CFD中文网

Thanks a lot. I wonder if the partial elimination algorithm mentioned in Passalaqcua's paper was implemented in the current version?

Thanks. I created two meshes with the same body size but different type of meshes,namely structured and unstructured, the results are different. Note that both of two cases are grid independent. Do you have any idea about that ? since I saw a thread you opened in cfd online.

Already solved.

Thanks a lot, Dongyue. Very helpful. I found in SN model CdRe was multiplied by another Re. Thus, the formula is consistent.

@bestucan Thanks for your attention.

Thanks a lot. The problem was solved.

Your simulations just blow up, its difficult to debug the reason from your info.

@kimy 在 Code production of k ? 中说：

@东岳 Many thanks. In fact, I understood this procedure. One confusion is that the formula implemented in openfoam is not consistent with that mentioned in the Spalding' paper.WeChat Screenshot_20200504110942.png

Nobody is interested in this issue?

I am checking that the variables outbut from source code in solver and I found the differences between output and paraFoam. I cannot understand why it happened? The follow is my codes in simpleFoam.C:
**const dimensionedScalar myZero("myZero", dimensionSet(0,0,0,0,0,0,0),scalar(0) );

volScalarField k0=myZeroturbulence->k();
volScalarField nut0=myZeroturbulence->nut();
volScalarField nu0=myZero*turbulence->nu();

const label patchID = mesh.boundaryMesh().findPatchID("wall");
const tmp<scalarField> tnuw = turbulence->nu(patchID);
const scalarField& nuw = tnuw();

const tmp<scalarField> tnutw = turbulence->nut(patchID);
const scalarField& nutw = tnutw();

const fvPatchVectorField& Uw = turbulence->U().boundaryField()[patchID];

const scalarField magGradUw(mag(Uw.snGrad()));

for (int j=0; j<nut0.size(); j++)
{
if (j<=(nutw.size()/2-1)) {nut0[j]=nutw[j]; nu0[j]=nuw[j];}
if (j>=(mesh.cells().size()-nutw.size()/2)&&j<=mesh.cells().size()-1) {nut0[j]=nutw[j-(mesh.cells().size()-nutw.size())];nu0[j]=nuw[j-(mesh.cells().size()-nutw.size())];}
}

volVectorField Uww=U*myZero;
const fvPatchList& patches = mesh.boundary();
labelList wallList;
wallList.clear();

forAll(patches, patchi)
{
const fvPatch& curPatch = patches[patchi];
if (isType<wallFvPatch>(curPatch))
{
forAll(curPatch, facei)
{
label faceCelli = curPatch.faceCells()[facei];
wallList.resize(wallList.size()+1);
wallList[wallList.size()-1]=faceCelli;
}
}
}

forAll(wallList,i)
{

Uww[wallList[i]] = U[wallList[i]];

}

volScalarField uz = Uww.component(vector::Z);
volVectorField SG = uz*vector(0,0,1);
volScalarField magGradUww=mag(SG)/myY;

k0=((nut0+nu0)*magGradUww/0.3);

Info<<"k0="<<k0<<endl;
Info<<"nut0="<<nut0<<endl;
Info<<"nutw="<<nutw<<endl;
Info<<"magGradUww="<<magGradUww<<endl;
Info<<"magGradUw="<<magGradUw<<endl;
Info<<"tauw="<<(nut0+nu0)magGradUww<<endl;
Info<<"Uw="<<Uw<<endl;*

nutw and magGradUw are copied from nutkwallfunction. In the end, wallshearstress in the paraFoam is not the same as (nut0+nu0)*magGradUww.

Thats a tmp, not a regular volScalarField, try turbulence->k().ref().boundaryFieldRef()[patchID] = 0;

alpha was solved in twoPhaseSystem.C

请教大家一个问题， 我正在用twofluidmodel计算solid-water flow。但是我们有自己的模型，需要在原有的基础上改写代码，将solid的effictive visicosity等于water的。其中有一个phasepressure model，可以在这里面改写phase2.nueff_, 并且包含一些header文件，但是我这样写了之后，提示没有declaration。请问怎么解决这个问题呢？

请问：固体颗粒的turbulent viscosity 怎么编程让其等于流体的？

@kimy 在 In solid-water flow 中说：

interphase slip agebrism

Never heard of that

That's internal field, a good approximation is good for convergence. From your depiction I feel you need to learn more about the general CFD knowledge. You can definitely feel that the lacking of CFD theory is holding you back now