[matzeee]

Hello guys, i have a fsae car with an radiator modelled as porous region. so i have the fluid region and the radiator as an porous region connected with interfaces for the radiator inlets and outlets. physik continium is "segregated flow", "k-omega modell", "turbulent" ofc and "isothermal".
i tried quite a lot to make the simulation run, but after 20 to 200 iterations i get the message "a floating point occurs..", or to say it easy there is divergence. i write you the different cases i tried to solve this problem:

first case:
i meshed everything with trimmed cells. first i tried one physik continium with the inital condition of -15m/s but the programm cant solve it. in my output window i get the message of "pressure corrections in my fluid" or "reversed flow in the outlet". no idea what i need to do or to help to solve the problem.
the same modell without the radiator(so only my fluid region) runs, so i guess it has smth to do with my radiator as a second region(porous region).

very showy is that in front of the driver, the velocity of the fluid goes up very extrem. i give you a picutre of that and the output window.
http://www.imagebanana.com/view/mdrbw6r1/asdff.JPG
http://www.imagebanana.com/view/qzpdcjn5/awera.JPG

second case:
i used polyhedral mesh and 2 physic continiums. the continium for the fluid has an velocity initial conditions with -15 m/s. the porous regoin of my radiator got the second physical continium and an initial condition of -2m/s. For my surprise the case run and "convergence" was reached.

Now i got a new car and i am using again a polyhedral mesh, and again nothing is running, so i think the example in my second case was a lucky shot.
in picture one its after i guess 30 iterations, velocity magnitute:
http://www.imagebanana.com/view/s9cuggjf/asdf.JPG
second picture is the same car one iteration further and the mesh displayed:
http://www.imagebanana.com/view/3n1bd3gg/asfd.JPG



If you need more specify informations pls tell me i will give you any information you need, till i am very frustraded now
Maybe generall information about the integration of a porous radiator part would be helpful too!

[fluid23]

It would be helpful if you shared images of your mesh.

[matzeee]

my generally way in meshing:
import geometry, splitting my car and the radiator(radiator is a box in our simulation) into two geometrys.
building a new box for the windtunnel in star ccm.
Wrapping my car (only car without radiator) geometry to eliminiate holes e.g..
Subtract operation between my Wrapped Car, Radiator and Windtunnel to get the Fluid Domain (Target part is windtunnel).
Creating two more boxes for BOI (bodys of influence).
Next i select my Fluid Domain part(which came from the subtract operation) and the radiator part and assign parts to regions(one boundary per surface and one region per part because i want to use the radiator region as a porous region).
Then i am using the Automated mesh operation, meshing my subtracted part (car, radiator,wt) which is my fluid domain and the radiator (using per part meshing). Using some custom controlls aswell, for example Bodys of Influence or the radiator.
At least the radiator inlet and outlet of the regions fluid,radiator are connected with an interface.


Here are some pictures of the finished mesh:
http://www.imagebanana.com/view/bink4c4o/asd1.JPG
http://www.imagebanana.com/view/0bw6i4z0/zzz1.JPG
http://www.imagebanana.com/view/sg708ysh/zz.JPG
http://www.imagebanana.com/view/wxmti8fk/z.JPG
http://www.imagebanana.com/view/hftpvqa3/zz2.JPG
http://www.imagebanana.com/view/kwelbb64/zzzzz.JPG




If more informations required pls contact, even if you have some thesis, we will check it to make progress!

[raskolnikov]

related to meshing of the cooler: the mesh size of interfaces couldnt have any transitions on their surfaces. for example abs. surf size (or min or max) = 1 mm in both of regions. you can then check the master and slave areas for your interface to find out the interface usability.