CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Problem with patch rotation using pointDisplacement (http://www.cfd-online.com/Forums/openfoam-solving/78064-problem-patch-rotation-using-pointdisplacement.html)

cwang5 July 12, 2010 08:23

Problem with patch rotation using pointDisplacement
 
Dear all,

I've been on this problem for ages, and could really use some help from you guys. I am working on a 1 DoF rotation solve adapted from the OF1.6.x pimpleDyMFoam solver. The rotational motion was set so that the patch would rotate an angle based on the fluid force until the patch comes within certain distance from the side wall. The output file from the solver looks like this:

Code:

Courant Number mean: 1.687486318e-07 max: 4.771403433e-06
Time = 0.005488


The current angle of rotation is 0.0101409026

The cummulative angle of rotation is 27.41162474

The rotational Tensor is (0.9999485815 -0.01014072879 0 0.01014072879 0.9999485815 0 0 0 1)
DICPCG:  Solving for cellDisplacementx, Initial residual = 8.323386496e-07, Final residual = 8.284646139e-09, No Iterations 5
DICPCG:  Solving for cellDisplacementy, Initial residual = 4.176329538e-07, Final residual = 4.176378834e-09, No Iterations 5
DICPCG:  Solving for cellDisplacementz, Initial residual = 6.907626607e-09, Final residual = 6.907626607e-09, No Iterations 0
DILUPBiCG:  Solving for Ux, Initial residual = 2.563012164e-06, Final residual = 2.563012164e-06, No Iterations 0
DILUPBiCG:  Solving for Uy, Initial residual = 2.83695964e-06, Final residual = 2.83695964e-06, No Iterations 0
DILUPBiCG:  Solving for Uz, Initial residual = 1.12315512e-07, Final residual = 1.12315512e-07, No Iterations 0
DICPCG:  Solving for p, Initial residual = 0.0002383668533, Final residual = 5.899082318e-06, No Iterations 2
time step continuity errors : sum local = 1.111412268e-17, global = 2.737806772e-21, cumulative = 1.364305262e-12
DILUPBiCG:  Solving for Ux, Initial residual = 2.728957463e-12, Final residual = 2.728957463e-12, No Iterations 0
DILUPBiCG:  Solving for Uy, Initial residual = 3.225840715e-12, Final residual = 3.225840715e-12, No Iterations 0
DILUPBiCG:  Solving for Uz, Initial residual = 4.912966005e-14, Final residual = 4.912966005e-14, No Iterations 0
DICPCG:  Solving for p, Initial residual = 5.876274734e-06, Final residual = 9.725726683e-07, No Iterations 4
time step continuity errors : sum local = 1.832042973e-18, global = -1.027336569e-19, cumulative = 1.364305159e-12
DILUPBiCG:  Solving for omega, Initial residual = 0, Final residual = 0, No Iterations 0
DILUPBiCG:  Solving for k, Initial residual = 0.0009829920571, Final residual = 3.05954052e-11, No Iterations 1

Initializing Moment Calculation...

Calculating Moment...

The force moment acting on the flap is -0.2993934284
ExecutionTime = 140.92 s  ClockTime = 147 s

As you can see from the output log, the cumulative angle that was passed down to the rotation matrix indicated that the patch would have rotated around the origin for a few times already, yet in the paraFoam visualization very little motion was noticed.

I am wondering if it has to do with the tolerance setting of pointDisplacement or with the way that I define the pointDisplacement field in each time step:

Code:

forAllConstIter(labelHashSet, patchList_, iter)
{
        label patchi = iter.key();
        vectorField oldCell = mesh.C().boundaryField()[patchi];
        pointField oldPoint = mesh.boundaryMesh()[patchi].localPoints();
        vectorField newCell = ((oldCell-CoR) & RotTen) + CoR;
        pointField newPoint = ((oldPoint-CoR) & RotTen) + CoR;
        pointDisplacement.boundaryField()[patchi] == newPoint - oldPoint;
        cellDisplacement.boundaryField()[patchi] == newCell - oldCell;
}

which was based on the codes by Erik Ekedahl on one of the OpenFOAM PhD course website
http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2008/

I would really appreciate some help in resolving this issue.

Best Regards,

John

Philer July 23, 2010 10:26

You should have applied the cumulative, rather than current, angle while overriding the operator ==. Looks like you are doing the opposite.

cwang5 July 24, 2010 07:05

Quote:

Originally Posted by Philer (Post 268710)
You should have applied the cumulative, rather than current, angle while overriding the operator ==. Looks like you are doing the opposite.

Umm... what do you mean by cummulative? Was the pointDisplacement value supposed to be a cummulative value from the start time to the current time?
I think I've fixed the motion solver by only applying the rotation algorithm every 5 timesteps for now, but I couldn't seems to get the rest of the grid points to move with the patch, any idea what might have cause it? Thanks.

John

Philer July 26, 2010 09:54

Yes, the pointDisplacement value is supposed to be a cummulative value from the start time to the current time.

The new version 1.7 is very good at this. Try it.

Philer


All times are GMT -4. The time now is 22:04.