CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Regards the circular cylinder threads (http://www.cfd-online.com/Forums/main/4467-regards-circular-cylinder-threads.html)

Andrew Parker February 26, 2002 11:24

Regards the circular cylinder threads
 
dear all, some questions

I have been reading the recent threads about flow round a circular cylinder. I am an undergrad, and I have just completed very good experiments for the flow around a circular cylinder in a subsonic wind tunnel. The results compare very well with Schlichting and although the Re is 120000 I have found that the boundary layer is laminar, even although the free-stream is turbulent.

Fluents un-steady laminar results compare very well with my experimetal data, and hence as a natural progression I wanted to trip the boundary layer and then use a turbulence model to predict those results.

As regards the recent post I am now rather worried that I don't know which turbulence model to use. Should I use unsteady RMS, or un-steady k-e, with or without this correction for C_mu.

Could really do with some input on this,

Cheers

Andy

Jonas Larsson February 26, 2002 11:50

Re: Regards the circular cylinder threads
 
I'd start with the Realizable k-epsilon model - Shih's realizability fix tends to reduce the problems with overproduction of turbulent energy which otherwise often kill unsteadyness.

sylvain February 26, 2002 12:01

Re: Regards the circular cylinder threads
 
You can also try the Spalart-Almaras model.

Adrin Gharakhani February 26, 2002 16:33

Re: Regards the circular cylinder threads
 
This is actually a question for you rather than an answer to your question!

I am curious to learn how well your experiments match your numerical data. The boundary layer of your problem is laminar but the aft-body shear layer is turbulent. Since, I gather from your post, you are not using any turbulence models I'd be interested to see some results. What do you consider to be a good match, for example? Based on what criteria? It would be beneficial for the rest of us, I assume, to learn about grid-resolution and some solid numbers.

I have to make one comment. I hope you are not comparing streaklines/pathlines because you can usually get good comparison here but your vorticity field may be way off. Can you post some vorticity contours or time-dependent drag data, etc. etc.?

Thanks in advance

Adrin Gharakhani

Andrew Parker February 27, 2002 07:52

Re: Regards the circular cylinder threads
 
Hi,

I have the pressure distribution over the cylinder for 180 degrees of rotation, experimentally taken by myself, Re=120000. I have Schlichting, which has the pressure distribution, for inviscid, subcritical and supercritical pressure distributions. From those results my pressure distribution lies on top of the laminar one or subcritical. To verify this I have used an un-steady laminar solution in Fluent, and the pressure distribution by Fluent lies on top of my experimental data. My time averaged drag coefficient is close but not identical to the my experimental data, which compares again very well with the subcritical value of Schlichting. I should mention that I am an undergrad, so my understanding is that if the boundary layer going round the cylinder is laminar, and it then separates and produces a Karman vortex street, then I do not understand how there can be a turbulent shear layer, or what you mean by aft body shear layer. I have used stream functions to look at the flow, but only to make an animation not for quantitative work.

Hope this answers you question I would like some feed back, especially about the turbulence, not sure if I will be able to send any data to the website, I will have to look into that.

Cheers again

Andy

Adrin Gharakhani February 27, 2002 16:13

Re: Regards the circular cylinder threads
 
> so my understanding is that if the boundary layer going round the cylinder is laminar, and it then separates and produces a Karman vortex street, then I do not understand how there can be a turbulent shear layer

"Transition" from laminar to turbulent flow is quite a common phenomenon (though not well understood, yet).

At your Reynolds number the wake becomes 3-D, for sure, and it surprises me that using a 2-D simulation (I'm assuming here) and with (apparently) no mechanism to account for transition to turbulence you are getting good comparisons. (Of course, "good" is a subjective measure - we need to look at the flow quantitatively)

Perhaps others with more direct experience with flow over circles can shed some light here.

Adrin Gharakhani


All times are GMT -4. The time now is 13:34.