New scalar in phaseModel.c

bestucan

I guess you use myBeta_ just like alphaMax: myBeta(), or just by its name?
the position where you put the declaration of myBeta_ shows that myBeta_ is a private variable.
myBeta_(phaseDict_.lookuporDefault("myBeta", 1.0)), is OK, because it's just initialization.

you should define a member function to return a private variable just like alphaMax().

the errors seems like lost address of something. Maybe the private variable's address without authority.

kimy

@bestucan Thanks. I defined "myBeta" according to the definition of alphaMax in all positions where it occurs.
Also

        scalar alphaMax() const
        {
            return alphaMax_;
        }

        scalar myBeta() const
        {
            return myBeta_;
        }

bestucan

@kimy try this way: just define myBeta_ and do not use it.

see is there any error. if not, the errors are caused by function myBeta().
otherwise, keep on delete the part of code you added until the errors disapear.
you'll find the "error" code.

if the error occurs after you delete all the modification. you should check the makefile.

principle: make sure origin code works well. then, add your code slowly. (add one or two lines and run it )
so it's easy find which part of code is wrong.

kimy

I really don't know why the error always happen in which I did not used myBeta only definition. Even thought, I only added one piece of codes, such as "scalar myBeta". The same error occurs. Could you please help me to try to define one scalar in those file and see what will happen? Really thanks for your help!

bestucan

here is diff result of lib file and case file compared with origin files.
diff.zip

output:

PIMPLE: Iteration 1
alphaMax():0.62alphaMax lue lue lue lue lue lue
alphamyBeta():1.23myBeta ha ha ha ha ha ha ha ha
alphaMax():1alphaMax lue lue lue lue lue lue
alphamyBeta():1myBeta ha ha ha ha ha ha ha ha
MULES: Solving for alpha.particles
MULES: Solving for alpha.air
particles fraction, min, max = 0.275003 0 0.55
air fraction, min, max = 0.724997 0.45 1
Phase-sum volume fraction, min, max = 1 1 1
alphaMax():0.62alphaMax lue lue lue lue lue lue
alphamyBeta():1.23myBeta ha ha ha ha ha ha ha ha
alphaMax():1alphaMax lue lue lue lue lue lue
alphamyBeta():1myBeta ha ha ha ha ha ha ha ha

my way(without mention the modification shown in diff file ):

copy entire multiphaseEulerFoam into $FOAM_RUN/applications/solvers (bu not ./Allwamke)

can@x260-TC ~/Documents/code/C/openfoam/can-8/applications/solvers %tree -L 2                                                                                                                      (git)-[master]-
.
└── multiphaseEulerFoam
    ├── Allwclean
    ├── Allwmake
    ├── functionObjects
    ├── include
    ├── interfacialCompositionModels
    ├── interfacialModels
    ├── multiphaseCompressibleMomentumTransportModels
    ├── multiphaseEulerFoam
    ├── multiphaseThermophysicalTransportModels
    └── phaseSystems

9 directories, 2 files

find and replace all $(FOAM_LIBBIN) with $(FOAM_USER_LIBBIN)

find ./ -type f -exec sed -i 's/\$(FOAM_LIBBIN)/\$(FOAM_USER_LIBBIN)/' {} \;

enter phaseSystems run wmake to build libmyphaseSystem.so
enter multiphaseEulerFoam run wmake(not ./Allwamke) to build solver mymultiphaseEulerFoam
run the case

kimy

@bestucan Many Thanks. I have a confusion that you defined the following two scalar and function in phaseModel.H

<         scalar myBeta_;
195,196d193
< 
<         scalar myBeta() const;

While in original file, according to the definition of alphaMax

        scalar myBeta() const
       {
            return myBeta_;
        }

and you put this code in phaseModel.C ?

