Grid independence check in transient simulation

abhi084 · December 4, 2014, 02:50

hi everyone,

i am new to CFD and this site. i wanted to know how to check the grid independence when you have to do transient simulation.

whether we check grid independence in steady simulation only and then use the same in the transient case. Please help.

Thanks for the help in advance.

sbaffini · December 4, 2014, 16:55

Well, it is quite simple indeed. For steady problems you compute the error at fixed space locations only. For unsteady ones you should compute the error for fixed space locations at a fixed time.

Then, suddenly follows another question: how to reach the the fixed time if you changed your grid? Is correct to keep the same time step size?

As a matter of fact, if you are dealing with general transport equations (diffusion + convection), your dimensionless numbers u*dt/dx (Courant number) and k*dt/dx^2 (with k diffusion coefficient) can't both stay the same. So, usually, you choose one of them to be constant and let the other one vary (the one supposed to be less relevant). More simply, you can just keep a constant dt. In general, however, you should check that, while grid refining, you are not going in regions in the space of these parameters such that your scheme is not anymore valid.

December 4, 2014, 02:50	Grid independence check in transient simulation	#1
abhi084 New Member Abhi Join Date: Dec 2014 Posts: 28 Rep Power: 12	hi everyone, i am new to CFD and this site. i wanted to know how to check the grid independence when you have to do transient simulation. whether we check grid independence in steady simulation only and then use the same in the transient case. Please help. Thanks for the help in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PBC breaks down, is there a grid check tool?	gerritgroot	SU2	8	June 13, 2013 19:24
Grid check failed .... run simulation useless?	Zigainer	FLUENT	2	May 17, 2012 05:31
OF 1.6 \| Ubuntu 9.10 (64bit) \| GLIBCXX_3.4.11 not found	piprus	OpenFOAM Installation	22	February 25, 2010 14:43
Problems in compiling paraview in Suse 10.3 platform	chiven	OpenFOAM Installation	3	December 1, 2009 08:21
how to identify transient simulation converged	littlelz	CFX	5	January 27, 2009 19:13

December 4, 2014, 16:55		#2
sbaffini Senior Member Paolo Lampitella Join Date: Mar 2009 Location: Italy Posts: 2,195 Blog Entries: 29 Rep Power: 39	Well, it is quite simple indeed. For steady problems you compute the error at fixed space locations only. For unsteady ones you should compute the error for fixed space locations at a fixed time. Then, suddenly follows another question: how to reach the the fixed time if you changed your grid? Is correct to keep the same time step size? As a matter of fact, if you are dealing with general transport equations (diffusion + convection), your dimensionless numbers udt/dx (Courant number) and kdt/dx^2 (with k diffusion coefficient) can't both stay the same. So, usually, you choose one of them to be constant and let the other one vary (the one supposed to be less relevant). More simply, you can just keep a constant dt. In general, however, you should check that, while grid refining, you are not going in regions in the space of these parameters such that your scheme is not anymore valid.