Inverse heat transfer problem with Conjugate Gradient Method
Posted: 21 Jun 2013, 18:24
Hello dear fellas...
i'm back.... i'm quite busy and my research leave is going to its end so.. i won't be here quite often but don't hesitate to PM me or email me.
This is the last program i did (and i checked it). From a writing point of view it is quite shortly detailed. i apologize.
Here is a compressed folder with my files: i use Elmer for defining my physical problem and then i wrote a python program with my optimization procedure based on the computation of the gradient with an adjoint problem (for who knows).
The aim is to identify/estimate an unknown heat flux with regards to the time evolution from given temperature measurements.
This objective is achieved through a minimization procedure that consists to minimize the quadratic temperature difference between measured and calculated (from a direct problem / elmer). Than the minimization method is based on the conjugate gradient iterative regularization method (short details in my pdf). this method requires to solve to other problems: adjoint and sensitivity. I defined the direct problem with elmer gui while i made two copies of it for defining directly in sif files the adjoint and sensitivity problems. The python file centralize the call of these 3 problems and the minimization technique is programmed inside two. You just have to run it (with python 2.7 of course on your pc).
Have fun!
If you have any questions.. contact me.
Regards and long life to Elmer!!!!
i'm back.... i'm quite busy and my research leave is going to its end so.. i won't be here quite often but don't hesitate to PM me or email me.
This is the last program i did (and i checked it). From a writing point of view it is quite shortly detailed. i apologize.
Here is a compressed folder with my files: i use Elmer for defining my physical problem and then i wrote a python program with my optimization procedure based on the computation of the gradient with an adjoint problem (for who knows).
The aim is to identify/estimate an unknown heat flux with regards to the time evolution from given temperature measurements.
This objective is achieved through a minimization procedure that consists to minimize the quadratic temperature difference between measured and calculated (from a direct problem / elmer). Than the minimization method is based on the conjugate gradient iterative regularization method (short details in my pdf). this method requires to solve to other problems: adjoint and sensitivity. I defined the direct problem with elmer gui while i made two copies of it for defining directly in sif files the adjoint and sensitivity problems. The python file centralize the call of these 3 problems and the minimization technique is programmed inside two. You just have to run it (with python 2.7 of course on your pc).
Have fun!
If you have any questions.. contact me.
Regards and long life to Elmer!!!!