What are the internal numerical data types used by FLUENT?

Leb · July 16, 2015, 07:18

Hi everyone,

I was wondering what is the default numerical data type FLUENT is using. In the manual I can find this:Lhttp://www.afs.enea.it/project/neptunius/docs/fluent/html/udf/node252.htm

but it's for UDFs.

The reason I'm asking is because I wonder if running a simulation on a mesh that has micro meter scale system expressed in meters (i.e. length scale etc. units are read as meters in FLUENT) would be worse/not converge as fast as one using micro meter scale in FLUENT (i.e. instead of having 0.000001m I would have 1um when FLUENT is calculating). I guess if FLUENT is using at least float precision, then this wouldn't be a problem...

Cheers

Kokemoor · July 16, 2015, 12:08

The default is single (float) precision, but double is supported and easy to enable. In the GUI launcher, there's a checkbox for it. If you're launching from command line, add "dp" to the solver flag, e.g. -3ddp instead of -3d.

Leb · July 21, 2015, 11:15

Thanks, I realised that after I have posted the question. I'm a bit surprised that it makes a difference in convergence ( I thought floats "dynamic range" would be more than enough...).

July 16, 2015, 07:18	What are the internal numerical data types used by FLUENT?	#1
Leb New Member V.Leb Join Date: Jun 2015 Posts: 8 Rep Power: 10	Hi everyone, I was wondering what is the default numerical data type FLUENT is using. In the manual I can find this:Lhttp://www.afs.enea.it/project/neptunius/docs/fluent/html/udf/node252.htm but it's for UDFs. The reason I'm asking is because I wonder if running a simulation on a mesh that has micro meter scale system expressed in meters (i.e. length scale etc. units are read as meters in FLUENT) would be worse/not converge as fast as one using micro meter scale in FLUENT (i.e. instead of having 0.000001m I would have 1um when FLUENT is calculating). I guess if FLUENT is using at least float precision, then this wouldn't be a problem... Cheers

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
high continuity Fluent modeling ABL with WRF/CALMET data	xitao21	FLUENT	1	May 2, 2019 10:20
OpenFoam to Fluent data conversion Problem	vemps	OpenFOAM	1	August 8, 2011 02:30
How to read tabulated data in Fluent	Siddhartha Pal	FLUENT	0	August 30, 2007 05:18
reading data into fluent	co2	FLUENT	0	December 28, 2003 01:31
Boundary Types in Fluent 5/6	Oresko	FLUENT	3	July 23, 2002 19:23

July 16, 2015, 12:08		#2
Kokemoor Senior Member Andrew Kokemoor Join Date: Aug 2013 Posts: 122 Rep Power: 13	The default is single (float) precision, but double is supported and easy to enable. In the GUI launcher, there's a checkbox for it. If you're launching from command line, add "dp" to the solver flag, e.g. -3ddp instead of -3d.

July 21, 2015, 11:15		#3
Leb New Member V.Leb Join Date: Jun 2015 Posts: 8 Rep Power: 10	Thanks, I realised that after I have posted the question. I'm a bit surprised that it makes a difference in convergence ( I thought floats "dynamic range" would be more than enough...).