[sharonyue]

Hi,

Nowdays I intend to make a test of my mrf case's result with the paper.But I find the result vary with the different diameter of my rotating size.I made a comparision in two situation in which only the MRF zones size is different.As you can see in the image.(its a slice of a 3D case)

The first one's MRF zone is smaller than the other.expect this everything is the same.But after 2000 iterations,the velocity is a little different.I dont know if it gets convergence but in 2500s iterations,the velocity looks dont change anymore.

in this thread linnemann happend to run into this situation. but I didnot try his case.http://www.openfoam.org/mantisbt/view.php?id=409

Time = 2093

smoothSolver:  Solving for Ux, Initial residual = 0.000157922, Final residual = 1.01677e-05, No Iterations 2
smoothSolver:  Solving for Uy, Initial residual = 0.000172252, Final residual = 9.70874e-06, No Iterations 2
smoothSolver:  Solving for Uz, Initial residual = 0.000168408, Final residual = 1.10209e-05, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.00286416, Final residual = 6.07327e-05, No Iterations 3
time step continuity errors : sum local = 0.000134333, global = -1.86372e-05, cumulative = 0.176384
smoothSolver:  Solving for epsilon, Initial residual = 5.02636e-05, Final residual = 1.76898e-06, No Iterations 2
smoothSolver:  Solving for k, Initial residual = 0.000214597, Final residual = 1.16413e-05, No Iterations 2
ExecutionTime = 3003.18 s  ClockTime = 3019 s

MRF1
{
    type            MRFSource;
    active          true;
    selectionMode   cellZone;
    cellZone        innerCylinder;

    MRFSourceCoeffs
    {
	nonRotatingPatches ();
        origin      (0 0 0);
        axis        (0 1 0);
        omega       31.4159;
    }
}

[Dothan]

Hi,
I found a similar result issue in a different geometry MRF problem.
Are there some news?

Dothan

[pechwang]

Hi everyone,

Same thing happens to me as well. Different size of MRFzones gives different results.
And for me, none of these gives me the same results as the analytical solution.

[Dothan]

Hi pechwang,
Do you have a turbomachinery analytical solution to compare CFD solutiuons?
I know only Eckardt Impeller.

Dothan

[pechwang]

Hi Dothan,

No, the analytical solution is for lubrication model. Unfortunately, MRF cannot give me correct results.

Thanks,
Pengchuan

[be_inspired]

Same problem here. When MRFZone limit is near walls ( sometimes it is required because stator zone is quite near) the pressure field around the walls is modified. The pressure lines are modified to accommodate its shape to the limit of the MRF Zone so suction lobe is affected and so all the aerodynamics of the wall (wake,....). The solution is wrong.

I do not have AMI interfaces so the solution proposed does not work.
Did you start a bug report?

[Krao]

Hi all,

I am also confused about the size of MRF region. With all the parameters constant, each time I change my MRF region size I get different results. I have made several simulations in this regard and found out that, whenever MRF domain is very close to upper part of the propeller, It usually gives wrong results and wrong representation of flow. A domain size of around 1.1D width and 0.3-0.5D height have given good results, which are closer to the experimental values. Very small MRF regions have given poor results.

Because of having two MRF region closer to each other I am unable to find a proper solution to the above problem of domain sizing. It would be great if anyone of you can share your experience with MRF domain size and how you overcome this problem.

Thank you
Krao

[anthony761]

I have observed the same with centrifugal compressor calculations on a commercial software. I saw that with increasing rotational domain (downstream of the impeller), the compression ratio increases.

[Wolfram]

well, okay ... I am not the only one facing this issue....

Did anyone get to any solution?

[sylvester]

See this old thread and the excellent explanation of Eugene here: Should an empty MRF zone really induce this much cross-flow?

[Wolfram]

Originally Posted by sylvester

See this old thread and the excellent explanation of Eugene here: Should an empty MRF zone really induce this much cross-flow?

This is the answer. Although the MRF-Approach is not really suitable for my case, the post is calming me at least searching for my mistake in the setup.
Thank you!