Computational Technology Resources - CCP

Keywords: computational fluid dynamics, large eddy simulation, coupled, implicit, parallel, PETSc, computational.

Summary

In this research a computational fluid dynamics (CFD) code: MICSc (Momentum interpolation coupled solver using PETSc), for solving the filtered Navier-Stokes (N-S) equations with the large eddy simulation (LES) approach is presented. MICSc uses the finite volume method on co-located grids to discretize the equations and solves the resulting algebraic system in a coupled manner. A (Poisson-like) equation for pressure is derived from the continuity equation, by calculating the face fluxes from a compact formulation of the momentum interpolation technique (which improves the Rhie and Chow original formulation for unsteady flows and relaxation). The temporal terms are implicitly integrated and the efficiency of the coupled solver has been shown by solving the vortex shedding after a square cylinder [1].

The LES code is validated for open channel flows by solving fully turbulent flow in a plane channel (at Re_tau=224). Results are compared with reference solutions obtained by Moser et al. from a direct numerical simulation (DNS) [2]. This is a well-known validation case. The subgrid-scale (SGS) stresses, resulting from unresolved motions, are modeled using the Smagorinsky approach. The predicted mean velocity components, as well as the turbulent fluctuations, show good agreement with the DNS results from Moser's simulations.

MICSc is a parallel code based on the MPI message-passing standard, intended to run efficiently on large supercomputers. Its design follows a modern software engineering methodology, aiming at creating a modular, flexible and portable software. MICSc is based on the data structures provided by PETSc, the Portable, Extensible Toolkit for Scientific Computation, a toolkit for the parallel solution of PDE's developed at Argonne National Laboratory [3]. MICSc also relies on PETSc's linear system solvers, that are required to compute velocity and pressure fields at each time step of the simulation. We include a comparative study of the performance of different iterative linear solvers (GMRES and BiCGStab), as well as different preconditioners (Jacobi and Block Jacobi). We also analyze the behavior of the code in terms of parallel efficiency and scalability to large number of processors. The best preconditioner is Block Jacobi and the parallel efficiency of the whole simulation is reasonably good (more than 64% with 64 processors).

References

1: A. Cubero, N. Fueyo, "Preconditioning based on a partially implicit implementation of momentum interpolation for coupled solvers", Numerical Heat Transfer-B, 53(6), 510-535, 2008. doi:10.1080/10407790802035281
2: R.D. Moser, J. Kim, N. Mansour, "Direct Numerical Simulation of turbulent channel flow up to Re_tau=590", Physics of Fluids, 11(4), 943-945, 2007. doi:10.1063/1.869966
3: S. Balay, K. Buschelman, V. Eijkhout, W. Gropp, D. Kaushik, M. Knepley, L.C. McInnes, B. Smith, H. Zhang, "PETSc Users Manual", Technical Report ANL-95/11 - Revision 3.0.0, Argonne National Laboratory, 2008.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £125 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 94 PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON ENGINEERING COMPUTATIONAL TECHNOLOGY Edited by: Paper 12 MICSc: a PETSc-Based Parallel Code for Large Eddy Simulation A. Cubero¹, V. González², J.E. Román³, N. Fueyo¹ and G. Palau-Salvador² ¹Numerical Fluid Dynamics Group, University of Zaragoza, Spain ²Department of Rural Engineering, ³Department of Information Systems, Universidad Politécnica de Valencia, Spain doi:10.4203/ccp.94.12 purchase the full-text of this paper Full Bibliographic Reference for this paper , "MICSc: a PETSc-Based Parallel Code for Large Eddy Simulation", in , (Editors), "Proceedings of the Seventh International Conference on Engineering Computational Technology", Civil-Comp Press, Stirlingshire, UK, Paper 12, 2010. doi:10.4203/ccp.94.12 Keywords: computational fluid dynamics, large eddy simulation, coupled, implicit, parallel, PETSc, computational. Summary In this research a computational fluid dynamics (CFD) code: MICSc (Momentum interpolation coupled solver using PETSc), for solving the filtered Navier-Stokes (N-S) equations with the large eddy simulation (LES) approach is presented. MICSc uses the finite volume method on co-located grids to discretize the equations and solves the resulting algebraic system in a coupled manner. A (Poisson-like) equation for pressure is derived from the continuity equation, by calculating the face fluxes from a compact formulation of the momentum interpolation technique (which improves the Rhie and Chow original formulation for unsteady flows and relaxation). The temporal terms are implicitly integrated and the efficiency of the coupled solver has been shown by solving the vortex shedding after a square cylinder [1]. The LES code is validated for open channel flows by solving fully turbulent flow in a plane channel (at Re_tau=224). Results are compared with reference solutions obtained by Moser et al. from a direct numerical simulation (DNS) [2]. This is a well-known validation case. The subgrid-scale (SGS) stresses, resulting from unresolved motions, are modeled using the Smagorinsky approach. The predicted mean velocity components, as well as the turbulent fluctuations, show good agreement with the DNS results from Moser's simulations. MICSc is a parallel code based on the MPI message-passing standard, intended to run efficiently on large supercomputers. Its design follows a modern software engineering methodology, aiming at creating a modular, flexible and portable software. MICSc is based on the data structures provided by PETSc, the Portable, Extensible Toolkit for Scientific Computation, a toolkit for the parallel solution of PDE's developed at Argonne National Laboratory [3]. MICSc also relies on PETSc's linear system solvers, that are required to compute velocity and pressure fields at each time step of the simulation. We include a comparative study of the performance of different iterative linear solvers (GMRES and BiCGStab), as well as different preconditioners (Jacobi and Block Jacobi). We also analyze the behavior of the code in terms of parallel efficiency and scalability to large number of processors. The best preconditioner is Block Jacobi and the parallel efficiency of the whole simulation is reasonably good (more than 64% with 64 processors). References 1 A. Cubero, N. Fueyo, "Preconditioning based on a partially implicit implementation of momentum interpolation for coupled solvers", Numerical Heat Transfer-B, 53(6), 510-535, 2008. doi:10.1080/10407790802035281 2 R.D. Moser, J. Kim, N. Mansour, "Direct Numerical Simulation of turbulent channel flow up to Re_tau=590", Physics of Fluids, 11(4), 943-945, 2007. doi:10.1063/1.869966 3 S. Balay, K. Buschelman, V. Eijkhout, W. Gropp, D. Kaushik, M. Knepley, L.C. McInnes, B. Smith, H. Zhang, "PETSc Users Manual", Technical Report ANL-95/11 - Revision 3.0.0, Argonne National Laboratory, 2008. purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £125 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions