I try to run an analysis with flow-driven motion. The object is to rotate a body by airflow. The setup looks like this: One solid volume is surrounded by a rotating region that is surrounded again by a large fluid volume (air). The rotating region is set up as free-spinning with an inertia value. Two BC defined at the surrounding fluid: flow speed and pressure.
Actually nothing special so far. But my problem is: After meshing the solver closes unexpectedly. I tried my best with the CFD-help, but could not find the problem.
Btw: If defining the rotating region with a "known rotational speed", the analysis runs fine.
Is there any special parameter or trick?
Anyone have ideas?
I work with CFD2013. It's the 6th release I gain experience with since CFDesign v9.
When running 0 iterations the meshing process runs w/o problems. After that the model appears in the results window for a second or two. The then error "solver closed unexpectedly" pops up. After clicking "OK" at the popup-message, the model switches back to setup mode (pre-processing).
That is the way it runs, no matter how may iterations set.
There is an issue but it should work OK if you run for 0 iterations first, then continue from t(0).
I just tested it here and it's OK, this was after replicating your issue if running a number of iterations from 0 (it worked with a known rpm also as you say).
Can I confirm that you definitely ran this from 0 for 0 iterations? There should be no error if you do.
Also, what fluid are you using, just air_fixed?
The model runs now from 0 for 0 iternations. But, tryin to solve any further iteration results in the described solver error.
I added a support file of my model. Maybe you could take a quick look.
Also, I notice that this is compressible. You need to change the inlet conditions. A velocity is unsuitable, try a mass flow rate (this will not change throughout the domain). Also, set a pressure and temp. Is your Total Temperature in the Solve box definitely correct? Take a look at Total T in the help guide.
Also, the domain will likely need to be larger.
I suggest you test with an incompressible analysis first