|
[Sponsors] |
May 23, 2012, 09:30 |
Adjoint Methods in OpenFOAM
|
#1 |
Member
Join Date: Mar 2012
Location: Munich, Germany
Posts: 67
Rep Power: 14 |
Hello,
I`ve calculated some adjoint equations for optimization problems of a few equations. An easy example is the heat conduction. Now Iīm trying to implement them in OpenFOAM. For the heat conduction example my idea is to extend the solver laplacianFoam. Is this possile? Perhaps in an easy way? And how does it look like for other solvers? If you have a look at adjointShapeOptimizationFoam, is this an extension or modification of an existing solver? Do anybody have some experience on this field and have done anything similar? Are there any tutorials relating to this? Regards treima |
|
May 23, 2012, 10:30 |
|
#2 |
Member
Roland
Join Date: Mar 2009
Location: Netherlands
Posts: 92
Rep Power: 17 |
Hi treima,
AdjointShapeOptimizationFoam is based on the simpleFoam solver. It should be possible to build an adjoint heat conduction solver from an existing solver, for example laplacianFoam as you mentioned. Which solver is best suited will depend on the formulation of your adjoint equations and possibly also the intended application of your solver. Best regards, Sylvester |
|
May 23, 2012, 10:51 |
|
#3 |
Member
Join Date: Mar 2012
Location: Munich, Germany
Posts: 67
Rep Power: 14 |
Thank you!
One task for heat conduction is, for example: There is a domain, which boundary can be divided into three parts. One Dirichlet-part with given temperature, one Neumann-part, with given derivation of the temperature, and on the last part part, you are able to control the temperature. Itīs a Dirichlet-part, too. Now it should be possible to control the temperature, that you get a given temperature on the Neumann-part. I know, this is a very mathematical question and donīt always reflect the reality. But for test, if itīs possible to implement a adjoint solver in OpenFOAM, i think itīs just right. While Iīm new to programming OpenFOAM (except making some little changes in solvers), is there some good place where i can get some information? Especially about adjoint methods? |
|
May 23, 2012, 11:28 |
|
#4 |
Member
Roland
Join Date: Mar 2009
Location: Netherlands
Posts: 92
Rep Power: 17 |
The best place for assistance on programming OpenFOAM is this forum, in my opinion. With the help of Google you can find a lot of information here. The wiki might also be able to help you out.
The only information I could find on the implementation of the adjoint equations in OpenFOAM are the papers by Carsten Othmer. AdjointShapeOptimizationFoam.C gives the reference IIRC. Unfortunately I can't help you much more than this, as I have not delved into the subject very deeply. Adjoint methods are of interest to me though, so I try to follow (and participate in) discussions whenever I can. Best regards, Sylvester |
|
May 24, 2012, 06:57 |
|
#5 |
Senior Member
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,907
Rep Power: 33 |
For the record, I run a Summer School in Zagreb for this reason: to improve people's knowledge of CFD, OpenFOAM, numerics, programming, physical modelling etc. We have closed the application for this year, but there's going to be another one in 2013.
Hrv
__________________
Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk |
|
Tags |
adjoint, extension, implementation, modification, solver |
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Any questions about Runge-Kutta methods | Runge_Kutta | Main CFD Forum | 33 | September 9, 2019 17:32 |
New OpenFOAM Forum Structure | jola | OpenFOAM | 2 | October 19, 2011 07:55 |
Cross-compiling OpenFOAM 1.6 on Linux for Windows 32 and 64bits with Mingw-w64 | wyldckat | OpenFOAM Announcements from Other Sources | 7 | January 19, 2010 16:39 |
Summer School on Numerical Modelling and OpenFOAM | hjasak | OpenFOAM | 5 | October 12, 2008 14:14 |
comments on FDM, FEM, FVM, SM, SEM, DSEM, BEM | kenn | Main CFD Forum | 2 | July 18, 2004 19:28 |