[kumar2]

Hello

I have a question on the turbForce term in the liftDrag.C (../src/postprocessing/incompressible/liftDrag/liftDrag.C ).

Laminar Flow
--------------
In the laminar case the drag is the pressure drag + x_component of the wall shear stress . although i am not an expert in C++ , i could make out that
-mu.value()*U.boundaryField()[patchLabel].snGrad()*mesh.magSf().boundaryField()[ patchLabel]

is approximately meu*velocityGradient*area . Am i correct ?

2. turbulent drag
--------------------
In turbulent flow the wall shear stress is (according to ferziger and peric) rho*(u_tau)^2.

which equals rho*(C_meu)^0.25*k*sqrt(k)*vel_parallel_wall/(ln(n+E))

But the formulation given in liftDrag.C is totally different . first the laminar drag is found out ( even in turbulent case ) . then the turbulent drag is found ( this does not at all resemble the expression given by say ferziger , peric). then these 2 are added.

could someone tell me why this procedure is adopted and where i am wrong ? i am stuck with my comparison.

thanks a lot

kumar

[kumar2]

Hello all,

Could someone comment on this post ?

i am stuck with my results

thanks in advance

kumar

[anja]

Hi kumar,

I'm very sorry that I can't help you, but maybe you can answer some of my questions?

- where do I get the mu field from?
- how do I have to change the code for a compressible fluid?
- where do I have to write the code?
- how can I actually use it?
- where do I find the results

thanks
Anja

[hartinger]

Hi Kumar,

the difference between laminar and turbulent navier-stokes equations are the reynolds-stresses.
So you can simply take the laminar forces and add the forces due to the reynolds-stresses.

laminar:
mu.value()*U.boundaryField()[patchLabel].snGrad()* mesh.magSf().boundaryField()[patchLabel]
means:
viscosity * velocity gradient normal to surface * area

turbulent:
- mesh.Sf().boundaryField()[patchLabel]
& turbulence->R()().boundaryField()[patchLabel]
means:
surface vector(length equal to area) &(dot-product) reynolds stress tensor
the calculation is general, given a stress tensor which is calculated from your choice of turbulence model

regards
markus

[kumar2]

Hi Anja,

Let me try to answer your questions .

1. mu field
-------------

the liftDrag utility uses nu ( mu/density ). this is set in the transportProperties in your case directory.

2. Compressible fluid

since liftDrag utitlity uses nu , you may have to modify the utility itself.

3. writing the code.

The liftDrag.C ( & associated codes ) code resides in /OpenFoam-1.2/application/utilities/postProcessing/miscellaneous/liftDrag/ . This program while executing calls another program also called liftDrag.C ( & associated programs ) in ../src/postProcessing/incompressible/liftDrag/

the liftDrag.C in /miscellaneous/../ calls createNu.H . createNu.H then opens transportProperties file ( this resides in the case directory )gets nu ( mu/rho ) . You can give a different nu by either modifying the transportProperties file . however you can also modify the later part of createNu.H to say for example read a certain different value , but here again , we are still talking about nu ( NOT mu)

to modify the code . copy the code to your /applications directory and follow the instructions in user manual . also refer other wonderful posts ( search with liftDrag in openFoam )

Once you compile your program you are ready to go!

hope this helps

regards

kumar

[kumar2]

Hi Markus

Thanks a lot for your reply. i hope i can trouble you with one more question.

In the case of ke models the reynolds shear stress are not calculated . this means that the wall shear stress given in liftDrag utility is an approximate formulation because in the turbulent case - while using ke model - the wall shear stress is (according to ferziger and peric) rho*(u_tau)^2. which equals rho*(C_meu)^0.25*k*sqrt(k)*vel_parallel_wall/(ln(n+E)) , but the utility does not calculate this and approximates with the laminar formulation.

Am i correct ?

regards

kumar

[hartinger]

Hi Kumar,

if you look for the k-eps model in src/turbulenceModels/incompressible/kEpsilon the reynolds stresses are calculated according to the boussinesq approximation common to all k-eps models.

((2.0/3.0)*I)*k_ - nut_*2*symm(fvc::grad(U_))

the turbulence viscosity nut is calculated like
nut_ = Cmu*sqr(k_)/(epsilon_ + epsilonSmall_)

with standard wall function approach to get the wall values for nut (wallViscosityI.H) with yPlusLam = 11.63

if (yPlus > yPlusLam_)
{
nutw[facei] =
nuw[facei] *(yPlus*kappa_/log(E_*yPlus) - 1);
}
else
{
nutw[facei] = 0.0;
}

regards
markus and pierre

[kumar2]

Hi Markus & Pierre

Thanks a lot for the replies

regards

kumar

[anja]

Hi all,
also thanks from my side.

Anja

Hi, I want to implement the viscous dissipation rate in my application. the term which im hopeing to get implemented is 1/2*1/Cp*(R()() & R ()())

on acount that R()() = tau/rho.

However the implementation of the dot product of R()() is not accepted.

Does anyone know how to do this?
Thanks
/erik

[83_Ale_83]

Hi to everybody, does anybody know where this equation

nuw[facei] *(yPlus*kappa_/log(E_*yPlus) - 1);

is coming from??

Thanks in advance

[83_Ale_83]

Originally Posted by 83_Ale_83

Hi to everybody, does anybody know where this equation

nuw[facei] *(yPlus*kappa_/log(E_*yPlus) - 1);

is coming from??

Thanks in advance

Anyone?
Thanks again