Jacobi: error (divergence) solving for Velocity

[BLM]

Hello,

As you noticed, there is a typo in the documentation of v4.0. The correct expression of the evaporation/condensation source term is the one of v4.2 documentation. It differs slightly from the original Merkle model. I will made the Code_Saturne documentation more explicit on this point.
Note that there is no need to specify irovar = 1 in the cavitation module. irovar is related to monophasic flow while cavitation module is based on the homogeneous two-phase mixture approach.

Concerning your simulation of cavitation in the venturi, you should not pay too much attention to the warning message about the time step, espescially at the beginning of the simulation. The warning indicates that the mass conservation could be deteriorate due to void fraction clipping but actually the global mass budget is log in the listing and in your case it is not that bad (about 1E-7).
Concerning the low values of void fraction in your simulation, you are wright : it is because the total pressure you impose at outlet is about 1 bar thus there is no cavitation in the venturi in these conditions.

Kind regards,
Benoît

[Luciano Garelli]

Hello,

As Benoît mention the problem was in the pressure. In your case if you set a and the outlet pressure according to the cavitation number you will get the cavitation of the fluid at the throat. I have run you case with a dt=5e-6 [s] and the results are reasonable according to the paper "Development of algorithm based on the fractional time step for the simulation of cavitation".

Also I notice that the dimensions of your venturi are smaller that the used in this paper, so you have to be careful if you want to compare the results, because you Reynolds number will be different. It is a very interesting case to be solved, there are a lot of experimental information. Perhaps can be created a tutorial with this case, like the Shear driven cavity flow.

I attach a figure of the void fraction being advected.

Regards,

Luciano

[amir]

Hello,

I was under the impression that the problem is due to the high Pressure, so I run two different cases with two different pressure (5.7e4+P0 (Pa) and 2.2e4+P0 (Pa)), however the results are still without change, where the solution is clipped by the code and the void fraction equals to 1e-10 for the Min / Max values.

By default, I hold the Pref=101325(Pa) in GUI and applied the gauge Pressure in subroutine (e.g. x+P0 at outlet) so I believe I have any problem with my Pressure setting.

Moreover, I observed in exactly the same cases, when I decrease the reference time, the void fraction increases. As for the cases above, I applied 5e-7 as my reference time where the solution was clipped. But with a really smaller reference time (i.e. Tref=1e-10) the void fraction is calculated, even if it is far from good.

Please note that the pressure is less than 2000 (Pa) at the edge (Paraview image is attached) before beginning the cavitation simulation, therefore I suppose I should have the cavitation in this zone.

Could you also please specify the version which you are working with?

Please go to the following link to download the listings, SRC files and Paraview image:

https://cloud.utc.fr/public.php?service ... 809db2e993

I would appreciate it if you could give me some more feedback on this issue.


Kind regards,
Amir

[Luciano Garelli]

Hello,

The cavitation number and vapor pressure are absolute pressures. So, if you want to use a , the outlet gauge pressure for a given cavitation number is given by, , wherewith .

In the figure that you attach you are showing the gauge pressure, not the absolute pressure. The lower pressure is -58500 [Pa] wherewith the absolute pressure is

I think that this is the problem in your model. I'm using CS 4.0

Regards,

Luciano

[amir]

Hello Luciano,

I do not think so, maybe I did not explain well; I would say when I apply P0=Pref, I also sum the Pref at my outlet, so the outcome is the same with the case where I do not use Pref neither in GUI nor at outlet.

By the way :
with P0=Pref : Outlet Pressure = (0.5*Sigma*Rho*Vref*Vref) + Psat + Pref
And with P0=0 : Outlet Pressure = (0.5*Sigma*Rho*Vref*Vref) + Psat

To be sure I tested two cases:
1-I took my old non-cavitation case (with P0=Pref and Pref at outlet) and rerun it with icavit=1, P0=0 in GUI and without Pref at outlet but I obtained the same result as the other.

2-As your suggestion, I began another case from zero with P0=0 , outlet pressure = 5.4e4 (related to sigmsa=2.6) and reference time = 5e-6 till a good convergence, then I activated the cavitation. The result is always with clipping.

I am really stuck on this problem. I am pretty sure this is due to the high Pressure but I can not find where I am wrong?

please go to following link for downloading the listing and SRC files:
https://cloud.utc.fr/public.php?service ... 3914bbe214

Could you also please send me your case and your listing file?

Many thanks
Kind regards
Amir

[Luciano Garelli]

Hello,

I attach the case, the listing and the source files. I noted that you changed the Cdest and Cprod parameters. I hope that you find it useful and can solve this problem

Regards,

Luciano

[amir]

Hello Luciano,

I would like to thank you for your suggestions and your interest in my case.
Finally I could get a pretty good result.


Kind regards,
Amir

[Luciano Garelli]

Hello Amir,

You are welcome. I'm glad you were able to solve your problem.

Regards

Luciano