Development Of An Amr Code For The Advanced Simulation Of Incompressible Turbulent Flows With...

Universities and Institutes of France

France

July 22, 2022

Description

  • Organisation/Company: CEA
  • Research Field: Computer science › Modelling tools Engineering › Simulation engineering
  • Researcher Profile: Recognised Researcher (R2)
  • Application Deadline: 22/07/2022 21:00 - Europe/Athens
  • Location: France › Talence
  • Type Of Contract: Temporary
  • Job Status: Full-time
  • Hours Per Week: 35
  • Offer Starting Date: 01/10/2022
  • The objective of this project is the realization of advanced simulations of incompressible turbulent flows with the consideration of complex geometries in an immersed boundary code with mesh refinement techniques, in an HPC framework. It is part of a collaboration between the CEA and the Institute of Mechanics and Engineering of Bordeaux. I2M develops the Notus massively parallel computational code (https: // notus-cfd.org) for the study of incompressible single or multiphase flows, on Cartesian meshes with complex geometries taken into account by Immersed Boundary Methods (IBM) 1-5. It was recently used in the context of a major national HPC challenge on the Irene-rome supercomputer of GENCI/TGCC (up to 131,072 processors) 6.

    The CEA is a key player in the field of supercomputing and is preparing the next exaflop-scale supercomputers. We are interested here in the simulation of incompressible turbulent flows in complex geometries. These flows require the consideration of a large variety of spatio-temporal scales (boundary layer meshing in high dimensional geometries for flows with high Reynolds number). This project aims to remove scientific and technological barriers by proposing a tool based on adaptive mesh refinement (AMR) techniques taking advantage of the computing power of the latest generation of supercomputers.

    The physical modeling will be based on the incompressible Navier-Stokes equations. Turbulence will be modeled by LES and RANS (k-ω SST) approaches. A hybrid DES approach could be considered. The need to solve the finest near- wall scales requires adaptive meshes coupled with immersed boundary methods and CAD consideration of the geometry. It should be noticed that the proposed IBM/AMR coupled approach avoids the costly step of creating an unstructured mesh. The AMReX library 7 developed at LBNL, NREL and ANL offers a complete environment oriented towards exascale computing and hybrid architectures (CPU/GPU). It will be coupled with the Notus code whose schemes will be adapted to the new data and mesh structure.

    The duration of the post-doct position is 24 months. Initially, a bibliographic study will focus on numerical schemes, solvers and the most efficient libraries to achieve the objective of performing advanced simulations on geometries of interest in an exascale context. The next step will be to couple the selected tools and approaches in a new Notus-AMR code. The steps of implementation, verification, validation of models and methods will be pursued on the latest supercomputers, before carrying out advanced studies and simulations on complex geometries.

    Bibliography

    1 J. Picot, S. Glockner, Discretization stencil reduction of direct forcing immersed boundary methods on rectangular cells: the Ghost Node Shifting Method, Journal of Computational Physics, 364, pp18-48, 2018

    2 A. M. D. Jost and S. Glockner, Direct forcing immersed boundary methods: Improvements to the Ghost Node Method, Journal of Computational Physics, volume 438, 110371, 2021

    3 A. Lemoine, S. Glockner, J. Breil, Moment-of-Fluid Analytic Reconstruction on 2D Cartesian Grids, Journal of Computational Physics, 328, 131-139, 2017

    4 A. Lemoine. Analytic Gradient for the Moment-of-Fluid Method in Axisymmetric and on General Polyhedrons in Any Dimension. Journal of Computational Physics, Volume 422, 109741, 2020

    5 T. Milcent, A. Lemoine, Moment-of-Fluid Analytic Reconstruction on 3D Rectangular Hexahedrons, Journal of Computational Physics, Volume 409, 109346, 2020

    6 S. Glockner, A.M.D. Jost, A. Erriguible, Advanced petascale simulations of the scaling up of mixing limited flow processes for materials synthesis, Chemical Engineering Journal, volume 431, Part 1, 1 March 2022, 133647.

    7 AMReX, A software framework for massively parallel, block-structured adaptive mesh refinement (AMR) applications, https: // amrex- codes.github.io/amrex/

    Offer Requirements Skills/Qualifications

    The candidate must have a solid background in numerical methods and associated development on parallel architectures. The candidate will also have to be organized, take initiatives, and have a strong taste for code development and simulation of turbulent flows. Excellent skills of different languages is required, such as FORTRAN (2003), MPI, OpenMP, OpenACC, etc.

    Contact Information
  • Organisation/Company: CEA
  • Organisation Type: Private with public mission
  • Website: https: // notus-cfd.org
  • E-Mail: glockner@bordeaux-inp.fr
  • Country: France
  • City: Talence
  • Postal Code: 33400
  • Similar Jobs

    Inria

    France Jun 17, 2022

    Add to favorites Read more...

    Post-Doctoral Research Visit F/M Numerical methods (MPM and DEM) for gravity flows in mountains

    2022-05100 - Post-Doctoral Research Visit F/M Numerical methods (MPM and About the research centre or Inria department The Inria Grenoble - Rhône-Alpes research center groups together almost 600 other research and higher education institutions (University...

    Inria

    France Jun 8, 2022

    Add to favorites Read more...

    Post-Doctoral Research Visit F/M Kinetic approximation of diffusive and dispersive conservation laws

    2022-05023 - Post-Doctoral Research Visit F/M Kinetic approximation of diffusive and dispersive conservation laws Fonction : Post-Doctoral Research Visit The kinetic approach is a general method for approximating systems of physical wave speeds, the kinetic...

    Inria

    France Jun 3, 2022

    Add to favorites Read more...

    PhD Position F/M Koopman operator modelling of non-linear dynamical systems for ensemble methods.

    2022-05026 - PhD Position F/M Koopman operator modelling of non-linear About the research centre or Inria department laboratories of excellence, technological research institute, etc. This PhD will be supervised by Gilles Tissot and Étienne Mémin,...

    Inria

    France Jun 23, 2022

    Add to favorites Read more...

    Post-Doctoral Research Visit F/M Surface reconstruction and approximation with low-poly 3D models

    2022-05146 - Post-Doctoral Research Visit F/M Surface reconstruction and About the research centre or Inria department The Inria Université Côte d'Azur center counts 36 research teams as well as 7 The center's staff (about 500...