Switching to implicit scheme after explicit startup

Obad · November 20, 2016, 06:31

Hi guys,

I'm currently doing CFD simulations of supersonic flow.

My calculations are not time accurate since I'm trying to converge to a steady state. However, due to stability reasons I do the startup with an explicit scheme with very low CFL numbers (ramping up from 0.01). Usually I let the calculation run with a second order explicit scheme until it reaches a reasonably steady state.

Now I would like to refine my solution with a second order implicit scheme running at higher CFL, but no matter how low I go with the CFL number (CFL < 0.001) the solver crashes pretty much immediately.

Can anyone enlighten me please

agd · November 21, 2016, 11:30

There is no reason why a switch should not work, since all you are doing is feeding the implicit solver a different set of initial conditions compared to starting from freestream. I suspect you have a bug in your code, probably in some variable that needs an initial value that is not being set right. For example, if you are running second-order time, are you providing sufficient time levels in the restart? Are your boundaries being set to the correct restart values? Your description of your problem sounds like a code error.

November 20, 2016, 06:31	Switching to implicit scheme after explicit startup	#1
Obad Member Obad Join Date: Sep 2013 Posts: 42 Rep Power: 13	Hi guys, I'm currently doing CFD simulations of supersonic flow. My calculations are not time accurate since I'm trying to converge to a steady state. However, due to stability reasons I do the startup with an explicit scheme with very low CFL numbers (ramping up from 0.01). Usually I let the calculation run with a second order explicit scheme until it reaches a reasonably steady state. Now I would like to refine my solution with a second order implicit scheme running at higher CFL, but no matter how low I go with the CFL number (CFL < 0.001) the solver crashes pretty much immediately. Can anyone enlighten me please

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Newbie question about explicit and implicit methods	Yuby	Main CFD Forum	10	May 29, 2015 14:16
Explicit vs Implicit VOF in Fluent	MachZero	FLUENT	5	August 7, 2013 19:17
Is it feasible to apply implicit scheme in LES	EagleGao	Main CFD Forum	2	October 24, 2011 15:04
Euler (explicit or implicit)	anybody	Main CFD Forum	2	May 8, 2006 03:12
implicit vs explicit	pXYZ	Main CFD Forum	2	April 21, 2006 10:48

November 21, 2016, 11:30		#2
agd Senior Member Join Date: Jul 2009 Posts: 358 Rep Power: 19	There is no reason why a switch should not work, since all you are doing is feeding the implicit solver a different set of initial conditions compared to starting from freestream. I suspect you have a bug in your code, probably in some variable that needs an initial value that is not being set right. For example, if you are running second-order time, are you providing sufficient time levels in the restart? Are your boundaries being set to the correct restart values? Your description of your problem sounds like a code error.