How to model a turbulent flow over a cylinder?
 

I want to investigate a fluid flow over a circular cylinder for Re=3900. I use k-ebsilon turbulent model and unsteady solver. But I could not find the approprite stream function (I found symmetry form). It doesn't show any vorticity in the wake.

Could you please help me to model any turbulent flow over a circular cylinder?

Re: How to model a turbulent flow over a cylinder?
 

See this exemple: http://www.fluent.com/software/sf_me...l_cylinder.htm

Re: How to model a turbulent flow over a cylinder?
 

Dear Thanks for your reply. Yes, I tried according to that tutorial but I am confused- In this tutorial 'Laminar' viscous model is considered for Re=150. Is it applicable for Re=3900? I know, at Re=3900 the flow becomes turbulent, so how can it capture a turbulent flow with that laminar model? I selected the model 'define_model_viscous_k-ebsilon', but it didn't demonestrate any vortex shedding. I could not understand what is my fault. Please help me if you know.

Thanking you mahbub

Re: How to model a turbulent flow over a cylinder?
 

Hi,

I notice in the tutorail they use periodic BC. Why is that? What if we use axis BC, will we still get the fluctutaing lift in the monitor panel? I have the problem that my lift and drag is not simulating unsteady behaviour i.e. fluctuation but only converge to a certain value. Could you pls clarify.

Sham.

Re: How to model a turbulent flow over a cylinder?
 

you dont need to use periodic or axis. just use velocity inlet and outflow. regarding the k-epsilon, it should worked well. i used it before and the results were fine. you better make the grid really fine near the wall. and make sure you use the right time step size (you can read it in the fluent tutorial).

or, you can change the rho to V (velocity ratio).

For example: you Re=4000 you have a fix d (cylinder diameter) and miu (viscosity), then you can change the rho and V as long as the Re remains the same.

Lets say the d=1 and miu is 1.10-5. your true value of rho is 1 and V is 0.04 m/s.

change the ratio between rho and V. lets say, rho to V is 50:1 then rho is 1.4142 and V is 0.02828. usually, higher rho makes the shedding easily visible

Re: How to model a turbulent flow over a cylinder?
 

Another way you can use depends on what post-processor you are using. If using Fluent's display, you can not change the max value of the vorticity, that is why you can not see anything. If you use Tecplot, you can export the Fluent data file into tecplot data file including the vorticity magnitude. At first, you might not see any vorticity in Tecplot. After you change the display level of the vorticity from very large to a little smaller, you will see the vorticity much clearer.

I am not very sure about the relationship between vorticity magnitude and Re. Maybe when Re increses, the vorticity magnitude should increase too.

Hope this will help you!

Re: How to model a turbulent flow over a cylinder?
 

Dear Naomi, Haoyin and others

Thanks for your co-operations but I still couldn't display the vortex shedding in the contour of stream function or velocity magnitude. I used the mesh and followed the instruction from (Just select k-ebsilon turb. instead of laminar and change density to 3900 to change the Re number others are as usual) http://www.fluent.com/software/sf_me...l_cylinder.htm what's the wrong with me. Please help me.

Mahbub

Re: How to model a turbulent flow over a cylinder?
 

Mahbub,

Actually, I simulated the flow over a cylinder couple months ago. In my understanding and my experience, you can not simulate the vortex shedding at a large Re with K-epsilon model. It is straightforward that the vortex shedding should be more significant with the increase of Re, however, it would not be the truth. I am not very sure about this. I did simulate the vortex shedding and Von Karman vortex street at Re = 140. I think that is why the tutorial you mentioned use that specific Re.

You can do further research about the relationship between high Re and vortex shedding degree.

Re: How to model a turbulent flow over a cylinder?
 

I think you should read the following post carefully:

http://university.fluent.com/forum/v...6ef5bb5a23e1e8

Hi,
I am modelling the turbulent flow over the cylinder by fluent(k-epsilon model)
I have a basic question about defining the boundary conditions for the domain. The domain is rectangular. for specifying velocity inlet for domain, I have to define the turbulent intensity as well as hydraulic diameter, I want to know I should use the diameter of the cylinder as the Hydraulic diameter or the cross length of the inlet face?!
Thank you

Default settings should work for most of the problems. Regarding your particular question ,appropriate value is dia of cylinder.

Default settings should work for most of the problems. Regarding your particular question ,appropriate value is dia of cylinder.

tutorial
 

Hi every body
I need a tutorial for turbulent flow over a cylinder...Can you help me?
please help me..(sasanghomi@yahoo.com)
thanks.

Hi dear mahbub
In this raynold number we don't have vortex shading.
Note that vortex shading hapend jaust in laminar flow,Re>40 and <300
Good luck

Problem with modeling turbulent flow around a cylinder
 

Hi everybody
I'm modeling a turbulent flow around a cylinder but I have a problem
my Reynolds number is 1000 and I'm using k-omega SST model when I use steady model for my simulation, Residuals don't converge and come down then become constant and then start to oscillate (all of them).
but when I use transient model it converges very well and reaches to a steady condition after e.g. 5 seconds
but I want to model a steady flow. can anyone plz help me to solve this problem.
even when I make my mesh better I still have this problem. and even when I reduce URFs the result doesn't change. I even checked y-plus and was 0.29 that was OK for this turbulence model.
Is it possible to simulate steady flow in this Reynolds number? (or higher, cuz I still have the same problem with higher Reynolds-numbers)
and another point that when I use first order solution it converges but for second order solution, I have the so-called problem.