CFD Online Discussion Forums - How to integrate a parameter in whole domain(domainIntegrate(alfa))

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)

- OpenFOAM Programming & Development (https://www.cfd-online.com/Forums/openfoam-programming-development/)

- - How to integrate a parameter in whole domain(domainIntegrate(alfa)) (https://www.cfd-online.com/Forums/openfoam-programming-development/235515-how-integrate-parameter-whole-domain-domainintegrate-alfa.html)

How to integrate a parameter in whole domain(domainIntegrate(alfa))

Dear foamers
I want to use domainIntegrate(alfa) to integrate a parameter in all my domain. I checked the main code for domainIntegrate(alfa) in below address;

https://cpp.openfoam.org/v3/a05655_source.html

and it is written as below;

Code:

   

   82 

   83 template<class Type>

   84 dimensioned<Type>

   85 domainIntegrate

   86 (

   87     const GeometricField<Type, fvPatchField, volMesh>& vf

   88 )

   89 {

   90     return dimensioned<Type>

   91     (

   92         "domainIntegrate(" + vf.name() + ')',

   93         dimVol*vf.dimensions(),

   94         gSum(fvc::volumeIntegrate(vf))

   95     );

   96 }

   97 

   98 

   99 template<class Type>

  100 dimensioned<Type> domainIntegrate

  101 (

  102     const tmp<GeometricField<Type, fvPatchField, volMesh> >& tvf

  103 )

  104 {

  105     dimensioned<Type> integral = domainIntegrate(tvf());

  106     tvf.clear();

  107     return integral;

  108 }

  109 

  110 

  111 template<class Type>

  112 dimensioned<Type> domainIntegrate

  113 (

  114     const DimensionedField<Type, volMesh>& df

  115 )

  116 {

  117     return dimensioned<Type>

  118     (

  119         "domainIntegrate(" + df.name() + ')',

  120         dimVol*df.dimensions(),

  121         gSum(fvc::volumeIntegrate(df))

  122     );

  123 }

  124 

  125 

  126 template<class Type>

  127 dimensioned<Type> domainIntegrate

  128 (

  129     const tmp<DimensionedField<Type, volMesh> >& tdf

  130 )

  131 {

  132     dimensioned<Type> integral = domainIntegrate(tdf());

  133     tdf.clear();

  134     return integral;

  135 }

  136 

  137

Now, my question is very simple;
if I want to integrate alfa over whole domain, I should use domainIntegrate(alfa), but what is alfa? I mean it is volScalarField? or Scalar? or dimensionedScalar?

Thanks,
Farzad

Extra explanation

For clarification, I used pressure to integrate which is a volScalarField and it gives me error;

Code:

In file included from DPMFoamFarzadd.C:216:0:

YO2cEqnYN2cEqn.H: In function ‘int main(int, char**)’:

YO2cEqnYN2cEqn.H:20:40: error: cannot convert ‘Foam::dimensioned<double>’ to ‘Foam::scalar {aka double}’ in assignment

   sumcloudSYO2 = fvc::domainIntegrate(p);

and it gives me this error;

Code:

YO2cEqnYN2cEqn.H:20:40: error: cannot convert ‘Foam::dimensioned<double>’ to ‘Foam::scalar {aka double}’ in assignment

Do you know why?

Thanks,
Farzad

Ok, instead of writing

Code:

fvc::domainIntegrate(p)

it must be written;

Code:

fvc::domainIntegrate(p).value();

Thanks,
Farzad

The clarification of the problem was discussed within 3 text messages in LinkedIn between him and myself., 😉

Yes, Thanks a lot Tobias Holzmann. I am still working on other parts of domainIntegration and if I find something new, I will write it down in this post.

Quote:

Originally Posted by Tobi (Post 801863)

The clarification of the problem was discussed within 3 text messages in LinkedIn between him and myself., 😉

New problem after integration

I am doing a lagrangian particel tracking using DPMFoam, and inside lagrangian part(src/lagrangian/intermediate/parcels/Templates/KinematicParcel), you need to add

Code:

#include "fvCFD.H"

in order to use fvc::domainIntegrate(alfa).value();

I did this, and code compiles and works fine, but still I have a problem after doing the integration. Let's I explain it;

Code:



        scalar np0 = 1

        Info<< "dMass[0] = " << dMass[0] << endl;

        

        scalar dm = np0*dMass[0];                                                              

        label gid = 0;      

        cloud.rhoTrans(gid)[this->cell()] += dm;

        

        Info<< "cloud.rhoTrans(gid)[this->cell()] = "<<cloud.rhoTrans(gid)[this->cell()]  <<endl;

                 

        scalar sumKinematicParcelYO2 = 0.;



         sumKinematicParcelYO2 = fvc::domainIntegrate(cloud.rhoTrans(0)).value();



        Info<<"WaterfordMizz(insideKinematicParcel.C) - sumKinematicParcelYO2 ="<< sumKinematicParcelYO2 <<endl;

And the result is this one

Code:

        dMass[0] = 3.5044329e-28 

        cloud.rhoTrans(gid)[this->cell()] = 3.5044329e-28

        WaterfordMizz(insideKinematicParcel.C) - sumKinematicParcelYO2 = 3.4299845e-34

        

        Or in the next Lagrangian time step(2) 

        

        dMass[0] = 2.6911542e-15 

        cloud.rhoTrans(gid)[this->cell()] = 2.6911542e-15

        WaterfordMizz(insideKinematicParcel.C) - sumKinematicParcelYO2 = 2.6339832e-21

        



        And in the next Lagrangian time step(3) 

        

        dMass[0] = 3.1145925e-15 

        cloud.rhoTrans(gid)[this->cell()] = 8.7192363e-15

        WaterfordMizz(insideKinematicParcel.C) - sumKinematicParcelYO2 = 8.5340044e-21

Now, we have the value inside one cell, and the remaining cells are empty(initialized zero). After integration, integration results is 6 order magnitude less than the value inside one cell, why?

Farzad

I expect that domainIntegrate is not suitable for particles. Check out any lagrangian code to see how they create the overall mass or whatever you want to do.