Grid Independency Study

pchoopanya · August 10, 2011, 05:55

Hey guys,

Hope people those are living in the UK are still able to run their simulations safely in their workplaces

while the riots are happening all over the island. I am in Brighton and luckily there is no sign of these nonsense anti-social behaviour, not yet!

I am a newbie to CFD and Fluent. My research lies on CFD anyway, I am trying to run some demo simulations and grid independency study.

I use a simple hexahedral cell (a cube) and reduce the size until the results do not change with the further reduction of the cell's size - am I doign this correctly?

Let's say, if I have eventually come to the final minimum size that the results do not further change, with the exactly the same geometry, can I use this final cell size as a reference and then run other simulations but with different operating conditions (parametric study?) Or do I need to find the minimum cell size at every single simulation when the operating conditions are changed?

Sorry for my poor English if you cannot understand my words...

Cheers,

Pattarapong

Tongji Environment · August 11, 2011, 00:09

Hi, in my opinion, there is a most appropriate grid density for one geometric model. Before the simulation, we should do grid independence research because the quality of grid can influence convergence speed and the computation precision.(the time consumpution of pre-process should take up 40% of the whole simulation)
You can create several models of different grid numbers with a certain proportion. Then go to caculate them respectively with the same condition setting. We can draw a curve with convergence iterations VS. grid density, choose the best one with the comprehensive consideration of accuracy and time requirement.
Once you choose a best grid density for a geometric model, you can use this grid to go to run other function simulations unless your geometric model will be changed.
May be useful for you!

August 10, 2011, 05:55	Grid Independency Study	#1
pchoopanya Member Join Date: Aug 2011 Posts: 33 Rep Power: 15	Hey guys, Hope people those are living in the UK are still able to run their simulations safely in their workplaces while the riots are happening all over the island. I am in Brighton and luckily there is no sign of these nonsense anti-social behaviour, not yet! I am a newbie to CFD and Fluent. My research lies on CFD anyway, I am trying to run some demo simulations and grid independency study. I use a simple hexahedral cell (a cube) and reduce the size until the results do not change with the further reduction of the cell's size - am I doign this correctly? Let's say, if I have eventually come to the final minimum size that the results do not further change, with the exactly the same geometry, can I use this final cell size as a reference and then run other simulations but with different operating conditions (parametric study?) Or do I need to find the minimum cell size at every single simulation when the operating conditions are changed? Sorry for my poor English if you cannot understand my words... Cheers, Pattarapong

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to do a grid independent study	xyq102296	ANSYS	16	November 11, 2021 09:05
Grid Independent Study for a car	John222	Main CFD Forum	1	January 26, 2011 20:05
About grid sensibilty study	Fabio	Main CFD Forum	0	October 15, 2009 12:23
Vorticity plot in grid refinement study	CH	Main CFD Forum	3	May 22, 2007 13:35
Combustion Convergence problems	Art Stretton	Phoenics	5	April 2, 2002 06:59

August 11, 2011, 00:09		#2
Tongji Environment New Member Calvin Join Date: Aug 2011 Posts: 4 Rep Power: 15	Hi, in my opinion, there is a most appropriate grid density for one geometric model. Before the simulation, we should do grid independence research because the quality of grid can influence convergence speed and the computation precision.(the time consumpution of pre-process should take up 40% of the whole simulation) You can create several models of different grid numbers with a certain proportion. Then go to caculate them respectively with the same condition setting. We can draw a curve with convergence iterations VS. grid density, choose the best one with the comprehensive consideration of accuracy and time requirement. Once you choose a best grid density for a geometric model, you can use this grid to go to run other function simulations unless your geometric model will be changed. May be useful for you!