What's the best order to run this simulation in?

siw · November 4, 2009, 05:24

Hi,

I'm planning to run a CFX simulation of a body that is free-falling which means that I'll use the ICEM replay remeshing and 6dof solver features.

But before I do this free-fall run I need to make sure my mesh and transient timesteps are okay for the stationary object before it starts to free fall. I've done one steady-state run with a mesh of "best guess" element sizing and residuals suggested the flowfield was transient. So I ran a transient simulation with a "best guess" timestep size and total time with the steady-state for initialisation.

So my question is this. Is it better to first keep the same time data the same and run the transient simulation on some finer resolution meshes. Or would it be better to first keep the same mesh and refine the timestep size? So should mesh or timestep independence be worked towards first?

It sounds from this that quite a few simulations must be conducted in order to get the single best setup for the run with the free-falling object, which'll have it's own set is assessments (i.e. remeshing criteria, numerics and RANS turbulence models)

Thanks

ghorrocks · November 4, 2009, 19:42

In stationary mesh simulations it is normally reasonably easy to get full convergence, a bit more work to get timestep convergence and difficult to get mesh convergence. In your case as a 6DOF thing it will be even more challenging again.

So normally I do timestep refinement first as it is easier and usually the timestep is OK for the other mesh sizes as well. But in you case that may not apply, can only just try it and find out.

November 4, 2009, 05:24	What's the best order to run this simulation in?	#1
siw Senior Member Stuart Join Date: Jul 2009 Location: Portsmouth, England Posts: 733 Rep Power: 25	Hi, I'm planning to run a CFX simulation of a body that is free-falling which means that I'll use the ICEM replay remeshing and 6dof solver features. But before I do this free-fall run I need to make sure my mesh and transient timesteps are okay for the stationary object before it starts to free fall. I've done one steady-state run with a mesh of "best guess" element sizing and residuals suggested the flowfield was transient. So I ran a transient simulation with a "best guess" timestep size and total time with the steady-state for initialisation. So my question is this. Is it better to first keep the same time data the same and run the transient simulation on some finer resolution meshes. Or would it be better to first keep the same mesh and refine the timestep size? So should mesh or timestep independence be worked towards first? It sounds from this that quite a few simulations must be conducted in order to get the single best setup for the run with the free-falling object, which'll have it's own set is assessments (i.e. remeshing criteria, numerics and RANS turbulence models) Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
1st order vs 2nd order	ken	FLUENT	8	March 14, 2013 03:43
Fluent Remote Simulation Facility Service (RSF) di	Rami	FLUENT	2	June 4, 2008 05:38
Simpliest way to run airfoil simulation diff AOA	Ben	FLUENT	0	November 19, 2006 02:11
Procedure to run unsteady simulation?	STN	Main CFD Forum	2	February 16, 2002 04:37
3-D Contaminant Dispersal Simulation	Apple L S Chan	Main CFD Forum	1	December 23, 1998 10:06

November 4, 2009, 19:42		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	In stationary mesh simulations it is normally reasonably easy to get full convergence, a bit more work to get timestep convergence and difficult to get mesh convergence. In your case as a 6DOF thing it will be even more challenging again. So normally I do timestep refinement first as it is easier and usually the timestep is OK for the other mesh sizes as well. But in you case that may not apply, can only just try it and find out.