mass flow rate and size of mesh

zandi · February 13, 2010, 15:33

Hi every body
I'm working on a model of spillway in cfx (free surface flow modeling). when I change the mesh size espacialy near wall bc type (bottom of domain), the mass flow rate passing the domain change. could anyone explain this for me that what and why have happened?!
thanks

ghorrocks · February 14, 2010, 03:59

Good to see you are starting a mesh sensitivity check - many people do not bother with this (or don't know it is important) and their results are rubbish because of the laziness.

When your results are changing between mesh refinements it means your mesh is not adequately refined. You need to use a finer mesh.

The seminal reference for this analysis is "Computational Fluid Dynamics" by Roache. I strongly recommend you read it.

zandi · February 14, 2010, 06:16

thank you
in the mesh study hat I have done the water surface, pressure, velocity and turbulent kinetic energy respectively reached to a stable point. But the mass flow rate did not. refining mesh make number of my elements more than 6,000,000 (my study started from 50000).the i7 system that I work on it, does runs with 6000000 elements in about 6 days with trouble. it is difficult to continuing mesh refinment.

ghorrocks · February 14, 2010, 17:37

CFD was never meant to be easy. Besides simulating nuclear explosions, one of the main reasons people build supercomputers is to run CFD.

feizaghaee · February 15, 2010, 16:09

i have had this problem i work with free surface flow but i could finally solve it. you could set lower physical time scale for continuity equation in solver control rather than general time scale you should try several times to find a effective cobimation of several tme scales for good result. setting the lower tme scale for volume fraction helps better covergence and it's very effective.

zandi · February 16, 2010, 02:33

realy thank you mr Feizaghaee
I'm going to test it

ghorrocks · February 16, 2010, 17:32

Sounds like beating about the bush to me. Either way you need to do a mesh sensitivity check and if that shows your computers are not big enough to solve your problem then you need bigger computers.

zandi · February 17, 2010, 06:10

thanks a lot friends i'm going to test time scale and other way I can do. still any more suggestion would be greatly appreciated

zandi · February 20, 2010, 15:36

Quote:

Originally Posted by feizaghaee

i have had this problem i work with free surface flow but i could finally solve it. you could set lower physical time scale for continuity equation in solver control rather than general time scale you should try several times to find a effective cobimation of several tme scales for good result. setting the lower tme scale for volume fraction helps better covergence and it's very effective.

thank you it works for discharge
but not for convergence
could help me please

feizaghaee · February 21, 2010, 04:31

if you work with cfx 12 go to the multiphase control in sovler control>advance tab check the box of multiphase control and select coupled mode. in sovler control>advance tab select hydro control and set the number of transient itration to for example 5.
can you tell me what time scales do you use for each equation and what is the total time scale?

February 13, 2010, 15:33	mass flow rate and size of mesh	#1
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	Hi every body I'm working on a model of spillway in cfx (free surface flow modeling). when I change the mesh size espacialy near wall bc type (bottom of domain), the mass flow rate passing the domain change. could anyone explain this for me that what and why have happened?! thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Gambit Unstructed (Tet) Mesh - size control	newbie	FLUENT	3	August 26, 2008 11:38
size function growth rate	Dong Fu	FLUENT	1	February 12, 2008 01:02
Mass flow of discrete phase at a point	dan	FLUENT	0	November 2, 2005 10:34
What's the point of imprting an stl mesh?	philip	FLUENT	8	July 6, 2005 16:51
liquid particle track and flow rate	Tim	FLUENT	1	November 28, 2001 21:41

February 14, 2010, 03:59		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Good to see you are starting a mesh sensitivity check - many people do not bother with this (or don't know it is important) and their results are rubbish because of the laziness. When your results are changing between mesh refinements it means your mesh is not adequately refined. You need to use a finer mesh. The seminal reference for this analysis is "Computational Fluid Dynamics" by Roache. I strongly recommend you read it.

February 14, 2010, 06:16		#3
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	thank you in the mesh study hat I have done the water surface, pressure, velocity and turbulent kinetic energy respectively reached to a stable point. But the mass flow rate did not. refining mesh make number of my elements more than 6,000,000 (my study started from 50000).the i7 system that I work on it, does runs with 6000000 elements in about 6 days with trouble. it is difficult to continuing mesh refinment.

February 14, 2010, 17:37		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	CFD was never meant to be easy. Besides simulating nuclear explosions, one of the main reasons people build supercomputers is to run CFD.

February 15, 2010, 16:09		#5
feizaghaee Senior Member moein Join Date: Dec 2009 Posts: 132 Rep Power: 16	i have had this problem i work with free surface flow but i could finally solve it. you could set lower physical time scale for continuity equation in solver control rather than general time scale you should try several times to find a effective cobimation of several tme scales for good result. setting the lower tme scale for volume fraction helps better covergence and it's very effective.

February 16, 2010, 02:33		#6
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	realy thank you mr Feizaghaee I'm going to test it

February 16, 2010, 17:32		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Sounds like beating about the bush to me. Either way you need to do a mesh sensitivity check and if that shows your computers are not big enough to solve your problem then you need bigger computers.

February 17, 2010, 06:10		#8
zandi Senior Member Fatema Zandi Goharrizi Join Date: Mar 2009 Posts: 158 Rep Power: 17	thanks a lot friends i'm going to test time scale and other way I can do. still any more suggestion would be greatly appreciated

February 21, 2010, 04:31		#10
feizaghaee Senior Member moein Join Date: Dec 2009 Posts: 132 Rep Power: 16	if you work with cfx 12 go to the multiphase control in sovler control>advance tab check the box of multiphase control and select coupled mode. in sovler control>advance tab select hydro control and set the number of transient itration to for example 5. can you tell me what time scales do you use for each equation and what is the total time scale?