FSI continuity defect
 

Hello all,

Im trying to simulate a Fluid-Structure-Interaction with Ansys Workbench 15.0. The fluid part with CFX and the structural part with ANSYS (Transient Structural).

The model has 3 parts:

1. The fluid domain (water) a tube with d=10mm and a length of 100mm
2. The Top Membrane a Shell-Element (structural, steel) with d=10mm and a thickness of 0.25mm
- the first Fluid Solid interface
- with a defined displacement:
0 [s] = 0 [m]
2.e-005 [s] = 0 [m]
7.e-003 [s] = 1.e-003 [m]
1 [s] = 1.e-003 [m]

on the center (a circle with d=1mm) only in Z-direction, longitudinal to the tube.
3. The Bottom Membrane also a Shell-Element (structural, steel) with d=10mm and a thickness of 0.25mm
- the second Fluid Solid interface
- it will be deformed due to the fluid forces.

So the tube is closed by the 2 Shell-Elements. The 2 Shells are deforming.

Analysis type: Transient
Total time: 10ms
Time steps: 1e-06s

My Problem:
The solution shows a large continuity defect, like you can see in the chart and picture.

http://www.directupload.net/file/d/3...ft2yq4_png.htmhttp://www.directupload.net/file/d/3...ft2yq4_png.htmhttp://www.directupload.net/file/d/3886/hift2yq4_png.htm
Chart 1: traversal node displacement, blue top membrane, red bottom membrane

http://www.directupload.net/file/d/3...8m5r6t_png.htmhttp://www.directupload.net/file/d/3...8m5r6t_png.htmhttp://www.directupload.net/file/d/3886/xb8m5r6t_png.htm
Picture 1: Time: 10ms, Displacement scale x8

Im already try to decrease the Conservation target (10e-02 to 10e-04), Residual target (10e-04 to 10e-06) and the Expert Parameter: max Continuity loops from 1 to 5. This steps were showing no significant difference.

Have somebody experience with that kind of FSI simulations and have an idea to solve this problem?

If someone needs more information about the model, it would be a pleasure for me to describe it more in detail.

Thanks for the help!

Best regards,
Martin Wisniewski

For a closed deforming domain you need to use a compressible fluid to maintain mass conservation. So you need to define your water density to be a function of pressure, e.g:

Pref= 1 [atm]
BulkMod= 2.2e9[Pa]
Density = 998[kg/m^3]*(1 +(Absolute Pressure - Pref) / BulkMod)

Once you combine this with FSI it will be really unstable. Look on the ANSYS customer portal for Solution 2022119 to see how to stabilize.