/home/can/.local/share/OpenFOAM/OpenFOAM-8/applications/solvers/multiphase/multiphaseEulerFoam/phaseSystems/phaseModel/phaseModel/phaseModel.C
88,89c88
<     alphaMax_(fluid.subDict(phaseName).lookupOrDefault("alphaMax", 1.0)),
<     myBeta_(fluid.subDict(phaseName).lookupOrDefault("myBeta", 1.0))
---
>     alphaMax_(fluid.subDict(phaseName).lookupOrDefault("alphaMax", 1.0))
143,147d141
< }
< 
< Foam::scalar Foam::phaseModel::myBeta() const
< {
<     retu

bestucan

@kimy yes, declaration in .H file and definition in .C file.

declaration of function myBeta() in phaseModel.H

//- Return the maximum phase-fraction (e.g. packing limit)
        scalar alphaMax() const;
        scalar myBeta() const;

and definition of function myBeta() in phaseModel.C

Foam::scalar Foam::phaseModel::alphaMax() const
{
    return alphaMax_;
}

Foam::scalar Foam::phaseModel::myBeta() const
{
    return myBeta_;
}

kimy

@bestucan But why alphaMax() was originally defined in the phaseModel.H without any error?

        scalar alphaMax() const
        {
            return alphaMax_;
        }

After I defined only scalar myBeta() in phaseModel.H, the following error occurs:

In file included from phaseModel/phaseModel.C:26:0:
phaseModel/phaseModel.H:158:29: error: invalid declarator before 'd'
         tmp<volScalarField> d() const;
                             ^
phaseModel/phaseModel.H:162:26: error: non-member function 'const Foam::PhaseCompressibleTurbulenceModel<Foam::phaseModel>& Foam::turbulence()' cannot have cv-qualifier
             turbulence() const;

bestucan

@kimy because the position of this part

 scalar alphaMax() const
        {
            return alphaMax_;
        }

in phaseModel.H belong to the declaration of class phaseModel.
so you do have to name alphaMax() as phaseModel::alphaMax()
as for "Foam::", just for rigorous.

"declaration in H and define in C" is not something like ban. just a rule free to follow. and it's a good habit.

you had edit your last post and delete the important part. I've seen it. you lost a ";". and error also said "invalid declarator before 'd'"

kimy

@bestucan Many thanks. The same error occurs, same as original when I run my case.

Could you please upload your phaseModel.C and H files here?

bestucan

pasheModel_and_systemsolve.zip

systemsolve is where I call myBeta()

kimy

@bestucan I am a little bit crazy because my files are totally same as yours.......Still error after run. Note that I did not change the default library into user's library (All of them I leave it as defualt "LIB = $(FOAM_LIBBIN)/liblibcompressibleTwoPhaseSystem"
).

kimy

@bestucan

Selecting default blending method: none
Selecting dragModel for (solids in water): SchillerNaumann
Selecting swarmCorrection for (solids in water): none
#0  Foam::error::printStack(Foam::Ostream&) at ??:?
#1  Foam::sigSegv::sigHandler(int) at ??:?
#2  ? in /lib64/libc.so.6
#3  ? at ??:?
#4  __libc_start_main in /lib64/libc.so.6
#5  ? at ??:?
Segmentation fault (core dumped)

bestucan

@kimy do it slowly.

1 find a normal case which works well. ensure basement is good
2 add myBeta() only, and not use it. ensure compile works
3 use cout << myBeta() << "somg flag". ensure function myBeta() works.
4 use myBeta_ as a constant value in calculation. ensure framework is well.
5 use myBeta_ as variable

divide int more step if needed

then use it as your wish. special case, special solver. If something wrong. there's nothing to do with myBeta(). maybe case? maybe solver? maybe variable overflow?

you combine all steps into one step. it's hardly to debug.
have you print myBeta_ value? just same as the value in case file?

the error message seems not a special error. something about cout (Ostream&)

kimy

@bestucan Thanks. Your suggestions are very helpful. However, the solver met the same error mentioned before. I uploaded my solver and case files. Hope you can see them if you wish (I cannot upload). I modified the solver based on twophaseeulerfoam.

kimy

2020twoPhaseEulerFoam.zip

bestucan

I've tried. seems like your OpenFOAM version is different with mine. my version is 8

kimy

@bestucan Thanks a lot. What you mean is that you did not meet any problem in your OpenFOAM version? Could you please upload the whole modified solver files here (twophaseeulerfoam)? The files I uploaed above was copied from blueCFD (maybe openfoam 5). Additionally, I am also using HPC where openfoam-v1906 is loaded.

kimy

This is the solver file of openfoam_v1906

slurrypipebendsolver.zip

kimy

I have already reached my goal in another way. Thanks for all helps.

CFD中文网

New scalar in phaseModel.c