2 way FSI problem
 

HI
i have a problem with two way FSI in cfx that I attach the error

i can't see the model and domains in cfx-pre in cfx.


please help me.

is the problem because of the mesh ?

Have you done the webinar on the FSI interaction made by ANSYS past winter ?

for the timestep error:
we can see on CFX that you didn't create any domain. Does that error appear at the launching of CFX in that case just go create the time step and everything should be fine (for the CFX part).
NOTE: you can't make an FSI simulation with transient/stationary case (it have to be stationary/stationary or transient/transient)

for the "wbloadmesh" error:
can you explain where do you use this variable and how do you define it please ?

tnx Thomas MADELEINE

it solved. :)

but now i have an another problem .

i want to define the rotational movement of rigid body around the origin , in specify path

but i don't now , how can I define it.

i attach all set ups of cfx and transient structural

help me, please
http://s2.img7.ir/xvRTU.gif
http://s2.img7.ir/iAwsi.gif
http://s2.img7.ir/OoznE.gif
http://s2.img7.ir/M61Oa.gif
http://s2.img7.ir/Ay1AB.gif
http://s2.img7.ir/cbjUS.gif
http://s2.img7.ir/y5W5V.gif
http://s2.img7.ir/zGRxs.gif

Mesh motion can be specified using the specified displacement option in the boundary conditions settings. Are you sure this is a 2-way FSI problem? Also, with that kind of large motion you will probably need to remesh during the run.

What are you trying to do?

totally i want to remesh but i don't know how ?????!!

quite difficult but possible in ANSYS CFX. If I remember correctly it will be with ICEM and some code, you can found some topic about it on this forum (key word remeshing) :

http://www.cfd-online.com/Forums/cfx...-icem-cfd.html

or you can use Fluent to do that...

but I don't really see where you want to go with a moving cylinder in 2.5D with a forced movement. Do you really need a 2-way FSI ?

Please can you explain what you are trying to simulate here ?

i want to simulate the thing like that this pic.

a cylinder with 50(c) rotate in a box , i want to calculate the time that the cylinder temp. decrease to 40(c)

i solve it by transient temperature but i cant define the rotating of cylinder in a circular path in the box.
look like the attached pic

well for that I will begin with something much simpler.

Firstly
Just two questions about your geometry:
- does the square of fluid is a fixed geometry or not ? I want to create a cylinder around the rigid body's trajectory

- Isn't the height quite important here ? like if a bath a hot cylinder in a water tank I will expect that the temperature will be higher at the surface of the tank.

Secondly
without any FSI, you can just create a model on CFX with MFX (CFX can handle some thermic simulation on solid), or export the HTC (heat transfer coefficient) form CFX on any thermo-mechanical software.

hi

i want to define the rotation path like that is shown instead of linear movement that i was define .

i dont have any prob. in moving mesh and something like that , biggest prob. is in defining the circular path that this cylinder can rotate in this box.


box is filled with water
around walls are adiabatic.
the cylinder temp. is 50 C
the water temp. is 20 C.

for the displacement you can set two expression like that:
dX = R*cos(Omega*t+ThetaIni)
dY = R*sin(Omega*t+ThetaIni)

R = radius of the trajectory
Omega = velocity (in rpm)
ThetaIni = initial angle

then you can use it to define mesh motion of the walls.


But again, the results will be horrible because in your case the boundary layout is really important and the mesh will be totally coarsen by the displacement.

Making a cylinder domain to have a sliding mesh rather a moving mesh is the best solution to my point of view. somebody did something similar here:
http://www.cfd-online.com/Forums/cfx...ddle-pool.html

One thing the minimal temperature that your pipe can achieve will be 20°C, so it will decrease by 30°C.

rigid body rotating in a box
 

hi
i tried to define two coordinate system that the firs one is rotating and the second one is stationary , and specify the rotational speed for rotating one but it didnt move and i attached all processes and errors .

help.

its simple , i just want to rotate the cylinder in a box.


is there any one that can help me ???? pleeeeeeeeeeeeeeaaaaaase:(:(:(

http://uupload.ir/files/n0m_1.gif

http://uupload.ir/files/f6r_2.gif

http://uupload.ir/files/8u6s_3.gif

http://uupload.ir/files/2nqa_4.gif

http://uupload.ir/files/netr_5.gif

http://uupload.ir/files/0qmo_7.gif

http://uupload.ir/files/lr2r_10.gif

http://uupload.ir/files/00xe_11.gif

http://uupload.ir/files/yk9p_12.gif

It appears you run on CFX-5. I have never tried this release. I don't know the coordinate system thing here, I don't thing it is useful.

for the rest, i don't understand:
Do you want to do the remeshing on CFX ? it is quite hard, really time consuming and not necessary here.
the moving mesh will not work. It can not handle big displacement like this and even if it works the mesh will be irrelevant.

the expression of motion for your cylinder are good. Apply them to the rigid body.
the fluid domain is stationary.

I don't understand your interface. If you fix the temperature and the HTC then why do you do an fluid simulation (except to have the temperature elevation of your fluid).

before doing anything I recommend to only solve the meshing movement in order to know if everything is correct. go to expert parameter (right-click on solver -> add) and go to the last tab and deactivate fluid and energy

hello sir
i want to practice one way and two way fsi problem in ansys
can you suggest any book or link from where i can get practice problem
thank you

Maybe this is a good solution for you as well regarding mesh rotation of the cyl.
Video: https://www.youtube.com/watch?featur...&v=lBVvx3YHzNI

Blog entry: http://www.edr.no/blogg/ansys_blogge..._body_modeling

Definitely use smth- like icem to get a proper hex mesh. If you want to use Ansys meshing... split the circle at 45° and extend this split point to the corners of the surrounding rectangle. This should help the auto-hex-mesher to get the blocking done correctly resulting in a o-grid around the circle.

thank you sir
really useful information