Computational Technology Resources - CCP

Keywords: computational fluid dynamics, high performance computing, petaflop.

Summary

The goal of this paper is to expand on why we need to develop a petaflop computational fluid dynamics (CFD) capability to study energy and environment and to highlight some of the potential difficulties inherent in this ambitious development. With increasingly powerful computers being built, accurately computing complex multiphysics problems is now feasible. This was not possible all that long ago, because the problems usually occur at different spatial and temporal scales. Code_Saturne, from EDF R&D, has been developed [1] with this specific objective because EDF has many operationally critical reasons (energy efficiency and environmentally compatible production) to further improve the fidelity of their modelling and simulation capability in areas such as, thermal fatigue due to hot fluid impacting pipes in their nuclear power plants, release of plumes from their plants in the atmosphere, and the fluid-structure interaction between marine turbines and water. However, the situations which require Petaflop capabilities raise problems that have yet to be resolved. It is still not clear how to generate meshes with billions of cells and, having created the mesh, it is unclear how the current partitioning tools are going to perform when having to partition a mesh into hundreds of thousands of subdomains. Another key issue concerns the debugging of any CFD code if only a few cells among billions are responsible for the failure of the code. Moreover, postprocessing is another hot topic as probably most of it will have to be carried out by the CFD code to avoid transfers from the Petaflop machine to the postprocessing one. Code_Saturne, which is now optimised to benchmark future European Petaflop machines within the PRACE project [2], will be used to demonstrate some preliminary applications and the case of a controlled flow around a wing will be investigated to show how lift can be increased to reduce energy consumption.

References

1: F. Archambeau, N. Mechitoua, M. Sakiz, "Code_Saturne: A Finite Volume Code for the Computation of Turbulent Incompressible Flows: Industrial Applications", International Journal on Finite Volumes, 1(1), 2004.
2: http://www.prace-project.eu

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 £72 +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: 90 PROCEEDINGS OF THE FIRST INTERNATIONAL CONFERENCE ON PARALLEL, DISTRIBUTED AND GRID COMPUTING FOR ENGINEERING Edited by: Paper 2 Developing a Petaflop Computational Fluid Dynamics Capability for Energy and Environment C. Moulinec¹, A.G. Sunderland¹, D. Emerson¹, X. Gu¹, Y. Fournier² and J.C. Uribe³ ¹STFC Daresbury Laboratory, Warrington, United Kingdom ²EDF R&D, MFEE, Chatou, France ³School of Mechanical, Aerospace and Civil Engineering, University of Manchester, United Kingdom doi:10.4203/ccp.90.2 purchase the full-text of this paper Full Bibliographic Reference for this paper C. Moulinec, A.G. Sunderland, D. Emerson, X. Gu, Y. Fournier, J.C. Uribe, "Developing a Petaflop Computational Fluid Dynamics Capability for Energy and Environment", in , (Editors), "Proceedings of the First International Conference on Parallel, Distributed and Grid Computing for Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 2, 2009. doi:10.4203/ccp.90.2 Keywords: computational fluid dynamics, high performance computing, petaflop. Summary The goal of this paper is to expand on why we need to develop a petaflop computational fluid dynamics (CFD) capability to study energy and environment and to highlight some of the potential difficulties inherent in this ambitious development. With increasingly powerful computers being built, accurately computing complex multiphysics problems is now feasible. This was not possible all that long ago, because the problems usually occur at different spatial and temporal scales. Code_Saturne, from EDF R&D, has been developed [1] with this specific objective because EDF has many operationally critical reasons (energy efficiency and environmentally compatible production) to further improve the fidelity of their modelling and simulation capability in areas such as, thermal fatigue due to hot fluid impacting pipes in their nuclear power plants, release of plumes from their plants in the atmosphere, and the fluid-structure interaction between marine turbines and water. However, the situations which require Petaflop capabilities raise problems that have yet to be resolved. It is still not clear how to generate meshes with billions of cells and, having created the mesh, it is unclear how the current partitioning tools are going to perform when having to partition a mesh into hundreds of thousands of subdomains. Another key issue concerns the debugging of any CFD code if only a few cells among billions are responsible for the failure of the code. Moreover, postprocessing is another hot topic as probably most of it will have to be carried out by the CFD code to avoid transfers from the Petaflop machine to the postprocessing one. Code_Saturne, which is now optimised to benchmark future European Petaflop machines within the PRACE project [2], will be used to demonstrate some preliminary applications and the case of a controlled flow around a wing will be investigated to show how lift can be increased to reduce energy consumption. References 1 F. Archambeau, N. Mechitoua, M. Sakiz, "Code_Saturne: A Finite Volume Code for the Computation of Turbulent Incompressible Flows: Industrial Applications", International Journal on Finite Volumes, 1(1), 2004. 2 http://www.prace-project.eu 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 £72 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions