CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   FSI problem (http://www.cfd-online.com/Forums/cfx/67580-fsi-problem.html)

 smagmon August 19, 2009 12:56

FSI problem

Hi all,
I am around a FSI simulation problem.
I have a closed tube with stationary water (no flow) initially. In one side of this tube, I set a BC as WALL with a mesh displacement . The specified displacement is given as a step function through CEL. All other boundaries are set to Stationary WALL. If I set water to be compressible (also using CEL), after running the model in CFX, I can observe pressure wave propagation.
As I have a converged solution for the fluid simulation, I started trying a FSI simulation, changing one of the stationary walls to ANSYS MULTI-FIELD option in mesh motion, and creating a /prep7 file in order to have a flexible wall in the other side (as a circular plate). I used ANSYS MFX for settings between Mechanical and CFX.
When I run the simulation many notices like bellow are given by CFX side, and then it crashes giving the following error.

+--------------------------------------------------------------------+
| ****** Notice ****** |
| While evaluating Static Enthalpy, |
| Static Pressure |
| went outside of its lower limit. Its minimum value was |
| -4.3869E+05. The bounds error was handled by clipping. |
| If this situation persists, consider increasing the table range. |
+--------------------------------------------------------------------+
----------------------------------------------------------------------
| COUPLING/STAGGER ITERATION = 3 |
----------------------------------------------------------------------
| SOLVING : Mesh Displacement |
----------------------------------------------------------------------
| Equation | Rate | RMS Res | Max Res | Linear Solution |
+----------------------+------+---------+---------+------------------+
| X-Disp | 1.00 | 3.0E-02 | 1.0E+00 | 8.6E-03 OK|
| Y-Disp | 1.00 | 1.0E-03 | 2.2E-02 | 1.1E-02 OK|
| Z-Disp | 1.00 | 1.0E-03 | 2.2E-02 | 8.8 1.1E-02 OK|
+----------------------+------+---------+---------+------------------+
| X-Disp | 0.01 | 4.3E-04 | 3.7E-03 | 1.0E-01 OK|
| Y-Disp | 0.01 | 1.4E-05 | 1.7E-04 | 5.2E-02 OK|
| Z-Disp | 0.01 | 1.4E-05 | 1.6E-04 | 24.3 5.2E-02 OK|
+----------------------+------+---------+---------+------------------+
| X-Disp | 0.18 | 7.7E-05 | 9.9E-04 | 9.8E-02 OK|
| Y-Disp | 0.14 | 2.0E-06 | 4.6E-05 | 3.4E-02 OK|
| Z-Disp | 0.14 | 2.0E-06 | 4.6E-05 | 24.3 3.3E-02 OK|
+----------------------+------+---------+---------+------------------+
| X-Disp | 0.21 | 1.6E-05 | 2.7E-04 | 8.5E-02 OK|
| Y-Disp | 0.16 | 3.2E-07 | 7.7E-06 | 3.0E-02 OK|
| Z-Disp | 0.16 | 3.1E-07 | 7.8E-06 | 24.3 2.9E-02 OK|
+----------------------+------+---------+---------+------------------+
| X-Disp | 0.20 | 3.3E-06 | 5.1E-05 | 8.2E-02 OK|
| Y-Disp | 0.16 | 5.1E-08 | 1.3E-06 | 2.4E-02 OK|
| Z-Disp | 0.16 | 4.9E-08 | 1.2E-06 | 24.3 2.4E-02 OK|
+----------------------+------+---------+---------+------------------+

+--------------------------------------------------------------------+
| ERROR #002100011 has occurred in subroutine cVolSec. |
| Message: |
| A negative SECTOR volume has been detected. Execution will proceed |
| but this is a possible cause of robustness problems. |
| The location of the first negative volume is reported below. |
| Volume : -0.1307E-12 |
| Location : ( 0.16305E-01, -0.47543E-03, -0.27501E-02) |
| This warning may be made fatal by setting the expert parameter |
| 'negative volume option = 1'. |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #002100012 has occurred in subroutine cVolSec. |
| Message: |
| A negative ELEMENT volume has been detected. This is a fatal |
| error and execution will be terminated. The location of the first |
| negative volume is reported below. |
| Volume : -0.3001E-11 |
| Location : ( 0.16099E-01, -0.74617E-03, -0.25828E-02) |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| CFX encountered the error: |
| 0. A folded mesh has oc- |
| curred. This process will shut down as soon as possible. |
| |
| |
| |
+--------------------------------------------------------------------+

Is the "Notice" related to the error? I already tried to do what it advices, changing the range table, but it doesn't work.
When I change the fluid to air as ideal gas instead of water the the notice and error don't come.

Is my description clear?

Somebody knows what can be my problem?

Any tip or constructive commentary will be extremely appreciated, or even questions about my model, please let me know.

Thanks

 smagmon August 20, 2009 08:44

I forgot to say that I am using ANSYS Mechanical to generate my mesh.

Thanks!

Smagmon

 ghorrocks August 20, 2009 18:46

As the message says, the mesh is generating negative volume elements. This can be caused by an invalid initial mesh but most of the time it is caused by excessive mesh motion.

Try a smaller timestep. Also do a simulation where you output the mesh from each timestep and see if you can see where the mesh is folding.

 smagmon August 21, 2009 05:19

Hi Mr. Horrocks,
first of all, thanks for your answer. I already tried using smaller timestep, and even an adaptive timestep, but it does not work. I will try to do my simulation outputing the mesh...
But I could observe that even that the simulation works with air ideal gas, some strange behavior can be observed ( http://www.youtube.com/watch?v=zz0pN3vVQRM ). The meshes seem to not fit each other. I used a hex mapped mesh, gotten from extruded mesh area. I am thinking about the effects of the mesh motion model. Which one should I use? I used in my lest run the model option "value" and set it to 1[m^3 s^-1]/wall distance. Is it fit with the advice (1/wall distance, does not fit expected dimensions)? Or did I misunderstand something?

 ghorrocks August 21, 2009 09:05

What strange behaviour can you see? I can't see anything obviously weird on the video, but I don't know what I am looking at.

 smagmon August 21, 2009 10:49

Sorry.. I meant, in the video, the right down image shows the mesh shape of fsi region in cfx. When the stimulus stops, it shows something as a cross, in a no homogeneous displacement. It seems to be caused by mesh stiffness variation along the boundary. I hope I explained now.
Observing the crashed simulation (error file for water simulation), and observing the solution for air, I could observe some similarities in the mesh shape. The negative elements in the crashed solution seems to be in that region where the crossed displacement appears.

 dhxlxz August 22, 2009 07:31

Nenative Volumes

Because there is a negative volume, so CFX stopped running. You can first improve your grid quality, make sure that the grid orthognality is good. Then you decrease your time step; Also you can make your grid deformation technique in the setting items of domain.
Good luck!

 All times are GMT -4. The time now is 16:19.