CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   LES : ERROR #004100018 (tried a few things) (https://www.cfd-online.com/Forums/cfx/25172-les-error-004100018-tried-few-things.html)

KK February 5, 2008 02:37

LES : ERROR #004100018 (tried a few things)
 
Hi,

I get the following error while simulating single phase flow in a cyclone with LES model. The initial condition is a solution from RSM turbulent model which had reached steady state.

ERROR #004100018 has occurred in subroutine FINMES. Message: Fatal overflow in linear solver."

If I change my boundary conditions from OPENING to OUTLET, solution proceed and solver does not fail. (Though temporary walls are raised in almost 100% boundary area.)

I searched all the the possible threads in this forum related to this problem and also read the manual for advice on flow modeling (convergence)

My doubts are

1) Why does it fail with OPENING type of boundary conditions while OPENING has addition degree of freedom as compared with OUTLETS ?

2) I changed my time step from 0.0005 sec to adaptive time steping with maximum inner loop iterations to be 2 and maximum 5. Time step update takes place at every 6th time step. But it still FAILED after a while.

3) The same mesh worked well with RANS (SSG-RSM). I know LES needs finer and better quality mesh. Could my mesh still be responsible for solver failure?

I would highly appreciate any comments,

Thanks,KK.


Glenn Horrocks February 5, 2008 16:43

Re: LES : ERROR #004100018 (tried a few things)
 
Hi,

I suspect your boundary is too close to the action so it is having problems converging. Have you tried moving your boundary further away?

Glenn Horrocks

KK February 5, 2008 17:28

Re: LES : ERROR #004100018 (tried a few things)
 
Hi Glenn,

I had already moved both the outlets (or openings, whatever) by 35 mm and 25 mm from cylindrical and conical regions, respectively. The total length of domain (before extending outlets) was 266 mm. Could you please suggest if I had extended by a sensible length ?

My geometry's image is posted at following link.

http://imgplace.com/image/view/44cd2...4897905ff1e80b

Thanks, KK

Glenn Horrocks February 6, 2008 16:25

Re: LES : ERROR #004100018 (tried a few things)
 
Hi,

Looks too close to me. I would try extending them both by quite a lot further.

Glenn Horrocks

KK February 7, 2008 16:35

Re: LES : ERROR #004100018 (tried a few things)
 
Thank your very much Glenn. I shall implement the suggestion.

KK.


All times are GMT -4. The time now is 00:16.