[sharonyue]

hey guys,

In OpenFOAM, icofoam solve N-S equations with no additional models, AFAIK, its the same with DNS, so can I use icoFoam to simulate turbulence? Just by set delta t is very small such as:1e-7?

Thanks.

[santiagomarquezd]

http://www.tfd.chalmers.se/~hani/kur...de_Mare_v3.pdf

[sharonyue]

Originally Posted by santiagomarquezd

http://www.tfd.chalmers.se/~hani/kur...de_Mare_v3.pdf

Hi Santiago,

I know there is a dnsFoam, but to my experience and Versteeg's book page 110:

The instantaneous continuity and Navier–Stokes equations (3.23) and
(3.24a–c) for an incompressible turbulent flow form a closed set of four equations
with four unknowns u, v, w and p. Direct numerical simulation
(DNS) of turbulent flow takes this set of equations as a starting point and
develops a transient solution on a sufficiently fine spatial mesh with
sufficiently small time steps to resolve even the smallest turbulent eddies and
the fastest fluctuations.

Equations 3.23 and 3.24 a-c is just what is implanted into icoFoam, so does it mean icoFoam= DNS?

Anyway Im not familar with dnsDoam.
Thanks.

[santiagomarquezd]

If you check the code for dnsFoam you will see that it is similar to icoFoam, except for a source term which adds turbulence to the system. Without it the problem would evolve to a resting system due the dissipative effect of the viscosity.

Regards

[ooo]

Originally Posted by santiagomarquezd

If you check the code for dnsFoam you will see that it is similar to icoFoam, except for a source term which adds turbulence to the system. Without it the problem would evolve to a resting system due the dissipative effect of the viscosity.

Regards

I would appreciate if you answer these 2 questions :

1)should we except same result if we solve a laminar case by icoFoam and dnsFoam?Does that extra source term influences on the result for a laminar case?

2)Is there any difference between the cost of dnsFoam and icoFoam for a identical laminar test case?
for example that extra source term might decrease the cost or not...?

Thanks in advance

[santiagomarquezd]

Hi,

1) It depends on the relative intensity of the source term, but in general, yes.
2) They are the same solvers, the source terms add a little more cost.

Regards

[huangxianbei]

Originally Posted by santiagomarquezd

If you check the code for dnsFoam you will see that it is similar to icoFoam, except for a source term which adds turbulence to the system. Without it the problem would evolve to a resting system due the dissipative effect of the viscosity.

Regards

Hi,Santiago:
Ichecked the DNS solver and found a force add to the r.h.s of the UEqn, and find the definition:

 volVectorField force
    (
        U/dimensionedScalar("dt", dimTime, runTime.deltaTValue())
    );

What's the reason to add this term? And what does this term mean?

I've looked into someone's code in this forum about channel flow DNS, the 'force' is not added, so I'd like to know is it necessary to do so?

[itchy]

Hi,

in the description of dnsFoam.c you can find:

Application
dnsFoam

Description
Direct numerical simulation solver for boxes of isotropic turbulence

In my opinion you only need this force, if you don't have shear or a wall that gives you shear. So dnsFoam is proper for the boxturb16 and for no other case.

In the tutorial of Martin de Mare, he mentioned that the force term is to keep the turbulence alive.
You only need this force, if you have no shear of e.g.shear at a wall? Because if you have a wall, turbulence keep alive by the shear, right?

I want to simulate a quasi-DNS of pipeflow. From the description, dnsFoam solver is for DNS of isotropic turbulence in boxes. So this solver is not suitable for my case.

Now my question:
Is icoFoam the right solver for DNS in complex geometries with shear at a wall??

kind regards
Florian

[Nidalsb]

Hi, have you figured this out yet? As I have a similar problem to solve and Im not sure whether I can use icofoam for such a case.