grid dependency

sri31049 · July 9, 2009, 13:38

hi all

any one can please help me in my problem.

i am doing the simulation of flow past circular cylinder at Re = 100. i have done the grid independence study and compared the flow parameter values such as drag, strouhal number and rms lift. Those are comparable for all grid resolutions.
Now i kept some sensors along the wake center line in the flow domain for sensing the cross stream velocity at that sensor points. Based on that i calculate variance and plot it. surprisingly those plots are not similar in both pattern and values.

Here i have a doubt, even tough i kept the sensors at same points for all the grids and performed the grid independence study, why i am getting different variance values (means cross stream velocity values) for different grids.

what i thought about grid independence study is that flow variables, wake behavior and flow patterns are same for all the grid resolutions. is it false?

anyone can please help me in this problem. Thanks in advance

wadebogs · July 26, 2009, 17:55

As your grid gets finer, if the solution is steady, your variables should stop changing at some point and you can say that the solution is no longer changing with the grid density and use that grid for calculations.

However, if you have an unsteady case then you will need to look at time-averaged quantities such as pressure etc. Re=100 cylinder will give you vortex shedding and as such means that solution varies with time. You have to model this "unsteady" or time-dependent and time average your post-processing. If you don't do this you will certainly see different values for your monitors.

Type "re=100 cylinder" into google and see the movies which come up on youtube. That will given you an idea of what is going on!

nvtrieu · March 3, 2010, 02:47

Hello everyone,

I'm doing simulation on circular cylinder too. For result analysing and in oder to compare with previous studies, I want to determine the Strouhal number (which defined as St=f*d/u, where d: referent length, u: velocity and f: frequency of vortex shedding). But I still confuse with the "f" parameter, how can we determine the value of "f" ?
Is there any one can tell me how to estimate the Strouhal number?
Thanks alot!

sri31049 · March 3, 2010, 08:12

We need to calculate the frequency from the lift history. First you choose some no of cycles and find out the total time for all the cycles from that you can directly determine the frequency as no of cycles/total time. good luck

July 9, 2009, 13:38	grid dependency	#1
sri31049 New Member paaresh Join Date: Mar 2009 Posts: 19 Rep Power: 17	hi all any one can please help me in my problem. i am doing the simulation of flow past circular cylinder at Re = 100. i have done the grid independence study and compared the flow parameter values such as drag, strouhal number and rms lift. Those are comparable for all grid resolutions. Now i kept some sensors along the wake center line in the flow domain for sensing the cross stream velocity at that sensor points. Based on that i calculate variance and plot it. surprisingly those plots are not similar in both pattern and values. Here i have a doubt, even tough i kept the sensors at same points for all the grids and performed the grid independence study, why i am getting different variance values (means cross stream velocity values) for different grids. what i thought about grid independence study is that flow variables, wake behavior and flow patterns are same for all the grid resolutions. is it false? anyone can please help me in this problem. Thanks in advance

March 3, 2010, 08:12	hai	#4
sri31049 New Member paaresh Join Date: Mar 2009 Posts: 19 Rep Power: 17	We need to calculate the frequency from the lift history. First you choose some no of cycles and find out the total time for all the cycles from that you can directly determine the frequency as no of cycles/total time. good luck

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
OpenFOAM on MinGW crosscompiler hosted on Linux	allenzhao	OpenFOAM Installation	127	January 30, 2009 20:08
DxFoam reader update	hjasak	OpenFOAM Post-Processing	69	April 24, 2008 02:24
DecomposePar links against liblamso0 with OpenMPI	jens_klostermann	OpenFOAM Bugs	11	June 28, 2007 18:51
Combustion Convergence problems	Art Stretton	Phoenics	5	April 2, 2002 06:59
Troubles modelling flow through a grid	Hans Klaufus	CFX	1	June 28, 2000 17:43

July 26, 2009, 17:55		#2
wadebogs New Member Join Date: Jul 2009 Posts: 16 Rep Power: 17	As your grid gets finer, if the solution is steady, your variables should stop changing at some point and you can say that the solution is no longer changing with the grid density and use that grid for calculations. However, if you have an unsteady case then you will need to look at time-averaged quantities such as pressure etc. Re=100 cylinder will give you vortex shedding and as such means that solution varies with time. You have to model this "unsteady" or time-dependent and time average your post-processing. If you don't do this you will certainly see different values for your monitors. Type "re=100 cylinder" into google and see the movies which come up on youtube. That will given you an idea of what is going on!

March 3, 2010, 02:47		#3
nvtrieu Member Join Date: Jul 2009 Posts: 70 Rep Power: 17	Hello everyone, I'm doing simulation on circular cylinder too. For result analysing and in oder to compare with previous studies, I want to determine the Strouhal number (which defined as St=f*d/u, where d: referent length, u: velocity and f: frequency of vortex shedding). But I still confuse with the "f" parameter, how can we determine the value of "f" ? Is there any one can tell me how to estimate the Strouhal number? Thanks alot!