[shipman]

Dear Foam Users,

I am trying to simulate cavitation phenomena inside of nozzle and using interphasechangefoam. As you know that there are 3 transport models (Merkle, Kunz and Sauer) I have tried all of these 3 models with kwSST turbulence model however i have problems with the obtaining of cavitation region as follows;

1. If i set the initial cond. according to Velocity as follows:
==> 0/U

internalField   uniform (0 3.104 0);
boundaryField
{
    inlet
    {
        type            fixedValue;
        value           $internalField;
    }
      outlet
    {
        type            pressureInletOutletVelocity;
        phi             phi;
        value           $internalField;
    }
    wall
    {
        type            fixedValue;
        value            uniform (0 0 0);
    }

==>0/P

internalField   uniform 1e5;
boundaryField
{
   inlet
    {
         type           zeroGradient;
    }
    outlet
    {
        type            fixedValue;
         value           $internalField;
    }
    wall
    {
        type           zeroGradient;
    }
}

cavitation starts to form until to some level, however after that it disappears although flow rate is increasing. I didnt understand why this happens? (as seen in the attached fig.)

2.if i set the initial cond. according to pressure as follows:
==>0/P

internalField   uniform 1e5;
boundaryField
{
    inlet
     {
        type            fixedValue;
        value           uniform 2.8e5;
    }
    outlet
    {
        type            fixedValue;
         value           $internalField;
    }
   wall
    {
        type           zeroGradient;
    }

==>0/U

internalField   uniform (0 0 0);
boundaryField
{
    inlet
    {
        type            zeroGradient;
        value           uniform (0 0 0);
    }
    outlet
     {
        type            inletOutlet;
        inletValue       uniform (0 0 0);
    value         uniform (0 0 0);
    }
    wall
    {
        type            fixedValue;
        value            uniform (0 0 0);
    }

in this case no cavitation form obtained. Therefore i am confused. Could someone advice me something or give some insight what kind of reasons can result in these results? I increased mesh number, changed mass transport model( means i tried 3 of models) still gives same results...

In addition; For interphasechange solver (or more generally if we try simulation multiphase with mass transition which initial condition setting is more logical; pressure or velocity based?

Thanks in advance...

[shipman]

Sorry I forgat to say that initial velocity and pressure values are taken from the experiment...

[alexB]

Hi Baris,

are you still working on that issue?

regards
Alex

[shipman]

Hi Alex,

Yes I am still working about same topic.

Do you have any question?

B

[alexB]

I didn't realize, that you are sure pretty experienced, so I assume that you aren't stuck at the same place like one year before...

I think i wanted to help you yeserday.

I currently try different things with InterPhaseChangeDyMFoam, but my problem is more a lack of c++ knowledge... thanks for your offered help and sorry for the unnecessary disturbance.

regards
Alex

[shipman]

Thank you for your kind consideration. Yes, i faced stability problem for a long time. However, finally i fixed and solved the problem.

I think that C++ is main problem for all new beginners. This can be also fixed in time with hard work

Please keep in touch and let me know about your challenge and brief findings in InterPhaseChangeDyMFoam.

BR

Baris

[alvariten]

Hi people,
I have seen you are using interPhaseChangeFoam to simulate cavitation phenomena inside nozzle. Actually I'm working in this topic too. I have a question about this solver I hope you could answer me: interPhaseChangeFoam works with a hydrostatic pressure definition, isn't it? Did you do your simulations with this pressure definition or did you do any change?
And... to make this change, is it necessary to modify the pEqn.H file, or is there another easy way?
Regards,
Alvaro.

[alexB]

Hi Alvaro,

did you try to work with the cavitatingBullet tutorial multiphase/InterPhaseChangeFoam/cavitatingBullet ?
You can change the geometry and customize it for your case.

regards
Alex

[shipman]

Hi Alvora,

Here hydrostatic pressure definition is just a kind of trick to simplify the boundary conditions and also increase the stability when the gravity should be considered. I just didnt understand why you want to change it? Finally, solver also calculate the total pressure as calculated and indicated in "CreateFields.H" P=p_rgh+rho*gh. As Alex said you can look at cavitatingBullet tutorial or last year i attended to CFD open course in Chalmers and already written a tutorial which explains cavitatingFoam and interPhaseChangeFoam for nozzle cavitation. You can also download and check it here: https://www.researchgate.net/publica...njector_Nozzle

BR.

Baris

[alvariten]

Hi,

I've seen your work Baris, and looks fine. I've tried to implement the code but I have a problem in the step to compile "phaseChangeTwoPhaseMixtures". The terminal returns me the following error and I don't know how to solve it.

alvaro@AlvaroPC:~/OpenFOAM/alvaro-2.3.0/applications/solvers/multiphase/TransportCavitatingFoam/phaseChangeTwoPhaseMixtures$ wmake
linux64GccDPOpt/options:5: *** missing separator.  Stop.
wmake error: file 'Make/linux64GccDPOpt/objectFiles' could not be created in /home/alvaro/OpenFOAM/alvaro-2.3.0/applications/solvers/multiphase/TransportCavitatingFoam/phaseChangeTwoPhaseMixtures
alvaro@AlvaroPC:~/OpenFOAM/alvaro-2.3.0/applications/solvers/multiphase/TransportCavitatingFoam/phaseChangeTwoPhaseMixtures$

I have to warn you that I have no knowledge of C++ programming. This are the Make files.
Make/files

phaseChangeTwoPhaseMixture/phaseChangeTwoPhaseMixture.C
phaseChangeTwoPhaseMixture/newPhaseChangeTwoPhaseMixture.C
Kunz/Kunz.C

LIB = $(FOAM_USER_LIBBIN)/libphaseChangeTwoPhaseMixtures

Make/options

Kunz/Kunz.C

LIB_LIBS = \
    -L$(FOAM_USER_LIBBIN)

EXE_INC = \
    -I$(LIB_SRC)/transportModels/twoPhaseMixture/lnInclude \
    -I$(LIB_SRC)/transportModels \
    -I$(LIB_SRC)/transportModels/incompressible/lnInclude \
    -I$(LIB_SRC)/finiteVolume/lnInclude

LIB_LIBS = \
    -ltwoPhaseMixture \
    -ltwoPhaseProperties \
    -lincompressibleTransportModels \
    -lfiniteVolume

And the file options generated in the folder linux64GccDPOpt.

