[fernexda]

Dear all,

I have further questions concerning the forces calculation with moving mesh. I'm simulating a rotating wind turbine. My problem is that the resulting forces seem to be too high (I have a wind turbine with power coefficient much higher than 1...).

I was thinking maybe I had something wrong with my simulation, so I tried to run a simple 2D case of a cylinder with basically the same pre-processing data (cell size, boundary conditions, solver (pimpleFoam)). And I ended up with perfect values : I have a Cd error of 2%. So the simulation parameters seem to be in order.

To sum up, I have run the following simulations with the same input parameters (the main difference is the moving mesh) :

1. Cylinder simulation : perfect results, 2% error on the Cd
2. Moving mesh simulation : crazy results, forces too big

Therefore the error seems to be coming from the moving mesh. So my question is : does anyone have experienced wrong forces calculations because of a moving mesh ? Or does anyone have experienced right forces calculations ? Should I trust my results and look somewhere else for the error (even though I've looked into almost everything else...) ?

I would be glad if anyone having computed forces from a moving mesh could give their feedback (no matter wether it's positive or negative, I'd like to make myself an opinion on how confident I should be with my results) !

Thank you in advance for any answer!

Daniel
PS :

• Version OF 2.1.0
• Solver pimpleFoam / pimpleDyMFoam (for the moving mesh)

[Bazinga]

I just found your thread and I am having the same problems that you have. The forces are calculated quite well for the non-moving cylinder but they are getting too high for a moving mesh.

I am going through the simulation for some days now and I can not find the problem. I was trying 3D LES, 2D URANS but the values of the force are higher than expected.

Has anyone experienced similar behavior and found a solution to this? I am using the forceCoeffs function object in the controlDict file.

[fernexda]

Hi Bazinga,

I found the solution a long time ago but forgot to post it here.

In my case, it was simply an error in the numerical procedure, I applied the wrong boundary condition for U on the paddle.

• For the non-moving cylinder, I imposed U=0 on the cylinder surface, which is fine.
• For the moving paddle, I did the same, which is not correct. The relative velocity between the air and the paddle is zero, but the absolute velocity (air/fixed reference) is not zero on the paddle.

So the correct boundary condition for U on a moving solid with a no-slip condition is the movingWallVelocity. My input was:

type            movingWallVelocity;
value           uniform (0 0 0);

Hope this helps.

Regards,

Daniel

[Bazinga]

Thank you very much for your help. I made this exact mistake.

Best regards