
[Sponsors] 
July 26, 2001, 20:00 
Re: flow over a cylinder, turbulent and laminar

#21 
Guest
Posts: n/a

A loworder discretisation scheme may be stabilising the solution (nonphysically) somewhat. Try using a higher order scheme.


July 26, 2001, 20:04 
Re: flow over a cylinder, turbulent and laminar

#22 
Guest
Posts: n/a

I use second order schemes.
After some reading and discussions, it seems that the vortex shedding will not be as obvious for a turbulent solution. And due to the added randomness, steady cycles may not be achieved. So my solution may be more accurate than i once thought, but I still need o understand the physics here. 

April 4, 2010, 22:24 

#23 
New Member
unreal
Join Date: Apr 2010
Posts: 2
Rep Power: 0 
Steve has given some valuable info about flow over cylinder. I have a question for all. When you say drag value at Re=10^6, what is it ? Is it average drag? Because the flow is turbulent the average value of drag can change with time(Correct me if I am wrong,please). How do you calculate this average drag. and how to we calculate the mean flow.
I have one more question. I am currently doing a Re=800 simulation without using any turbulence models. I have the flow solutions at different times. Can anybody tell me how to post process a turbulent flow (turbulence in wake in this case). What do I need to do to calculate TKE and how do I get the E(k) vs k plot. Thanks in advance . I think this is one of the best threads for flow over a cylinder 

December 5, 2010, 19:16 

#24 
New Member
Join Date: Mar 2010
Posts: 6
Rep Power: 7 
Hello all,
I agree; this really is a good thread. I have a general question regarding the unsteady wake of the 3D circular cylinder in cross flow problem I'm interested in:  The mesh I'm using is very fine, worked on it for quite a while. Also, the boundary conditions I set and the domain size is accurate and sufficient. My problem lies in the unsteady solution. I first iterate the case for 4000 steps using steady state, and make sure I reach convergence. Then I run it unsteady for 300 time steps using a time step about 1/50 the time step obtained from inverting the frequency obtained by using the definition of the Strouhal number (St = f*L/V). The problem is that the unsteady oscillations of the wake disappear and I get a symmetric wake once I get convergence. I wait until all the residuals level out and stay constant for a good while. Any ideas why this is happening? Is it because of the turbulence model (komega SST) being diffusive and getting rid of the vortex shedding, or is it the time step that is too large? And one more thing, how many pseudotime steps are required for each time step? Cheers, Ozgur 

December 5, 2010, 19:17 

#25 
New Member
Join Date: Mar 2010
Posts: 6
Rep Power: 7 
To reply to some of your questions, cfdcfd123, here are my opinions:
 I haven't done calculations with Re magnitudes around a miliion, but I have worked on Re values around 10000 for quite a while.  Fluent calculates TKE on its own; so you should be able to get TKE contour plots easily in postprocessing.  As for the plot (is it epsilon vs. k that you're looking for), you can take the average of any parameter you're interested in over the solutions obtained at different time steps. I take data at time steps placed really close to each other and take an arithmetic mean of the properties I'm looking at. Although, if you take an arithmetic mean with too large a time step between steps, this may be inaccurate.  I am modeling 3D cross flow past a circular cylinder placed on a flat wall, but I'm not focused on drag that much; but I'll let you know if something comes to mind. Best of luck, Ozgur 

December 14, 2010, 07:19 

#26 
New Member
Matt Hollins
Join Date: Nov 2010
Posts: 9
Rep Power: 7 
Ozgur_,
I am also currently running a similar problem at Re = 90,000. I am attempting to predict noise, and as such have based my time step on the max resolvable frequency that I desire to have. I picked 4000Hz (close to upper limit of human range) and then decided that for each wavelength I wanted to resolve it with 10 points. So delta t = 1/(4000 x 10) = 2.5E5 s I am not sure this will help you, but could be of some use in the future maybe. TO reduce computational time I am currently only running 1520 iterations per time step. I am relatively new to CFD and have been using the Komega SST model, however after reading this thread I am considering trying Spallart. Matt 

April 18, 2011, 02:37 

#27 
Senior Member
Ugly Kid Joe
Join Date: Aug 2010
Posts: 193
Rep Power: 7 
I am trying to simulate a flow over a cylinder at Re= 10000 and using Kw SST as the turbulence model. I dont know how to select the BC for turbulence ???? Can anyone help me out.
Thank you 

December 8, 2012, 06:08 
flow over a circular cylinder.. @ Re=1000

#28 
New Member
Rishitosh Ranjan
Join Date: Dec 2012
Location: Trichirapalli, Tamilnadu, India
Posts: 18
Rep Power: 5 
hiii..
i m very new to in CFD area... i have worked on flow over a single cylinder, two clinder in tandem and tow cylinder in side by side @ diff Re with diff ... i got good results.... but when i started working for Re=1000 .. for single cylinder... i didnt get gud result.... i have use used mesh wid y~min <1 to Y+max<5.. wid Kwsst model... Cd wich i m getting is 1.4 and as per journals and different thesis, i went through, Cd should be 1~ or < 1. i hv tried 0.5% turbulence @inlet and hydraulic dia equal to dia of cylinder.... and @ outlet 0, 0.5 and 0.7 back flow turbulence and same hyd dia... but still... 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
CFX Treatment of Laminar and Turbulent Flows  Jade M  CFX  10  December 11, 2015 20:52 
Flow over 2D Cylinder, Laminar and Turbulent  Tsr63  FLUENT  5  November 13, 2014 13:13 
Laminar flow or Turbulent flow  mech  FLUENT  0  January 27, 2007 19:51 
2D cylinder, external, laminar and turbulent flow  Dave  FLUENT  0  December 6, 2006 21:42 
modelling laminar and turbulent flow in the same  Ol  FLUENT  2  November 25, 2005 03:52 