# 1 "options"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "options"
Kunz/Kunz.C

LIB_LIBS = -L$(FOAM_USER_LIBBIN)


EXE_INC = -I$(LIB_SRC)/transportModels/twoPhaseMixture/lnInclude -I$(LIB_SRC)/transportModels -I$(LIB_SRC)/transportModels/incompressible/lnInclude -I$(LIB_SRC)/finiteVolume/lnInclude

Could you help me to compile the case correctly?
Thanks.

[alvariten]

Hi again,
I ran the simulation with interPhaseChangeFoam and I obtained high negative pressure values. I think it's wrong, but why does this happen? How can I avoid that?
Regards

[alexB]

Hi alvaro,

negativ pressure values aren't bad if they don't exceed -1e+05 Pa (-1bar) . I never gave it a thought, because I think its effective pressure and not absolute pressure plotted by paraView.

regards
Alex

[alexB]

I think I have to correct myself.
I read in some posts, that p is the absolute pressure in InterPhaseChangeFoam, so you have discovered really an error.

Edit:
Now I have to correct/clarify my correction.
My first suggestion

negativ pressure values aren't bad if they don't exceed -1e+05 Pa (-1bar) . I never gave it a thought, because I think its effective pressure and not absolute pressure plotted by paraView.

is not wrong, but confusing, because my second statement

p is the absolute pressure in InterPhaseChangeFoam

is also right.

As read here: http://www.cfd-online.com/Forums/ope...m-error-2.html

Originally Posted by pEqn.H

p == p_rgh + rho*gh;

p is the absolute pressure
p_rgh is the relative/effective pressure that is plotted by paraFoam
rho*gh is the atmospheric pressure

so everything is fine in my computation.

regards
Alex

[Kozan]

Hello Alex,

I'm not completely sure that you are right.You can find something about the negative pressure issue in this tutorial:http://www.tfd.chalmers.se/~hani/kur...lenceModel.pdf
page 16


Regards
Jiří

[shipman]

Hi Jiri,

In that tutorial he also said that negative pressure is unrealistic condition.

. To prevent
from this unrealistic condition, one option is using a very high value for evaporation coefficient

Frankly speaking, according to my experience, absolute pressure can not be negative. However, somehow for a small duration a local pressure around bubbles can be measured negative in some experiments. This is definitely unrealistic conditions. (Because absolute pressure min can be 0 under the lab condition. look at the definition here: https://en.wikipedia.org/wiki/Pressu...nization_gauge) Therefore, if you have so small negative calculated value then you can ignore it.

BR

Baris

[Kozan]

Originally Posted by shipman

Hi Jiri,

In that tutorial he also said that negative pressure is unrealistic condition.



Frankly speaking, according to my experience, absolute pressure can not be negative. However, somehow for a small duration a local pressure around bubbles can be measured negative in some experiments. This is definitely unrealistic conditions. (Because absolute pressure min can be 0 under the lab condition. look at the definition here: https://en.wikipedia.org/wiki/Pressu...nization_gauge) Therefore, if you have so small negative calculated value then you can ignore it.

BR

Baris

Hello,

Yes, I agree with this statemt. It was reaction to the last post. I have to confess that I read the thread only cursorily. On the other hand I thought that the tutorial could be useful. Have a nice day!

Regards,
Jiří

[yhdbuaa]

Hi Shipman,I just cant't understand why the pressure could below 0,some people say that may be induced by compressible property,can you explain more details? thanks a lot

[er_ijaz]

Dear all,


Sorry for reopening this discussion.


I'm working on pump sump models and trying to model the cavitation at the suction side of the pump.


I found in this forum that pressure should be applied as absolute and hence 1e5 value is used in tutorial


I have following doubt,
Since we have to input p_rgh value instead of p, then it should be 1e5-(1000(water)*-9.81*height) right instead of just 1e5?

[MiguelGC]

Dear all,
I'm triying to reproduce using OF the cases of asymmetrical nozzle that Baris present in his PhD Thesis (Sigma 1.91 and 1.19). for the latter case, I have similar troubles that BAris report here, concerning that the cavity appears, but then it dissapear. I try to model times longer that 10 times the particle residence into the nozzle (t_res=nozzle_lenght/c_nozzle= 8e-3m/(10-12)m/s=8e-4s), i.e., 8e-3s in order to have results with physical sense. I have initially a 2D mesh but with high resolution (cell size 4e-5m) that prove to have asymptotic behavior (fullfiling the CGI index and Richardson extrapolation checking). I set the same setups of Baris simulations but I check that the bubble dissapear as Baris initially has reported in this CFD forum. Dear Baris, you say that solved the problem, could you comment how you solved the problem? I also use Fluent soft obtaining good results with Shingal and Schnerr Sauer models, using SST kOm model, because the bubble remains. Also, I know that Baris' mesh is 3D, but I check the mesh quality results by using a 3D mesh too, and obtained the same flow pattern than the 2D case for Sigma 1.91, has was expected. At the present, I'm using OF .com, but when you reported the problem, both OF (.com, .org) sources were similar...