HR: 0830h
AN: S41E-0137    [PDF]
TI: Parallel Simulation of Wave Propagation in Three-Dimensional Poroelastic Media
AU: * Sheen, D
EM: dsheen@indiana.edu
AF: School of Earth and School of Earth and Environmental Sciences, Seoul National University, San 56-1, Shillim-dong, Gwanak-gu, Seoul, 151-742 Korea, Republic of
AU: * Sheen, D
EM: dsheen@indiana.edu
AF: Laboratory for Computational Geodynamics, Indiana University, 800 E. Kirkwood Ave., Bloomington, IN 47405 United States
AU: Baag, C
EM: baagce@snu.ac.kr
AF: School of Earth and School of Earth and Environmental Sciences, Seoul National University, San 56-1, Shillim-dong, Gwanak-gu, Seoul, 151-742 Korea, Republic of
AU: Tuncay, K
EM: ktuncay@indiana.edu
AF: Laboratory for Computational Geodynamics, Indiana University, 800 E. Kirkwood Ave., Bloomington, IN 47405 United States
AU: Ortoleva, P J
EM: ortoleva@indiana.edu
AF: Laboratory for Computational Geodynamics, Indiana University, 800 E. Kirkwood Ave., Bloomington, IN 47405 United States
AB: Parallelized velocity-stress staggered-grid finite-difference method to simulate wave propagation in 3-D heterogeneous poroelastic media is presented. Biot­_s poroelasticity theory is used to study the behavior of wavefield in fluid saturated media. In the poroelasticity theory, the fluid velocities and pressure are included as additional field variables to those for the pure elasticity in order to describe the interaction between pore fluid and solid. Discretization of governing equations for finite-difference approximation is performed for total of 13 components of field variables in 3-D Cartesian coordinates: six components for velocity, six components for solid stress, and a component for fluid pressure. The scheme has fourth-order accuracy in space and second-order accuracy in time. Also, to simulate wave propagation in an unbounded medium, the perfectly matched layer (PML) method is used as an absorbing boundary condition. In contrast with the pure elastic problem, the larger number of components to describe the poroelasticity requires a huge sum of core memory inevitably. In the case of modeling in a realistic scale, the computation is hardly to run on serial platforms. Therefore, the computationally efficient scheme to run on a large parallel environment is required. The parallel implementation is achieved by using a spatial decomposition and the portable message passing interface (MPI) for communication between neighboring processors. Direct comparisons are made for serial and parallel computations. The inevitability and efficiency of parallelization for the poroelastic wave modeling are also demonstrated using model examples.
DE: 0689 Wave propagation (4275)
DE: 3210 Modeling
DE: 5114 Permeability and porosity
DE: 7260 Theory and modeling
SC: Seismology [S]
MN: 2003 Fall Meeting