[lovelybugatti]

Hi
I need some help in simulating JET flow.
As you know 2d planar Jet has analytical solution, I want to test a 2d planar incomprehensible laminar jet with Open-foam and comparing it with analytical result,
The results that I obtained till now are far away from analytical solution. I changed the gird size, domain size, Boundary condition and tried different kind of solvers such as icoFoam, pisoFoam and rheoFoam (Newtonian) case, but still results are not even close. Can you please help me in this regards.
Best regards

[piu58]

Normally, such refrence cases are well covered by OF. We may only help you, if you give the case and the result you expect und got.

[lovelybugatti]

Originally Posted by piu58

Normally, such refrence cases are well covered by OF. We may only help you, if you give the case and the result you expect und got.

Please see the attached, this is the case in Rheofoam. I used the same case for Icofoam. I already compared my DNS code result with analytical solution and they are match. But OF results are far away.

blockMeshDict.doc

constitutiveProperties.doc

controlDict.doc

p.doc

U.doc

I will upload some result in next post.
Best

[lovelybugatti]

Originally Posted by piu58

Normally, such refrence cases are well covered by OF. We may only help you, if you give the case and the result you expect und got.

As you can see DNS and analytical solution are matched but openfoam is not.
untitled.jpg
Yours

[piu58]

I let the case run. Before it works, I had some elements to add and the given slightly to change. IcoFoam crahses after a few rounds.

If you look at the results you see strange effects at the inlet. I assume this is caused by your inlet b.c. for pressure: The pressure is 0 all over the inlet. But there are different levesl of inlet velocity, which need different pressure levels.

[lovelybugatti]

Originally Posted by piu58

I let the case run. Before it works, I had some elements to add and the given slightly to change. IcoFoam crahses after a few rounds.

If you look at the results you see strange effects at the inlet. I assume this is caused by your inlet b.c. for pressure: The pressure is 0 all over the inlet. But there are different levesl of inlet velocity, which need different pressure levels.

Dear Pilz
I agree with you. But the results ,I sent for you, are based on assuming constant pressure at outlet. As I could not get comparable results with analytical solution,I thought the problem can be because of boundary condition so I decided to play with BC, inlet outlet and.. (the thing that you are discussing now)
Please recheck the results in my former posts (when
inlet is assumed zero gradient and outlet is constant pressure) you can see that the results are physical but can not be compared with analytical solution and my DNS data. That is my problem, I do not why? The velocity profile should have sharp gradient in the middle of domain (parabolic shape only in the middle) and remain almost constant at the rest(in analytical and DNS), but OF result has parabolic shape in all domain no matter how the domain length is.
I have not found any explanation or solution for it.
Best regards

[piu58]

Good morning,

it is not easy to help you if the information you give is incomplete. I recommend:
- Establish the case in a way you think it is ok
- clean all results and the polymesh folder
- zip it
- re-run it to be sure that the zipped case is complete
- store the zip file at a public place.

Beside this I don't understand your graphic fully. Did you calculate the DNS results by yourself? If not, then is your OF result the single one. May it possible that you made some scaling error? From which publication do you got the results?

[lovelybugatti]

Originally Posted by piu58

Good morning,

it is not easy to help you if the information you give is incomplete. I recommend:
- Establish the case in a way you think it is ok
- clean all results and the polymesh folder
- zip it
- re-run it to be sure that the zipped case is complete
- store the zip file at a public place.

Beside this I don't understand your graphic fully. Did you calculate the DNS results by yourself? If not, then is your OF result the single one. May it possible that you made some scaling error? From which publication do you got the results?

OK I did what you said.
These are the links of my cases.

1- RheoFoam case:
https://1drv.ms/u/s!AuzMDkqqTaK0hk1t93qyfZ8VEFSr

2-IcoFoam case:
https://1drv.ms/u/s!AuzMDkqqTaK0hkyVLPq_xuNgnqeo

I have a DNS code (six order ,..) and I compared its results with analytical solution, and they matched very well.(no publications were used to comparison, I only used analytical solution).
About scaling error I used the same algorithm to non-dimensionalize the results.
Yours

[lovelybugatti]

Originally Posted by piu58

Good morning,

it is not easy to help you if the information you give is incomplete. I recommend:
- Establish the case in a way you think it is ok
- clean all results and the polymesh folder
- zip it
- re-run it to be sure that the zipped case is complete
- store the zip file at a public place.

Beside this I don't understand your graphic fully. Did you calculate the DNS results by yourself? If not, then is your OF result the single one. May it possible that you made some scaling error? From which publication do you got the results?

Dear Mr. Pilz

As you can see in the attached files, I am simulating Laminar Planar Jet with different solvers (icoFoam and RheoFoam).

The initial conditions and boundary conditions are the same. The domain size and grid size are equal too. and also the Reynolds number. However,

the results of icoFoam are comparable with analytical solution in the far fields (although it has some perturbations)
but the result of Rheofoam are not valid in this case.

Can you please check the files and let me know what is wrong in the RheoFoam case?
The other thing is do you know how can eliminate the icoFoam perturbation case? Do you think if I assume finer grids, I can handle that?
Best regards

[piu58]

I started icoFoam and will say something about the solution tomorrow.

rheoFoam does not belong to my OF installation.

[piu58]

I run the simulation, but again something is missing: The sampleDict. So I give up.

But some last remarks:

I assume that you reference the theoretical solution given by Hermann Schlichting (1933).
There are some ways to get a factor in (which you have I think). One is the momentum and another one ist the mass density, a third one is the viscosity and a last one is the transformation for x, y any using eta.

Im am quite sure that you made a mistake. From your graph your calculated expands far more in the free fluid than expected. If I look at your solution with paraView I see that the jet widens at about 50% before in gets unstable. The jet cannot be much smaller, no way.

It would be useful if you gave your results in eta and not in U30-U0 .... because it is much closer to the theoretical solution.

[pachojazz]

Hi piu58 im with the same problem and I have doubts with the boundary condition in top and bottom patches ,What boundary conditions did you use in these ones ?