TY - GEN

T1 - A parallel immersed boundary method for blood-like suspension flow simulations

AU - Pacull, F.

AU - Garbey, M.

PY - 2011

Y1 - 2011

N2 - This paper presents a numerically efficient implementation of the Immersed Boundary Method (IBM), originally developed by [7] to simulate fluid/elastic-structure interactions. The fluid is assumed to be incompressib0le with uniform density, viscosity, while the immersed boundaries have fixed topologies with a linear elastic behavior. Based on the finite-difference method, a major numerical advantage of the IBM is the high level of uniformity of mesh and stencil, avoiding the critical interpolation processes of the cut-cell/direct methods. The difficulty of accurately simulating interaction phenomena involving moving complex geometries can be overcome by implementing large and parallel IBMcomputations on fine grids, as described in [1]. While this paper is restricted to a two-dimensional low-Reynolds-number flow, most of the concepts introduced here should apply to three-dimensional bio-flows.We describe here the decomposition techniques applied to the IBM, in order to decrease the computational time, in the context of the parallel Matlab toolbox of [3]. Finally, we apply the method to a blood-like suspension flow test-case.

AB - This paper presents a numerically efficient implementation of the Immersed Boundary Method (IBM), originally developed by [7] to simulate fluid/elastic-structure interactions. The fluid is assumed to be incompressib0le with uniform density, viscosity, while the immersed boundaries have fixed topologies with a linear elastic behavior. Based on the finite-difference method, a major numerical advantage of the IBM is the high level of uniformity of mesh and stencil, avoiding the critical interpolation processes of the cut-cell/direct methods. The difficulty of accurately simulating interaction phenomena involving moving complex geometries can be overcome by implementing large and parallel IBMcomputations on fine grids, as described in [1]. While this paper is restricted to a two-dimensional low-Reynolds-number flow, most of the concepts introduced here should apply to three-dimensional bio-flows.We describe here the decomposition techniques applied to the IBM, in order to decrease the computational time, in the context of the parallel Matlab toolbox of [3]. Finally, we apply the method to a blood-like suspension flow test-case.

UR - http://www.scopus.com/inward/record.url?scp=78651557699&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78651557699&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-14438-7_16

DO - 10.1007/978-3-642-14438-7_16

M3 - Conference contribution

AN - SCOPUS:78651557699

SN - 9783642144370

T3 - Lecture Notes in Computational Science and Engineering

SP - 153

EP - 160

BT - Parallel Computational Fluid Dynamics 2008 - Parallel Numerical Methods, Software Development and Applications

T2 - 20th International Series of Meetings on Parallel Computational Fluid Dynamics, CFD 2008

Y2 - 19 May 2008 through 22 May 2008

ER -