Flow over a blunt body - SST turbulence model

bharath · January 28, 2016, 23:35

Hi to all
I am doing a CFD simulation of flow over a blunt body in CFX 16.1(like flow over a cylider).

I am using SST turbulence model and expecting Vortex shedding behind the blunt body.I ran both steady and unsteady runs but i am not getting the Vortex shedding in CFX. However in Fluent with the same setup and with SST turbulence model Vortex Shedding is happening behind blunt body in Steady state itself.

Now what controls i can try to get Vortex shedding in CFX?

Thanks in advance

ghorrocks · January 29, 2016, 04:53

Go through the general FAQ on accuracy first: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

To capture vortex shedding you will need a good quality mesh, an appropriate time step size (determined by a sensitivity study or some other valid method, not just guessed), good convergence and second order time differencing.

bharath · January 29, 2016, 07:03

Hi Glenn
Thanks for your reply.I under stood the things in the link.

But how Fluent predicted the Vortex well with the same mesh (7 million nodes with Boundary Layers) as well as same setup of CFX ?

ghorrocks · January 29, 2016, 07:13

When properly set up both solvers are capable of modelling this flow. You obviously have not set CFX up correctly yet.

Opaque · January 29, 2016, 08:01

Not sure what you mean by Vortex Shedding, but I am sure there is no way to predict a transient feature using a standard steady state solver.

You may be seeing a solution that is not converging to steady state, but oscillating between two possible solutions, but that is NOT vortex shedding by any means.

Be careful how you interpret your results (with any solver).

Hope the above helps,

turbo · January 29, 2016, 10:16

As you know, the vortex shedding itself is unsteady, and thus the unsteady RANS should be applied. At first, get a steady-state CFX solution, and then switch to the transient. If, even though you already did this, you cannot see the shedding pattern, you had better try RSM (Reynolds stress model) as a turbulence closure that will predict much better than two-equations. Or you can manipulate some factors in the two-equation, like the production limiter (in Advanced control setting) or the curvature scaling factor. There would be some gaps in Cd values between the steady and unsteady solutions, but both should show the large flow separation.

January 28, 2016, 23:35	Flow over a blunt body - SST turbulence model	#1
bharath Senior Member Bharath kumar Join Date: Apr 2009 Posts: 169 Rep Power: 18	Hi to all I am doing a CFD simulation of flow over a blunt body in CFX 16.1(like flow over a cylider). I am using SST turbulence model and expecting Vortex shedding behind the blunt body.I ran both steady and unsteady runs but i am not getting the Vortex shedding in CFX. However in Fluent with the same setup and with SST turbulence model Vortex Shedding is happening behind blunt body in Steady state itself. Now what controls i can try to get Vortex shedding in CFX? Thanks in advance

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Wrong flow in ratating domain problem	Sanyo	CFX	17	August 15, 2015 06:20
Low Mixing time Problem	Mavier	CFX	5	April 29, 2013 00:00
SST K- omega turbulence model	mb.pejvak	Main CFD Forum	8	September 16, 2011 08:52
Advice on turbulence model for complex internal flow	JPBodner	Main CFD Forum	1	September 3, 2011 03:07
Natural convection - Inlet boundary condition	max91	CFX	1	July 29, 2008 20:28

January 29, 2016, 04:53		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,933 Rep Power: 145	Go through the general FAQ on accuracy first: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F To capture vortex shedding you will need a good quality mesh, an appropriate time step size (determined by a sensitivity study or some other valid method, not just guessed), good convergence and second order time differencing.

January 29, 2016, 07:03		#3
bharath Senior Member Bharath kumar Join Date: Apr 2009 Posts: 169 Rep Power: 18	Hi Glenn Thanks for your reply.I under stood the things in the link. But how Fluent predicted the Vortex well with the same mesh (7 million nodes with Boundary Layers) as well as same setup of CFX ?

January 29, 2016, 07:13		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,933 Rep Power: 145	When properly set up both solvers are capable of modelling this flow. You obviously have not set CFX up correctly yet.

January 29, 2016, 08:01		#5
Opaque Senior Member Join Date: Jun 2009 Posts: 1,897 Rep Power: 33	Not sure what you mean by Vortex Shedding, but I am sure there is no way to predict a transient feature using a standard steady state solver. You may be seeing a solution that is not converging to steady state, but oscillating between two possible solutions, but that is NOT vortex shedding by any means. Be careful how you interpret your results (with any solver). Hope the above helps,

January 29, 2016, 10:16		#6
turbo Senior Member Join Date: Jun 2009 Posts: 174 Rep Power: 17	As you know, the vortex shedding itself is unsteady, and thus the unsteady RANS should be applied. At first, get a steady-state CFX solution, and then switch to the transient. If, even though you already did this, you cannot see the shedding pattern, you had better try RSM (Reynolds stress model) as a turbulence closure that will predict much better than two-equations. Or you can manipulate some factors in the two-equation, like the production limiter (in Advanced control setting) or the curvature scaling factor. There would be some gaps in Cd values between the steady and unsteady solutions, but both should show the large flow separation.