[mo-ca]

Hi,
Im trying to compute a pressure driven movement of a simple steel beam and facing the problem, that both sides of one interface (solid - fluid) move differently.

I enclosed a picture for better understanding:


Both sides share the same expression of my defined movement and as one can see, it moves quite the expected way. But if you look closer, youll see the "original" position, which is the other part of the interface (solid).

I tried different meshing methods and definition of movements, but nothing really worked.

Its ansys cfx 14.5.

Thanks in advance

[ghorrocks]

I do not understand your question. What are you asking?

[mo-ca]

Originally Posted by ghorrocks

I do not understand your question. What are you asking?

Sorry if I didnt made my self clear enough.
The main problem is, that both sides of modelled interface move differently while having the same definition and I need some hints to find my error ..

[ghorrocks]

Do you mean the FEA bit of the simulation is moving but the mesh of the CFD is not moving? Have you done the FSI tutorials which come with CFX and the ones available on the ANSYS customer site?

[mo-ca]

Yes I did the fsi tutorials (Im trying 1way FSI, not 2way). The solid part ist not moving according to defined movement.

[ghorrocks]

Did the tutorials run as expected? Isn't your application similar to the tutorials?

[mo-ca]

the tutorial uses a 2way fsi, so cfx computes the fluid with resulting pressure forces and mechanical tries to deform you solid and returns the displacement, which is passed back to cfx for mesh deformation / displacement. what im trying to do is 1way fsi because of the computation time. so cfx is doing all the work (its a beta feature to deform solids using cfx).

during my runs it seems, that the not moving interface may be caused due to some meshing issue. if the mesh is coarse, sometimes it runs perfectly, but if you simply increase the number of elements, the simulation failes. unfortunately there is no "best practice" guideline. i tried using edge sizing, so that both sides of the interface do have the same amount of elements and spacing, but it didnt really help.

[edit]
I enclosed 4 images ... same geometry, meshing method and pre-setup was used

1. Mesh with Relevance 100; Solid part of interface blue; colored: fluid part
2. same as 1; other viewing angle
3. Mesh with Relevance 0; Solid part of interface
4. same as 3; fluid part of interface

I dont really understand, why the interface moves correctly with a coarse mesh and not with a fine one ...

[brunoc]

I'm guessing you're talking about the mesh deformation feature that's been in CFX for years.

If that's the case, how are you modelling the plate movement? Are you calculating it using forces and CEL Expressions? Or are they just moving based on a pre-set equation?

Either way, you probably have something wrong with your expressions. Post them here so we can see what you're doing and please share more information from your model.

[mo-ca]

hi and thanks for your help to this point.

within each side of my interface youll find the following displacement:

FLOW: Flow Analysis 1
  DOMAIN: Lamelle
    &replace BOUNDARY: INTF_Lamelle_Fluid Side 1
      Boundary Type = INTERFACE
      Interface Boundary = On
      Location = INTF_Lamelle_Fluid
      BOUNDARY CONDITIONS: 
        HEAT TRANSFER: 
          Option = Conservative Interface Flux
        END
        MESH MOTION: 
          Option = Specified Displacement
          DISPLACEMENT: 
            Displacement X Component = 0 [m]
            Displacement Y Component = Biegelinie2
            Displacement Z Component = 0 [m]
            Option = Cartesian Components
          END
        END
      END
      NONOVERLAP CONDITIONS: 
        Boundary Type = WALL
        BOUNDARY CONDITIONS: 
          HEAT TRANSFER: 
            Option = Adiabatic
          END
          MESH MOTION: 
            Option = Specified Displacement
            DISPLACEMENT: 
              Displacement X Component = 0 [m]
              Displacement Y Component = Biegelinie2
              Displacement Z Component = 0 [m]
              Option = Cartesian Components
            END
          END
        END
      END
    END
  END
END

and the corresponding expressions:

      Biegelinie2 = -.75 [N] * Lamellenlaenge / (6 * YoungModulus * Flaechentraegheitsmoment)*(zTransformiert)^2*(3-zTransformiert/Lamellenlaenge)
      KoordTransfWinkel = -90 [deg]
      Flaechentraegheitsmoment = Lamellenhoehe^3 * Lamellenbreite / 12
      Lamellenbreite = 45 [mm]
      Lamellenhoehe = 0.5 [mm]
      Lamellenlaenge = 90 [mm]
      Mesh Stiffness = Aspect Ratio^2 / max(0.01 [mm],Wall Distance) + volcvol ^(-1/3)
      YoungModulus = 2.06e5 [MPa]
      maxLamellenverformung = maxVal( Total Mesh Displacement )@REGION:Lamellenflaeche
      zTransformiert = -sin(KoordTransfWinkel) * (xGlobal - -0.024 [m]) + (zGlobal) * cos( KoordTransfWinkel )

the mentioned problem does also exist, if im using a specific displacement of 1 mm

[mo-ca]

anyone suggestions?

[stumpy]

You've set displacement on side 1 of the interface. What about side 2?

[mo-ca]

yes ... the specific displacement is set within the solid and fluid part of the interface ...