Dear Alexey,

What do you mean exactly by saying "we arrived to the point where you can describe the problem you are trying to simulate" and how did you understand that 0.1 MPa is to large. In my simulation, outside is environment. Therefore, i set it atmospheric pressure.

Ok, i will try 0 for pressure and will set the relaxation factors according to your advises.

I will let you know results.

Thanks so much

B.

Well,

I meant - "tell us about physical side of the problem you're trying to simulate", geometry, inflow/outflow conditions etc.

Cause I have almost no knowledge about your problem, I can only guess what's going on and give general hints like "change relaxation factors, change PIMPLE dictionary" but the problem maybe elsewhere (for example in wrong direction of g).

Hi
 

Dear Alexey,

I have tried following different cases:

1. 0/ k-epsilon I have changed wallFucntion setting =>as ZeroGradient

2. U inlet ==> FixedValue
outlet= intletOutlet and
outlet
{
type pressureInletOutletVelocity;
value uniform (0 0 0);
inletValue uniform (0 0 0);

3. i have used relaxation factors:
4. 0/p_rgh
inlet ==> zeroGradient
Outlet==> fixedValues ==>1e5, 1e4...0 (gradully decreased until 0)
and also for outlet i tried:

5. For gravity i checked direction it is true, but i also took it as pozitive..

All of these different challenges gives again following divergence error ::confused::confused::confused::confused:

I dont know what i can try..Please tell me some new ways..

Thanks in advance.

B

Hi,

well, as you don't want to share an exact description of the physical problem you are trying to simulate, let's continue guessing.

1. You run modified version of interPhaseChangeFoam, have you tried running original solver? Just to try to localize what's causing troubles with convergence.

2. The only modification you've made was that scaling factor for nut? Or you've done something else? Can you share your modified code?

I can continue with questions using just the pictures and solver output you've posted but think it won't be very productive.

hi
 

Dear Alexey,


Thanks so much for your reply. I think that you misunderstood. Of course i can share exact description of the physical problem. I am trying to simulate cavitation inside of nozzle. My nozzle shape is as you can see at thread #13.

1. yeah i have tried with original interphasechangefoam. I had one thread before you can see here: http://www.cfd-online.com/Forums/ope...foam_help.html

In original case, cavitation took place inside of nozzle, extended until some level and then disappeared. That was the problem.

After that in some papers i saw that RANS turbulence viscosity is calculated high and need to have some modification to decrease the nut Now what i am trying to do.

2. Yes i have only modification which is bold in the file of U.Eqn as follow:

sorry again for misunderstanding.

Please let me know what kind of info you want.

Thanks

B.

Well, cause modeling of cavitation is not something that I do, my thoughts will be quite general (i.e. maybe completely useless ;)).

1. When you were doing simulations with original interPhaseChangeFoam, what was in your PIMPLE dictionary? Are you sure that the solution converged on every time step?

2. If you need lower turbulent viscosity why not start with just modifying Cmu (if you're using k-epsilon model). Set it 100 time lower (9e-4) and see what will happen.

3. Finally are you sure that settings for cavitation model are correct?

In general you've got several points of possible errors: ICs/BCs, cavitation model settings, transport model settings, your modification of the code. See the influence of these things on the results, locate the source of the problem (well, I guess it'll be code modification as without it code was at least running and simulation wasn't blowing up). Plot muEff with your modification, compare it to the values without modifications. See if it even make any sense.