[david]

Hi all

I'm trying to simulate a pump (inlet casing, rotating impeller, diffuser and volute) with OpenFOAM-1.5-dev. The four domains are connected with GGI interfaces. The turbDyMFoam solver is used as it is a transient simulation. The forces acting on the hub, shroud and blades are monitored with the forces functionObject. Unfortunately the pressure forces oscillate with each time step. The attached image shows the pressure forces in y-direction. Rotation axis is the z-axis.


I could reproduce this behaviour in the icoDyMFoam tutorial cases mixerGgi and mixer2D. I tried to use different ddt-Schemes (Euler, backward, CrankNicholson) and tolerances without any improvement. Are there any possibilities to reduce the oscillations?


Thanks in advance

David

[louisgag]

Hi David,

I assume that by oscillations you mean the noise and not the sinusoidal oscillation of the signal? If so, I've read that using a smaller timestep might reduce the noise.

cheers,

-Louis

[david]

Hi Louis

Yes, the noise is the problem. Unfortunately, reducing the time step doesn't improve the results.

Regards

David

[markc]

Hello,

try to use some relaxation on p (or pd). First check whether relaxationFactors are implemented in the solver you use, otherwise you can easily implement them yourself.

Succes,

Mark

[sek]

Originally Posted by david

Hi Louis

Yes, the noise is the problem. Unfortunately, reducing the time step doesn't improve the results.

Regards

David

David,

I have similar prroblems you've had. With

[sek]

Originally Posted by david

Hi all

I'm trying to simulate a pump (inlet casing, rotating impeller, diffuser and volute) with OpenFOAM-1.5-dev. The four domains are connected with GGI interfaces. The turbDyMFoam solver is used as it is a transient simulation. The forces acting on the hub, shroud and blades are monitored with the forces functionObject. Unfortunately the pressure forces oscillate with each time step. The attached image shows the pressure forces in y-direction. Rotation axis is the z-axis.


I could reproduce this behaviour in the icoDyMFoam tutorial cases mixerGgi and mixer2D. I tried to use different ddt-Schemes (Euler, backward, CrankNicholson) and tolerances without any improvement. Are there any possibilities to reduce the oscillations?


Thanks in advance

David

=================================
Hi David,

I am having a similar "noise" problem. Were you able to resolve your problem?