[petr.f.]

Nice Monday afternoon to all!

Recently I've run into a problem with moveDynamicMesh. I'd like to perform a simple flutter analysis with openfoam. I'm using the standard NACA 0012 profile, which is attached to en elastic axis EA (i.e. in 2D it is a point located on the profiles chord) and can oscillate around this axis and vibrate in vertical direction.
In order to place the profile into right initial position (e.g. Alpha_init = 3°, h_init = -0.01m) I'm using moveDynamicMesh solver and mesh created with GridPro. With this mesh I can perform steady simulations without any problem (the only error checkMesh reports are highAspectRatio cells, which is normal. However, if I add prescribed motion it always ends with many bad cells (no matter what time-step or tolerance parameters I use)

Checking geometry...
    Overall domain bounding box (-2.69997 -3 0) (1.5 3 0.015)
    Mesh (non-empty, non-wedge) directions (1 1 0)
    Mesh (non-empty) directions (1 1 0)
    All edges aligned with or perpendicular to non-empty directions.
    Boundary openness (-1.83973e-18 1.14303e-19 7.89943e-15) OK.
 ***High aspect ratio cells found, Max aspect ratio: 1265.93, number of cells 6
  <<Writing 6 cells with high aspect ratio to set highAspectRatioCells
    Minimum face area = 3.73703e-09. Maximum face area = 0.00237795.  Face area magnitudes OK.
 ***Zero or negative cell volume detected.  Minimum negative volume: -2.45686e-09, Number of negative volume cells: 5
  <<Writing 5 zero volume cells to set zeroVolumeCells
    Mesh non-orthogonality Max: 179.673 average: 7.00679
   *Number of severely non-orthogonal (> 70 degrees) faces: 83.
 ***Number of non-orthogonality errors: 9.
  <<Writing 92 non-orthogonal faces to set nonOrthoFaces
 ***Error in face pyramids: 28 faces are incorrectly oriented.
  <<Writing 24 faces with incorrect orientation to set wrongOrientedFaces
    Max skewness = 0.92386 OK.
    Coupled point location match (average 0) OK.

Failed 4 mesh checks.

Problematic cells are located on and behind the trailing edge. The motion itself is prescribed in a standard way using the pointDisplacement file in 0 directory

boundaryField
{
     ".*"
     {
         type            fixedValue;
         value           uniform (0 0 0);
     }
     
    wing
    {
        type            angularOscillatingDisplacement;
        axis            (0 0 1);
        origin          (0.12 0 0.0075);
        angle0          0;
        amplitude       -0.05235987755983; //rad
        omega           3.14159; //rad/s
        value           uniform (0 0 0);
    }

    "(front|back)"
    {
        type empty;
    }

}

dynamicMeshDict in the constant directory

dynamicFvMesh      dynamicMotionSolverFvMesh;

motionSolverLibs ("libfvMotionSolvers.so");

solver displacementLaplacian;
// solver velocityComponentLaplacian z;

displacementLaplacianCoeffs
{
    diffusivity quadratic inverseDistance 1(wing);
}

and this solver for point motion in the fvSolution

    cellDisplacement
    {
        solver          GAMG;
        tolerance       1e-5;
        relTol          0.0;
        smoother        DICGaussSeidel;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 10;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

Do you have any idea, what am I doing wrong or how to move the profile without obtaining errors?

[HenningW]

Even though this message is a bit older:

It looks similar to an error I had recently with a 3D mesh. The mesh within the domain distorted without any significant problem. However, the connection to the wall resulted in inverse faces, looking similar to yours (the patch moved but the rest of the surface mesh of a wall did not). I assume that the B.C. on the wall prevents the mesh from any movement. This seems to result in an invalid mesh. I never figured out a clean solution for that and resulted in leaving a small gap. In this case you might be able to move the empty patches as well (a cylindrical domain might come in handy for that) to avoid the static mesh on it.

[i.sabahi]

Originally Posted by petr.f.

Nice Monday afternoon to all!

I'd like to perform a simple flutter analysis with openfoam. I'm using the standard NACA 0012 profile, which is attached to en elastic axis EA (i.e. in 2D it is a point located on the profiles chord) and can oscillate around this axis and vibrate in vertical direction.

Hello Petr, did you manage to solve your problem and simulate the airfoil?


Regards
Iman