[gu1]

Hello,

I'm having trouble calling a constant to be resolved within my turbulence model... could someone help me?

The problem is as follows, within the v2f.C file there is the function:

template<class BasicTurbulenceModel>
tmp<volScalarField> v2f<BasicTurbulenceModel>::Ts() const
{
    return max(k_/epsilon_, 6.0*sqrt(this->nu()/epsilon_));
}

However if I wanted to make this function receive a constant such as G:

const volScalarField G(this->GName(), nut*S2);

How would I do it?

I tried that way:

v2f.C

template<class BasicTurbulenceModel>
tmp<volScalarField> v2f<BasicTurbulenceModel>::Ts(const volScalarField& G) const
{
    return max((k_*G)/epsilon_, 6.0*sqrt(this->nu()/epsilon_));
}

...and:

v2f.H

volScalarField  Ts(const volScalarField& G) const;

...but I was not successful.

The CorrectNut() function requires the value of Ts() and since I didn't find out how to call the constant in the above function (Ts), I didn't progress.

Thanks

#OpenFoam 5.0

[t.teschner]

The cleanest solution I can see here is to use function overloading, i.e. provide two different functions with the same name and return type but with different function arguments.

// v2f.H file

//- default function declaration
volScalarField  Ts() const;

//- newly added function
volScalarField  Ts(const volScalarField& G) const;

// v2f.C

template<class BasicTurbulenceModel>
tmp<volScalarField> v2f<BasicTurbulenceModel>::Ts() const
{
    return max(k_/epsilon_, 6.0*sqrt(this->nu()/epsilon_));
}

template<class BasicTurbulenceModel>
tmp<volScalarField> v2f<BasicTurbulenceModel>::Ts(const volScalarField& G) const
{
    return max((k_*G)/epsilon_, 6.0*sqrt(this->nu()/epsilon_));
}

In this way both definition can happily coexist and the compiler will choose the correct function at compile time. For more info on function overloading, see this link: https://www.learncpp.com/cpp-tutoria...n-overloading/

[gu1]

Hello,

Thanks for the answer.

What about CorrectNut()? How would I do it? Since it calls the Ts() function return.

[t.teschner]

What exactly is the problem with CorrectNut()? Could you show the source code and explain what you are trying to do?

[gu1]

Originally Posted by t.teschner

What exactly is the problem with CorrectNut()? Could you show the source code and explain what you are trying to do?

template<class BasicTurbulenceModel>
void v2f<BasicTurbulenceModel>::correctNut()
{
    this->nut_ = min(CmuKEps_*sqr(k_)/epsilon_, this->Cmu_*v2_*Ts());
    this->nut_.correctBoundaryConditions();
    fv::options::New(this->mesh_).correct(this->nut_);

    BasicTurbulenceModel::correctNut();
}

The correctNut() function requires the response of the Ts() function. If I'm going to assign a constant to it, doesn't that constant need to be referenced? Like this:

template<class BasicTurbulenceModel>
void v2f<BasicTurbulenceModel>::correctNut(G)
{
    this->nut_ = min(CmuKEps_*sqr(k_)/epsilon_, this->Cmu_*v2_*Ts(G));
    this->nut_.correctBoundaryConditions();
    fv::options::New(this->mesh_).correct(this->nut_);

    BasicTurbulenceModel::correctNut();
}

[t.teschner]

well, the same logic as before applies here, you would need to have two function declarations in your header, i.e.

//- default
template<class BasicTurbulenceModel>
void v2f<BasicTurbulenceModel>::correctNut();

//- new declaration
template<class BasicTurbulenceModel>
void v2f<BasicTurbulenceModel>::correctNut(const volScalarField& G)

that propagates all the way down the call stack to where you actually define the constant G.

Since these are all part of the same class v2f, it might be worthwhile to consider making G a private / protected member of the class and then initialise it in the constructor or wherever it makes sense. In this way you don't have to pass the variable around all the time (that's a clear code smell and we should avoid it).

May I suggest to study up a bit more on c++ classes? I think there is some basic understanding missing, this is not a difficult problem (no offense) but with a bit more background in object orientated programming you will hit less of these problems.

[gu1]

Hi,

I appreciate your help and understand the reasons for making such a recommendation. I know how to do the basics, but unfortunately something I don't understand, because of how the OF is built. See the error below, this happens when I declare that correctNut() has to receive the constant G for it to be used in the function Ts():

In file included from /opt/openfoam5/src/TurbulenceModels/turbulenceModels/lnInclude/LESeddyViscosity.H:42:0,
                 from v2f.H:52,
/opt/openfoam5/src/TurbulenceModels/turbulenceModels/lnInclude/eddyViscosity.H:69:22: note: 	void Foam::eddyViscosity<BasicTurbulenceModel>::correctNut() [with BasicTurbulenceModel = Foam::LESModel<Foam::IncompressibleTurbulenceModel<Foam::transportModel> >]
         virtual void correctNut() = 0;
                      ^
/opt/openfoam5/wmake/rules/General/transform:25: recipe for target 'Make/linux64GccDPInt32Opt/makeTurbModel.o' failed

I understand that the relationship is in another file (eddyViscosity.H), but at the same time I don't know how to get around it.

I did it:

template<class BasicTurbulenceModel>
void v2f<BasicTurbulenceModel>::correctNut(const volScalarField& G)
{
    this->nut_ = min(CmuKEps_*sqr(k_)/epsilon_, this->Cmu_*v2_*Ts(G));
    this->nut_.correctBoundaryConditions();
    fv::options::New(this->mesh_).correct(this->nut_);

    BasicTurbulenceModel::correctNut(??); //I don't know if it's necessary.
}

[t.teschner]

The error is telling you that the file eddyViscosity.H only contains the definition for

virtual void correctNut() = 0;

So you need to add your new definition to that file as well. It may appear logical now to extend the file so that you have

virtual void correctNut() = 0;
virtual void correctNut(const volScalarField& G) = 0;

However, this is probalby going to result in an immediate or later compilation failure (because of the = 0 at the end, this is a pure virtual function and could result in other classes making use of that file to get a compilation error because you need to define this function then everywhere which is not what you want. This might all be confusing, but again, without background in polymorphism (in this case), i.e. virtual functions, this all may not make much sense)

TL;DR

modify the eddyViscosity.H file so that you have the following function declarations:

virtual void correctNut() = 0;
virtual void correctNut(const volScalarField& G);