[tonyuprm]

Hi all,

I wanted to know if there are any validation studies on OpenFOAM (DNS) using the flat channel from Kim and Moin 1987. I have been trying to do this but my results do not seem to converge towards their data.

Thanks,

Tony

[stevenvanharen]

Hi Tony,

I obtained good results for the channel flow with OpenFOAM.

If you send me your e-mail address I can send you my MSc. Thesis, this will be helpfull for you.

[pooyan]

Hi Steven

I have just started to learn open foam and use it for the channel flow problem. Would you kindly guide me on this problem or let me know of any useful reference?!
My email Address is: pooyanrazi@yahoo.com

Thanks,
Pooyan

Originally Posted by stevenvanharen

Hi Tony,

I obtained good results for the channel flow with OpenFOAM.

If you send me your e-mail address I can send you my MSc. Thesis, this will be helpfull for you.

[sunliming]

Originally Posted by stevenvanharen

Hi Tony,

I obtained good results for the channel flow with OpenFOAM.

If you send me your e-mail address I can send you my MSc. Thesis, this will be helpfull for you.

Hi Steven，
I'm working on that too，and my results seem not good，especially in the log-law zone. I'm a little bit confused, would you be kind to send me one copy of your MSc. Thesis. My email:sunliming59045@126.com or sunliming59045@hotmai.com.
Thank you so much!
Sunliming
Here is my Uf-y+ plot：

[francescomarra]

Dear Tony,

I too have performed comparative tests of the channel flow case in the framework of a national initiative in Italy (LESinItaly).
I adopted OpenFOAM while others groups adopted several others commercial, opensource and academic software. Main results are published in the proceedings of AIMETA 2011:
http://www.lamc.ing.unibo.it/aimeta2011/

All papers can be downloaded from:
http://www.lamc.ing.unibo.it/aimeta2...Aimeta2011.zip

Note that we chosen the approach to fix the forcing, not the flux as done in the channelFoam application. This makes the differences with DNS results larger.

From our results OpenFOAM performs as well as the average of commercial and multi-physics codes.

My best regards,

Franco

[Ohbuchi]

Hi Franco,
Thank you very much for your important information about DNS on OpenFOAM.
I've got the archive, but I couldn't read your paper; MEM-331-2.pdf.
It may be broken.

Best regards,

Masashi Obuchi

[francescomarra]

Dear Masashi,

you are right, the direct link to the paper is broken. Unfortunately I do not have the possibility to fix it. Two options are available to get the paper:
- download the entire cd rom of the proceedings (it should works, it is the second link I posted);
- send me your e-mail and I will reply with the paper in attach.

Let me know which option you prefer.

Best regards,

Franco

[Ohbuchi]

Dear Franco,

Thank you very much for your kind proposal.
I've got the whole CD archive from second link in your message.
But I couldn't open your paper(MEM-331-2.pdf) via Acrobat Reader.
Please send it to following e-mail address.
ohbuchi@amber.plala.or.jp

Best regards,

Masashi Obuchi

[francescomarra]

Dear Masashi,

I have just sent you by email the copy of the paper.
Sorry for the delay but I had to recollect it from source.

Best regards,

Franco

[Beaver]

Dear Franco,

Could you send me a copy of your paper?
My email: krasnopolsky@imec.msu.ru

Thanks!

[francescomarra]

Dear Boris,

I have just sent you by email the copy of the paper.
Best regards,

Franco

[kingjewel1]

This is a source for some benchmark CFD experimental data, which you might not know about:

http://cfd.mace.manchester.ac.uk/ercoftac/classif.html

Feel free to add links other sites or repositories if you know of them

[Keqi Ding]

Dear Franco,

Could you send me a copy of your paper?
My email: dingkq@gmail.com

Thanks!

[francescomarra]

Dear Keqi Ding,

I have just sent you by email the pdf copy of the paper.
Best regards,

Franco

[kerim]

Dear Franco,

Could you send me your papar too?
My email:kurbanaliev@rambler.ru

Thanks in advance.

[MaryBau]

Hi Franco and all;

In a previous post (#5) you mentioned:

Originally Posted by francescomarra

Note that we chosen the approach to fix the forcing, not the flux as done in the channelFoam application. This makes the differences with DNS results larger.

Why is the reason for that difference in the results?

Thanks for your help,

Mary

[kerim]

Hi Steven

I have just started to learn open foam for DNS for the channel flow problem. Would you kindly to send me a copy of your Master thesis?!
My email Address is: kurbanaliev@yahoo.com

[francescomarra]

Dear Mary,

sorry for my late answer. The question is simply explained. By fixing the flux you adjust at each time step the forcing function to ensure the correct flux. Ensuring the flux means to already satisfy the average of the velocity component normalt the coss section. Therefore errors in the computation of the stresses do not appears as flux errors.
Conversely, by just fixing the average pressure gradient, i.e. the force that "move" the fluid, and waiting for the average flux to equilibrate, it will be a results of how much well the stresses are computed. The computed numerical viscosity will affect the possibility for the medium to easely flow in the channel, and therefore the final value of the flux. If many errors are present in the computation they are anymore yet corrected during the computation to converge always to the same value of the flux, but accumulate and discrepancies can be quite large.

I hope this clarify the concept.

Best regards,

Franco

[huangxianbei]

Originally Posted by francescomarra

Dear Mary,

sorry for my late answer. The question is simply explained. By fixing the flux you adjust at each time step the forcing function to ensure the correct flux. Ensuring the flux means to already satisfy the average of the velocity component normalt the coss section. Therefore errors in the computation of the stresses do not appears as flux errors.
Conversely, by just fixing the average pressure gradient, i.e. the force that "move" the fluid, and waiting for the average flux to equilibrate, it will be a results of how much well the stresses are computed. The computed numerical viscosity will affect the possibility for the medium to easely flow in the channel, and therefore the final value of the flux. If many errors are present in the computation they are anymore yet corrected during the computation to converge always to the same value of the flux, but accumulate and discrepancies can be quite large.

I hope this clarify the concept.

Best regards,

Franco

Hi,
Just as you say, fixing the flux and fixing the force are two ways to simulate channel flow. If the flux is fixed, that means the velocity across the channel is identified. While if the pressure gradient is fixed, the flux is pushed to vary according to the pressure gradient. I've used the two ways to simulate in Ansys CFX, however, when the channel rotates, only the fixed pressure gradient seems correct, and a strange thing happens that the velocity is almost 20times the data in papers. So in your opinion, which way is recommended?

[kozden]

Hi Steven,

I know this is an old subject but I have just seen this topic in the forum. I'm quite new in OpenFOAM and trying to simulate channel flow in my thesis and it would be very good if I can have a look at your thesis. Could you please send your thesis to my e-mail which is kamil.ozden.me@gmail.com . I'm looking forward for your mail.

Best Regards,
Kamil