Run in fluent

Byukuya · March 12, 2010, 04:27

Hi people,

I have been investigating the flow around mounted bluff bodies and these are 3D problems. I want to ask that when i increase the mesh size of my problem, the capacity of a computer to solve it increases enourmusly, is it normal?
Let me explain it with example,
Cube - 500k mesh - need 2GB RAM
Cube - 2mil mesh - need more than 6B RAM (6GB very hardly and slow)

Well u can say that it is normal when u increase the amount of mesh size but i have friends doing CFD on more complex geometries and they can solve mil meshes in an average laptop. So i want to be guided that if im missing smth.

Thanks,

Byukuya

thecfduser · March 12, 2010, 05:40

Yes it is normal
(just try to see the number of faces on ure new mesh; it is much more than 4 times the number of faces on the old one

)
U cannot solve million meshes on average laptop... I have a very good laptop and with one million mesh, it is very slow...

Byukuya · March 12, 2010, 15:48

Firstly thank you for replying,

If you dont mind i would like to ask what kind of a problem are you trying to solve because i have some friends that working on combustion and they can solve mil meshes.

Byukuya

thecfduser · March 12, 2010, 21:42

In monophase flow, 1 million mesh is too much for me. It takes a lot of time, specially if ure flow is unsteady or complex ..... In multiphase it is even slower.. Me, i work on multiphase and unsteady flow...
Anyway, the time needed for calculation depends also on the mesh..If u are using a cartesion mesh, it will go faster than an tetrahedric mesh.
It depends also on the turbulence model (LES or RANS?), the number of equations to solve (with heat trnasfer? multiphase?) and on the number of iterations needed to converge. If they are modeling a very simple flow, it will be workable on a laptop

A small question: are u sure that u need 1 million mesh???? did u tried to see if a smaller mesh is enough???

Byukuya · March 13, 2010, 02:56

Well that's the most annoying part of it . When i use my case with everything installed in ( i mean turbulence model i boundary condition, equations etc) the run doesnt start even when the mesh size is 900k.

But i am trying to use laminar model for my mesh for the mesh dependence study to figure wether i need mil meshes. But it is taking too much time like 100 iteration 2 days. It didnt ended up yet so i'm not sure if i need mil mesh yet.

Byukuya

March 12, 2010, 04:27	Run in fluent	#1
Byukuya New Member Join Date: Mar 2010 Posts: 8 Rep Power: 16	Hi people, I have been investigating the flow around mounted bluff bodies and these are 3D problems. I want to ask that when i increase the mesh size of my problem, the capacity of a computer to solve it increases enourmusly, is it normal? Let me explain it with example, Cube - 500k mesh - need 2GB RAM Cube - 2mil mesh - need more than 6B RAM (6GB very hardly and slow) Well u can say that it is normal when u increase the amount of mesh size but i have friends doing CFD on more complex geometries and they can solve mil meshes in an average laptop. So i want to be guided that if im missing smth. Thanks, Byukuya

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
Working directory via command line	Luiz	CFX	4	March 6, 2011 20:02
Fluent 12 Beta Run Times	Allan Walsh	FLUENT	2	May 17, 2008 01:47
How to run fluent in terminal?	zjuADou	FLUENT	7	July 20, 2005 09:36
how to generate a long annulus then run in FLUENT	zhu chen	Main CFD Forum	0	March 27, 2004 18:42
Help: run FLUENT bechmark	SC	FLUENT	1	July 11, 2003 04:56

March 12, 2010, 05:40		#2
thecfduser Senior Member karine Join Date: Nov 2009 Posts: 158 Rep Power: 16	Yes it is normal (just try to see the number of faces on ure new mesh; it is much more than 4 times the number of faces on the old one ) U cannot solve million meshes on average laptop... I have a very good laptop and with one million mesh, it is very slow...

March 12, 2010, 15:48		#3
Byukuya New Member Join Date: Mar 2010 Posts: 8 Rep Power: 16	Firstly thank you for replying, If you dont mind i would like to ask what kind of a problem are you trying to solve because i have some friends that working on combustion and they can solve mil meshes. Byukuya

March 12, 2010, 21:42		#4
thecfduser Senior Member karine Join Date: Nov 2009 Posts: 158 Rep Power: 16	In monophase flow, 1 million mesh is too much for me. It takes a lot of time, specially if ure flow is unsteady or complex ..... In multiphase it is even slower.. Me, i work on multiphase and unsteady flow... Anyway, the time needed for calculation depends also on the mesh..If u are using a cartesion mesh, it will go faster than an tetrahedric mesh. It depends also on the turbulence model (LES or RANS?), the number of equations to solve (with heat trnasfer? multiphase?) and on the number of iterations needed to converge. If they are modeling a very simple flow, it will be workable on a laptop A small question: are u sure that u need 1 million mesh???? did u tried to see if a smaller mesh is enough???

March 13, 2010, 02:56		#5
Byukuya New Member Join Date: Mar 2010 Posts: 8 Rep Power: 16	Well that's the most annoying part of it . When i use my case with everything installed in ( i mean turbulence model i boundary condition, equations etc) the run doesnt start even when the mesh size is 900k. But i am trying to use laminar model for my mesh for the mesh dependence study to figure wether i need mil meshes. But it is taking too much time like 100 iteration 2 days. It didnt ended up yet so i'm not sure if i need mil mesh yet. Byukuya