex properties such as anisotropy, attenuation, and small-scale heterogeneities. Each of the named property requires a special treatment that increases the computational complexity of an algorithm in comparison with ideally elastic isotropic media. At the same time, such formations are typically relatively small, filling about 25% of the model, thus the local use of computationally expensive approaches can speed-up the simulation essentially. In this paper we discuss both mathematical and numerical aspects of the hybrid algorithm paying most attention to its parallel implementation. At the same time essential efforts are spent to couple different equations and, hence, different finite-difference stencils to describe properly the different nature of seismic wave propagation in different areas. The main issue in the coupling is to suppress numerical artifacts down to the acceptable level, usually a few tenth of the percent.
AB - This paper presents a problem-oriented approach, designed for the numerical simulation of seismic wave propagation in models containing geological formations with complex properties such as anisotropy, attenuation, and small-scale heterogeneities. Each of the named property requires a special treatment that increases the computational complexity of an algorithm in comparison with ideally elastic isotropic media. At the same time, such formations are typically relatively small, filling about 25% of the model, thus the local use of computationally expensive approaches can speed-up the simulation essentially. In this paper we discuss both mathematical and numerical aspects of the hybrid algorithm paying most attention to its parallel implementation. At the same time essential efforts are spent to couple different equations and, hence, different finite-difference stencils to describe properly the different nature of seismic wave propagation in different areas. The main issue in the coupling is to suppress numerical artifacts down to the acceptable level, usually a few tenth of the percent.
KW - Coupling of finite-difference stencils
KW - Domain decomposition
KW - Finite-difference schemes
KW - Group of processor units
KW - Local grid refinement
KW - Master processor unit
KW - MPI
KW - LEBEDEV SCHEME
KW - MESH REFINEMENT
KW - FINITE-DIFFERENCE
UR - http://www.scopus.com/inward/record.url?scp=85028709586&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-62932-2_4
DO - 10.1007/978-3-319-62932-2_4
M3 - Conference contribution
AN - SCOPUS:85028709586
SN - 9783319629315
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 42
EP - 57
BT - Parallel Computing Technologies - 14th International Conference, PaCT 2017, Proceedings
A2 - Malyshkin, Victor
PB - Springer-Verlag GmbH and Co. KG
T2 - 14th International Conference on Parallel Computing Technologies, PaCT 2017
Y2 - 4 September 2017 through 8 September 2017
ER -
ID: 25775210