Uncertainty Quantification and Data assimilation using OpenPALM
From Monday 22 May 2017 to Tuesday 23 May 2017
Training
Announced
Deadline for registration: 15 days before the starting date of each training
Duration : 2 days / (14 hours)
Pre-registration
Abstract
This is a two-day training course that presents some basic ideas on how implementing Uncertainty Quantification (UQ) and Data Assimilation (DA) methods with the dynamical coupling software OpenPALM (https://cerfacs.fr/globc/PALM_WEB/),
possibly combined with the Python libraryOpenTURNS (http://www.openturns.org/).
Python has the advantage of being a pre-compiled, portable, dynamical, object-oriented free language.
Target participants
This training session is for engineers, physicists, computer scientists
and numerical analysts who aim at developing or using modular uncertainty quantification and data assimilation applications based on numerical models.
Prerequisites
• Notions of data assimilation methods
• Notions of uncertainty quantification methods
• Notions of basic programming in Python, Fortran or other classical languages
• Training on OpenPALM (https://cerfacs.fr/en/event/code-coupling-using-openpalm-5/)
Scientific contact : Sophie RICCI Mélanie ROCHOUX
Fee
- Trainees/PhDs/PostDocs : 100 €
- CERFACS shareholders/CNRS/INRIA : 300 €
- Public : 600 €
Program
(Every day from 9h to 17h30)
Theoretical part
- General presentation of data assimilation
- Various applications at CERFACS
- Variational formulation of data assimilation
Practical part
- Implementation of a simplified ocean model in the OpenPALM coupler
- Formulation of the innovation vector
- Formulation of the different mathematical operator for the variational approach
- Formulation of the cost function and its gradient
- Validation and use of the adjoint model
- Validation and use of the cost function
- Implementation of the 3D-FGAT algorithm with OpenPALM
- Formulation and use of the 4D-Var and 3D-PSAS algorithms
