Transiet Animation - vortex shedding

derz · May 4, 2010, 23:46

Hi guys,
I'm trying to get on top of transient simulations in CFX. I've decided to try and model flow past a cylinder - I want to get CFX to show me some vortex shedding. So I'm using the same parameters as this guy used: here is his video. Inlet turbulence is 5%.

He seems to get some nice vortex shedding.

Here is my video...

I'm using the exact same parameters as him - same dimensions too. Pretty much everything he displays at the start of the video, I used. The time step is 0.025 seconds and 15 seconds in total - that's 600 timesteps. also I know I should probably be using a laminar model due to low Re, but I used k-omega because the youtube guy did too. With the laminar I don't get any shedding either...

Seeming as he is also using CFX, I can't understand why I'm not getting any vortex shedding. In fact, my animation shows no changes in the wake at all.

I ran this as a 2D simulation. However I've also tried 3D with similar results. I'm getting CFX to output results after each timestep - then when the animation is finished, I use the keyframe animation movie maker to make the movie of the changing velocity contours.

The only thing that does not match this youtube's guys simulation is the mesh - could my mesh be the reason I'm not getting any shedding? The mesh isn't very fine in the region behind the cylinder - but around the cylinder I've used some pretty fine inflation (here is an image).

Just wondering if anyone had any ideas where I'm going wrong?

Thanks,

Derwin

mvoss · May 5, 2010, 05:04

hey,

did you even look at the resolution of th ICEM mesh?
Your mesh is way too coarse. Do some refinement in the vortex regions.

neewbie

derz · May 5, 2010, 05:55

hi...yeah i noticed he used ICEM mesh. I have workbench ansys 12.0. i don't think i have ICEM mesh. i tried making it less coarse but still same result.

i would like to give my mesh the same type of shape as the one ICEM gives, but not sure how to in the CFX Workbench environment. It only gives me patch conforming/patch independent options.

thanks for your response...i'll focus on improving the mesh for now.

ghorrocks · May 5, 2010, 07:40

Some issues I can see...

1) Make the top and bottom walls art of the inlet boundary.
2) Your mesh is probably too coarse.
3) Your mesh is probably low quality (can tell by the jaggedness of your contours).
4) A hex mesh carefully fitted to the cylinder will help a lot.
5) And dump the turbulence model, it just damps stuff out. Run it laminar.

derz · May 5, 2010, 08:02

Thanks ghorrocks. Yes I have started to run it in laminar. It seems my mesh is the issue - I will attempt to make it more like the meshes I've seen in many papers I've read - like this one but with a cylinder (sorry for the large picture).

derz · May 10, 2010, 10:21

following up - i've improved my mesh a great deal. and its made a huge difference. before it took like 15 seconds until i saw a little bit of movement in the wake...now its moving noticeably after 4 secs or so. thanks for everyone's help...

May 4, 2010, 23:46	Transiet Animation - vortex shedding	#1
derz Member Derwin Parkin Join Date: Feb 2010 Posts: 35 Rep Power: 16	Hi guys, I'm trying to get on top of transient simulations in CFX. I've decided to try and model flow past a cylinder - I want to get CFX to show me some vortex shedding. So I'm using the same parameters as this guy used: here is his video. Inlet turbulence is 5%. He seems to get some nice vortex shedding. Here is my video... I'm using the exact same parameters as him - same dimensions too. Pretty much everything he displays at the start of the video, I used. The time step is 0.025 seconds and 15 seconds in total - that's 600 timesteps. also I know I should probably be using a laminar model due to low Re, but I used k-omega because the youtube guy did too. With the laminar I don't get any shedding either... Seeming as he is also using CFX, I can't understand why I'm not getting any vortex shedding. In fact, my animation shows no changes in the wake at all. I ran this as a 2D simulation. However I've also tried 3D with similar results. I'm getting CFX to output results after each timestep - then when the animation is finished, I use the keyframe animation movie maker to make the movie of the changing velocity contours. The only thing that does not match this youtube's guys simulation is the mesh - could my mesh be the reason I'm not getting any shedding? The mesh isn't very fine in the region behind the cylinder - but around the cylinder I've used some pretty fine inflation (here is an image). Just wondering if anyone had any ideas where I'm going wrong? Thanks, Derwin

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K-Epsilon for Vortex Shedding	Sham	FLUENT	33	March 29, 2017 06:48
Vortex shedding, FSI-analysis, turbulence numerics	tallknuseren	CFX	3	May 10, 2010 05:31
Vortex shedding?	rbel038	CFX	4	April 27, 2008 20:57
Vortex shedding behind cylider in cross flow	Muthu	FLUENT	0	March 6, 2006 11:29
basic vortex shedding	john	Main CFD Forum	4	November 6, 2000 14:23

May 5, 2010, 05:04		#2
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	hey, did you even look at the resolution of th ICEM mesh? Your mesh is way too coarse. Do some refinement in the vortex regions. neewbie

May 5, 2010, 05:55		#3
derz Member Derwin Parkin Join Date: Feb 2010 Posts: 35 Rep Power: 16	hi...yeah i noticed he used ICEM mesh. I have workbench ansys 12.0. i don't think i have ICEM mesh. i tried making it less coarse but still same result. i would like to give my mesh the same type of shape as the one ICEM gives, but not sure how to in the CFX Workbench environment. It only gives me patch conforming/patch independent options. thanks for your response...i'll focus on improving the mesh for now.

May 5, 2010, 07:40		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	Some issues I can see... 1) Make the top and bottom walls art of the inlet boundary. 2) Your mesh is probably too coarse. 3) Your mesh is probably low quality (can tell by the jaggedness of your contours). 4) A hex mesh carefully fitted to the cylinder will help a lot. 5) And dump the turbulence model, it just damps stuff out. Run it laminar.

May 5, 2010, 08:02		#5
derz Member Derwin Parkin Join Date: Feb 2010 Posts: 35 Rep Power: 16	Thanks ghorrocks. Yes I have started to run it in laminar. It seems my mesh is the issue - I will attempt to make it more like the meshes I've seen in many papers I've read - like this one but with a cylinder (sorry for the large picture).

May 10, 2010, 10:21		#6
derz Member Derwin Parkin Join Date: Feb 2010 Posts: 35 Rep Power: 16	following up - i've improved my mesh a great deal. and its made a huge difference. before it took like 15 seconds until i saw a little bit of movement in the wake...now its moving noticeably after 4 secs or so. thanks for everyone's help...