FSI Turek Benchmark- defoming wall will not deform :(

  • 185 Views
  • Last Post 26 July 2018
Florian Henscheid posted this 16 July 2018

Hey,

for my bachelor thesis i am working with ansys Workbench to simulate the Turek Hron Benchmark. I watched several YouTube tutorials to make myself familiar with ansys and system coupling problems. anyhow the wall wich has to be a deforming wall will not deform at any time.

On the following pictures u can see the mesh, the boundary conditions, the fluid solid interface and the created named selction.

Many thanks for your help

Order By: Standard | Newest | Votes
Kremella posted this 16 July 2018

Hello Florian,

Could you please provide some more details on the problem you are attempting to model (geometry, physics, BCs etc.)? Are you using one-way or two-way coupling? Can you elaborate on the usage of dynamic meshing in your model?

Thank you.

Best Regards,

Karthik 

 

Raef.Kobeissi posted this 16 July 2018

Hi,

Have a look at this tutorial and let us know if you missed any step.

Cheers

Raef Kobeissi

Florian Henscheid posted this 17 July 2018

Hi Raef,

I used this video as my template and I think that I did everything like him.

Florian H.

Florian Henscheid posted this 18 July 2018

Hey Karthik,

the coboid is 2.2 meters long with a height of 0.41 meters and a width of 0.1 meters. The cylinder has a diameter of 0.1 meter and at it is the flag with a length of 0.35 meter and a heigth of 0.02 meter. The maximum inlet velocity is 2.25 m/s with the viscous model "laminar". The flag is supposed to be the moving wall and is calculated like a cantilever in the transient structural with a max acceleration (in + y-direction) of 9.8066 m/s^2 instead of a force . For the solid i am using the standard Material Polyethylen like Raef. I make use of the 2 way coupled FSI. Like Raef I only choose the fluid solid interface for a smoothing mesh.

kind regards

Florian Henscheid

 

Ryan O'Connor posted this 18 July 2018

Hi Florian,

I recommend to take a look at the system coupling log file (sclog.scl) and the Mecahnical log file (solve.out). In the scl file you can see whether the mapping is good. 

 

 +--------------------------------------------------------------------+

 |                        MAPPING SUMMARY                             |

 +--------------------------------------------------------------------+

 | Data Transfer                    |                                 |

 |   Diagnostic                     |   Source Side     Target Side   |

 +----------------------------------+---------------------------------+

 | Data Transfer                    |                                 |

 |   Percent Nodes Mapped           |       100             100       |

 |   Percent Area Mapped            |       100             100       |

 | Data Transfer 2                  |                                 |

 |   Percent Nodes Mapped           |       N/A             100       |

 +--------------------------------------------------------------------+

 

In the out file, you can review the forces being sent to Mechanical and the displacements being sent to Fluent.

 

 

 *** FORCE SUM ACROSS TARGET INTERFACE . . . . .Fluid Solid Interface (FSIN_1)

     RECEIVING FORCE FX SUM =   0.0000    

     RECEIVING FORCE FY SUM =   0.0000    

     RECEIVING FORCE FZ SUM =   0.0000    

 

...

 

 *** AVERAGE DISP INCREMENT ACROSS SOURCE INTERFACE. Fluid Solid Interface (FSIN_1)

     AVERAGE SENDING DISPLACEMENT INCREMENT DX =  0.31655E-17

     AVERAGE SENDING DISPLACEMENT INCREMENT DY =  0.63077E-18

     AVERAGE SENDING DISPLACEMENT INCREMENT DZ =   0.0000    

 

 

Regards,

Ryan

Florian Henscheid posted this 19 July 2018

Hi Ryan, 

Thanks for ur Answer.

My mapping summary looks exactly like the one given by you

 +--------------------------------------------------------------------+
 |                        MAPPING SUMMARY                                 |
 +--------------------------------------------------------------------+
 | Data Transfer                    |                                            |
 |   Diagnostic                      |   Source Side     Target Side   |
 +----------------------------------+---------------------------------+
 | Daten�bertragungen               |                                      |
 |   Percent Nodes Mapped           |       N/A             100       |
 | Daten�bertragungen 2            |                                      |
 |   Percent Nodes Mapped           |       100             100       |
 |   Percent Area Mapped             |       100             100       |
 +--------------------------------------------------------------------

Do the Sum Forces has to sum up to zero, or is that just in ur example? 

 

*** FORCE SUM ACROSS TARGET INTERFACE . . . . .Fluid - Solid Schnittstelle (FSIN_1)

     RECEIVING FORCE FX SUM =  0.14915E-02

     RECEIVING FORCE FY SUM = -0.29228E-01

     RECEIVING FORCE FZ SUM = -0.13391E-04

 

*** AVERAGE DISP INCREMENT ACROSS SOURCE INTERFACE. Fluid - Solid Schnittstelle (FSIN_1)

     AVERAGE SENDING DISPLACEMENT INCREMENT DX = -0.91796E-09

     AVERAGE SENDING DISPLACEMENT INCREMENT DY = -0.18160E-05

     AVERAGE SENDING DISPLACEMENT INCREMENT DZ = -0.19323E-14

 

Regards,

Florian 

 

Ryan O'Connor posted this 20 July 2018

Hi Florian,

If the fluid is applying a force on the solid the forces will most likely not sum to zero. From your output, it looks like Mechanical is sending small (but non-zero) displacements. I would expect that the motion of the corresponding walls in Fluent to match that in Mechanical. So it's not that the walls will not deform, but instead that the deformations are not building up to produce observable deformations. I recommend to review your mesh and timestep to ensure that they're fine enough to adequately resolve the oscillations. From one of your images, it looks like your fluid mesh is very coarse.

Regards,

Ryan

Florian Henscheid posted this 23 July 2018

Hello Ryan, 

I have the calculation a few more times with several finer mesh and smaller time steps.
With one calculation I got up to 17% with the other only up to the first time steps.
The error message was repeatedly: negative cell volume detected. What can i do to fix the problem?

regards,

Florian

Ryan O'Connor posted this 26 July 2018

Hi Florian,

The good news is that negative cell volumes indicate that you're getting motion in Fluent. The bad news is that this is a common issue with deforming meshes that may require some trial and error. First try to track down the root cause of the issue. Perhaps Mechanical is sending unphysical deformations (this may be caused by incorrect settings, such as material properties, in Mechanical). Or perhaps Mechanical is sending unphysical deformations because Fluent is sending unphysical forces to Mechanical. This may be caused by solution instability, which can be addressed by introducing solution stabilization. Or perhaps the motion is reasonable and the problem simply that you do not have good smoothing/remeshing settings. I recommend to take the FSI and MDM training courses so that you can get a better understanding of these issues and how to troubleshoot them.

Regards,

Ryan

Close