LES Simulation Basic Concept

AS_Aero · February 21, 2017, 18:09

Hi All

I am doing a simple in-compressible flow (water as medium) inside a twisted pipe/cylinder. Here I am doing an LES computation using CFX and for that I am confused on what basis I should give my mesh size and time step. And can use Hex mesh or Tetrahedrals ?

The flow inlet is 5 bar and mass flow rate is 10 l/min
The geometry is total lenght of 30 mm, Inlet diameter of the pipe is 7mm
after 20mm distance the pipe is twisted and a difuser is connected at the end to act like a nozzle. Based on the flow conditions we can say the Reynolds number is around 10000

So in this case I am confused how to decide the mesh size and the time step. Kindly give me some advice please.

Thanks

ghorrocks · February 21, 2017, 23:57

Hex or tets: Either can work, but hexas have less dissipation providing you can make them high quality. Dissipation is very important in LES models.

For LES you should set your mesh size as a function of the Taylor length scale of the turbulence. With a low Re flow that will hopefully be fairly large which will help.

AS_Aero · February 22, 2017, 04:12

Hi Glenn

Thanks a lot for the information.

But as I said my Re number at the inlet is around 10000 but there is a diffuser at the outlet side where the velocity is higher compared to the inlet pipe. So in this case the Re will be higher at that region compared to the inlet. But out of all these how can we decide that this is what the minimum (dx) should be : How we can calculate and then when it comes to (dt) based on CFL condition should the ratio be less than or equal to 0.5 ?
Then I am facing an issue with the HEX mesh with multizone in Ansys mesh: The mesh doesnt generate an inflation layer and the mesh is clumsy in the central axis along the flow (I can see that when I do the cut sectional view) !!

ghorrocks · February 22, 2017, 06:53

Estimate the Taylor length scale from the turbulence conditions. That is a good estimate for your mesh size. Then do a mesh sensitivity study to check it.

Rather than using a CFL based time step I recommend you use adaptive time stepping homing in on 3-5 coefficient loops per iteration.

Try the geometry and meshing forum for the mesh questions.

AS_Aero · March 8, 2017, 04:04

Hi Glenn

Thanks for your message. But as you mentioned to calculate the Taylor Length Scales from Turbulence quantities, could you elaborate the process I could do it ? I am still not sure how to calculate it. And I am doing LES directly and not based on my RANS or URANS. Thanks in advance.

ghorrocks · March 8, 2017, 17:49

A turbulence textbook like "Turbulence Modelling for CFD" by Wilcox is essential reading before contemplating LES work. It will show the derivations of Taylor length scales and all the other parameters you need.

You need an initial estimate of the turbulence conditions so you know the turbulence properties so you can design the simulation properly to do the simulation. So you should do an initial simulation, preferably URANS, to get an estimate of the turbulence properties, which you can then use to set up your LES simulation.

February 21, 2017, 18:09	LES Simulation Basic Concept	#1
AS_Aero Senior Member Join Date: Mar 2013 Location: Germany Posts: 357 Rep Power: 15	Hi All I am doing a simple in-compressible flow (water as medium) inside a twisted pipe/cylinder. Here I am doing an LES computation using CFX and for that I am confused on what basis I should give my mesh size and time step. And can use Hex mesh or Tetrahedrals ? The flow inlet is 5 bar and mass flow rate is 10 l/min The geometry is total lenght of 30 mm, Inlet diameter of the pipe is 7mm after 20mm distance the pipe is twisted and a difuser is connected at the end to act like a nozzle. Based on the flow conditions we can say the Reynolds number is around 10000 So in this case I am confused how to decide the mesh size and the time step. Kindly give me some advice please. Thanks

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
How to make grid for LES simulation	dinhanh	Main CFD Forum	3	November 11, 2015 03:37
Les transient simulation (loading)	juliom	Tecplot	6	August 31, 2015 15:45
2D LES simulation of turbulent jet	adnan.nweilati	FLUENT	6	December 3, 2014 10:33
LES Simulation of 5-to-1 rectangular prism	tung.nguyen	OpenFOAM	1	July 2, 2014 05:56
LES for multiphase simulation	Kushagra	CFX	0	April 24, 2008 20:51

February 21, 2017, 23:57		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 18,010 Rep Power: 146	Hex or tets: Either can work, but hexas have less dissipation providing you can make them high quality. Dissipation is very important in LES models. For LES you should set your mesh size as a function of the Taylor length scale of the turbulence. With a low Re flow that will hopefully be fairly large which will help.

February 22, 2017, 04:12		#3
AS_Aero Senior Member Join Date: Mar 2013 Location: Germany Posts: 357 Rep Power: 15	Hi Glenn Thanks a lot for the information. But as I said my Re number at the inlet is around 10000 but there is a diffuser at the outlet side where the velocity is higher compared to the inlet pipe. So in this case the Re will be higher at that region compared to the inlet. But out of all these how can we decide that this is what the minimum (dx) should be : How we can calculate and then when it comes to (dt) based on CFL condition should the ratio be less than or equal to 0.5 ? Then I am facing an issue with the HEX mesh with multizone in Ansys mesh: The mesh doesnt generate an inflation layer and the mesh is clumsy in the central axis along the flow (I can see that when I do the cut sectional view) !!

February 22, 2017, 06:53		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 18,010 Rep Power: 146	Estimate the Taylor length scale from the turbulence conditions. That is a good estimate for your mesh size. Then do a mesh sensitivity study to check it. Rather than using a CFL based time step I recommend you use adaptive time stepping homing in on 3-5 coefficient loops per iteration. Try the geometry and meshing forum for the mesh questions.

March 8, 2017, 04:04		#5
AS_Aero Senior Member Join Date: Mar 2013 Location: Germany Posts: 357 Rep Power: 15	Hi Glenn Thanks for your message. But as you mentioned to calculate the Taylor Length Scales from Turbulence quantities, could you elaborate the process I could do it ? I am still not sure how to calculate it. And I am doing LES directly and not based on my RANS or URANS. Thanks in advance.

March 8, 2017, 17:49		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 18,010 Rep Power: 146	A turbulence textbook like "Turbulence Modelling for CFD" by Wilcox is essential reading before contemplating LES work. It will show the derivations of Taylor length scales and all the other parameters you need. You need an initial estimate of the turbulence conditions so you know the turbulence properties so you can design the simulation properly to do the simulation. So you should do an initial simulation, preferably URANS, to get an estimate of the turbulence properties, which you can then use to set up your LES simulation.