[alimea]

Hi all

I have a main question:

Is this true that under relaxation factors are not affected the solution in PISO algorithm?

[Santiago]

even worse: using Underrelaxation in PISO is an error

[Taataa]

It affects simulations. The purpose of implementing PIMPLE algorithm instead of PISO is possibility of having larger time steps. Say for example you have a model with large simulation time and instead of using a PISO loop which convergence is bounded by relatively small time steps you can utilize under-relaxation to achieve convergence in larger time steps e.g. instead of having a time step of 0.1 sec with PISO loop you can set it to 1 sec at the cost of more computation time per time step.

[AndreasPe]

As the others mentioned, in OpenFOAM, the PISO algorithm does not contain under-relaxation. PIMPLE in contrast uses under relaxation to enable the use of large timesteps.

To your question: You could investigate on that by yourself by simulating test cases with different underrelaxation configurations and check if the results are changing. There was also published a paper about that topic: https://www.sciencedirect.com/scienc...77042716301649

There also seems to be a timestep dependence of the solution which has to be tackled when conducting unsteady simulations.

Also, in Foam-Extend-4.0, Prof. Jasak changed the SIMPLE and PIMPLE implementations to cope with these problem, see for example his training course at the 11th OpenFOAM Workshop http://openfoam-extend.sourceforge.n...s/courses.html

[alimea]

Originally Posted by AndreasPe

As the others mentioned, in OpenFOAM, the PISO algorithm does not contain under-relaxation. PIMPLE in contrast uses under relaxation to enable the use of large timesteps.

To your question: You could investigate on that by yourself by simulating test cases with different underrelaxation configurations and check if the results are changing. There was also published a paper about that topic: https://www.sciencedirect.com/scienc...77042716301649

There also seems to be a timestep dependence of the solution which has to be tackled when conducting unsteady simulations.

Also, in Foam-Extend-4.0, Prof. Jasak changed the SIMPLE and PIMPLE implementations to cope with these problem, see for example his training course at the 11th OpenFOAM Workshop http://openfoam-extend.sourceforge.n...s/courses.html

thanks for your complete
I'm using viscoelasticFluidFoam that is written based on PISO algorithm.
Now I'm investigating the time step independency for my case(vortex shedding behind a cylinder). But when I decrease dt, I see a phase difference in my CD diagram between dt=0.001 and dt=0.0005!! you can see that in below pic:




I thought that it may because of relaxation factor, but sb told me that PISO has no relaxation. However you can see in all of viscoelasticFluidFoam tutorials that the urf is applied!

[alimea]

Originally Posted by AndreasPe

As the others mentioned, in OpenFOAM, the PISO algorithm does not contain under-relaxation. PIMPLE in contrast uses under relaxation to enable the use of large timesteps.

To your question: You could investigate on that by yourself by simulating test cases with different underrelaxation configurations and check if the results are changing. There was also published a paper about that topic: https://www.sciencedirect.com/scienc...77042716301649

There also seems to be a timestep dependence of the solution which has to be tackled when conducting unsteady simulations.

Also, in Foam-Extend-4.0, Prof. Jasak changed the SIMPLE and PIMPLE implementations to cope with these problem, see for example his training course at the 11th OpenFOAM Workshop http://openfoam-extend.sourceforge.n...s/courses.html

Hi
I have solved two problems with and without urf and compared them with a paper results (Oliviera). the results are different!





you can see that there is a big difference between them!
However I couldn't get any result that can we use urf in viscoelasticFluidFoam (that is written based on PISO) or not?

Thanks
Note: most of tutorials of this solver has urf (0.5 or 0.3).

[alimea]

Hi

please answer me:

according to what I said in my last posts, can we use URF in viscoelasticFluidFoam solver?