Cerfacs Enter the world of high performance ...

The 30 March 2020

High order discontinuous spectral methods for massively parallel LES

nasri |  

Postponed

Deadline for registration: 15 days before the starting date of each training
Duration : 1 day (7 hours)

Pre-registration

Abstract

Large Eddy Simulation (LES) needs accurate schemes with enhanced spectral properties (low dissipation and low dispersion). Even if standard high order schemes (WENO, compact…) respect this need for structured grids, the situation is less clear for complex geometries, where unstructured grids are required.
The discontinuous spectral methods were introduced in order to circumvent the drawbacks of standard methods on unstructured grids. In this case, any conservative quantity is defined as a local polynomial in a control volume through its definition in several degrees of freedom per cell.
During this training, we will introduce these approaches defines as spectral discontinuous. After a bibliographic review, attention is paid on applications performed with the in-house solver JAGUAR developed at CERFACS and based on a discontinuous spectral formalism.

Objective of the training

To learn about spectral discontinuous approaches and to know how to realize LES computations with them.

Learning outcomes

The participants should be able to :

  • explain the differences between the different numerical schemes,
  • chose the good number of iterations in order to reach the convergence,
  • use the good sub-grid model,
  • chose the good size of unstructured grids,
  • being able to simulate academicals configurations.

Target participants

This training session is for students, engineers, physicists and computer scientists who wish to reinforce or extend their theoretical background to the precise use and analysis of CFD simulations.

Prerequisites

In order to follow this course, you need to:

  • Knowledge of Unix commands.
  • Knowledge of numerical flow simulations.

To verify that the prerequisites are satisfied, the following questionnaires must be completed. You need to get at least 75% of correct answers in order to be authorized to follow this training session. If you don’t succeed it, your subscription will not be validated. You only have two chances to complete them.

Questionnaire 1 : https://forms.gle/VxkqRFfoqJSuoTBQ7

Questionnaire 2 : https://forms.gle/DL6QXxRDRNqKA5V1A

Scientific contact : Jean-François Boussuge

Fee

  • Trainees/PhDs/PostDocs : 70 €
  • CERFACS shareholders/CNRS/INRIA : 200 €
  • Public : 400 €

Program

Day 1 morning:
Introduction and analysis of discontinuous spectral methods
– Discontinuous Galerkin
– Spectral Volume
– Spectral Difference
– Flux Reconstruction

Day 1 afternoon:

Effects of spectral discontinuous methods on a computational environment:
– High order mesh and curved elements,
– Splitting the mesh
– Post-treatment

Introduction to JAGUAR:
– Mesh and input file
– Boundary conditions
– Serial and parallel performance of JAGUAR

– Application of JAGUAR on several test cases. It is also possible to work on an attendee’s application.

Final examination

A final exam will be conducted during the training.

Pre-registration

NEWS

Continuity of activity of the Cerfacs during the Covid-19 pandemic

superadmin |  20 March 2020

On Monday 16 March 2020, in the context of the rapidly evolving COVID-19 epidemic, Cerfacs decided to reorganize its activities by implementing a Business Continuity Plan (BCP) and deploying teleworking facilities for all its employees. All staff members thus continue to carry out their full mission.Read more


A fiery wakeup call for climate science

superadmin |  26 February 2020

The extent of the recent wildfires in Australia significantly exceeded the projections of any member of the multi-model CMIP archive.  This highlights how current multi-model ensembles may be under-representing the risks of natural disasters under climate change.  Limited coupled system process representation in most models coupled with a lack of parameter uncertainty exploration means that some risks are not explored by the existing international multi-model framework.  This calls for a reassessment of how to focus climate model development on providing robust risk quantification for those impacts which most directly affect society. Sanderson, B.M., Fisher, R.A. A fiery wake-up call for climate science. Nat. Clim. Chang. (2020) nature.com Media coverage BBC Sydney Morning Herald The Guardian Wired The Daily Express YahooRead more

ALL NEWS