Slow Convergence/No convergence of Continuity

ritwikdas18 · June 18, 2015, 03:14

Hello everyone,
I am facing a problem in 3 phase , transient, pressure based, implicit,2-D flow in Fluent. The problem is no convergence with the CONTINUITY RESIDUAL even when the timestep size is as small as 0.0005 sec.The continuity residuals do not diverge but rather oscillate about the same point with the above time step(like from 2e-3 to 3e-3). This has caused me to use a maximum timestep of 0.0003 sec.
Therefore around 48 hours of computational time is taken up even for a flow of 10 sec.

1.My mesh orthogonality is 0.41. Maximum aspect ratio is 6.16.The mesh is very very fine.

2.Under-relaxation factors.
Pressure- 0.9
Momentum-0.05
Rest all default

3.With a timestep of 0.0003 sec 1 timestep = ~2.5 sec

4.Schemes used:- Pressure velocity segregated
First Order Upwind for all and Least Squares Cell Based Gradient

5.The problem is as follows:-
Bed of solid particles(Density=2500kg/m3) in a tank with nitric acid. Air is sprayed from a nozzle to suspend the solid particles. Therefore the above flow is to be modelled.

6.I will be heavily grateful for help on the continuity residual. A lot of time is spent in computation because of small time step.

CeesH · June 18, 2015, 12:07

what is convergence? Are your residuals relative or absolute?
As can be read in the FLUENT manual on judging convergence - it may depend very much on the situation what is good enough. I'd like to take it one step further, and say: never, ever, ever judge your convergence solely on your residuals.

Check if integral parameters of interest reach a steady value. Depending on your problem, these may be Nusselt number, drag coefficient, power number, mean velocity, etcetera. Do you see that these reach a constant value within say, < 1%, or do you see oscillations in these numbers? That's a much better guide of convergence than the residual.

ritwikdas18 · June 19, 2015, 05:25

1.My residuals are absolute.

2.I at this point of time don't have much idea of what to monitor in my problems since for me the major concern is volume fraction of solid(to check dispersion of solids) which is already present in the residual.

3.Also Fluent says to use a time step which takes 20 iterations / step to converge. However when I use a time step size of 1e-4 then it converges in 2-3 iterations. I think this is normal. However when I use a time-step size of 1e-3 the continuity residual keeps oscillating and hence it never converges not even in 100 iterations.Is the above normal?

Thanks for your time.

June 18, 2015, 03:14	Slow Convergence/No convergence of Continuity	#1
ritwikdas18 New Member Ritwik Das Join Date: May 2015 Posts: 4 Rep Power: 11	Hello everyone, I am facing a problem in 3 phase , transient, pressure based, implicit,2-D flow in Fluent. The problem is no convergence with the CONTINUITY RESIDUAL even when the timestep size is as small as 0.0005 sec.The continuity residuals do not diverge but rather oscillate about the same point with the above time step(like from 2e-3 to 3e-3). This has caused me to use a maximum timestep of 0.0003 sec. Therefore around 48 hours of computational time is taken up even for a flow of 10 sec. 1.My mesh orthogonality is 0.41. Maximum aspect ratio is 6.16.The mesh is very very fine. 2.Under-relaxation factors. Pressure- 0.9 Momentum-0.05 Rest all default 3.With a timestep of 0.0003 sec 1 timestep = ~2.5 sec 4.Schemes used:- Pressure velocity segregated First Order Upwind for all and Least Squares Cell Based Gradient 5.The problem is as follows:- Bed of solid particles(Density=2500kg/m3) in a tank with nitric acid. Air is sprayed from a nozzle to suspend the solid particles. Therefore the above flow is to be modelled. 6.I will be heavily grateful for help on the continuity residual. A lot of time is spent in computation because of small time step.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to write k and epsilon before the abnormal end	xiuying	OpenFOAM Running, Solving & CFD	8	August 27, 2013 16:33
Force can not converge	colopolo	CFX	13	October 4, 2011 23:03
Upgraded from Karmic Koala 9.10 to Lucid Lynx10.04.3	bookie56	OpenFOAM Installation	8	August 13, 2011 05:03
IcoFoam parallel woes	msrinath80	OpenFOAM Running, Solving & CFD	9	July 22, 2007 03:58
Could anybody help me see this error and give help	liugx212	OpenFOAM Running, Solving & CFD	3	January 4, 2006 19:07

June 18, 2015, 12:07		#2
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	what is convergence? Are your residuals relative or absolute? As can be read in the FLUENT manual on judging convergence - it may depend very much on the situation what is good enough. I'd like to take it one step further, and say: never, ever, ever judge your convergence solely on your residuals. Check if integral parameters of interest reach a steady value. Depending on your problem, these may be Nusselt number, drag coefficient, power number, mean velocity, etcetera. Do you see that these reach a constant value within say, < 1%, or do you see oscillations in these numbers? That's a much better guide of convergence than the residual.

June 19, 2015, 05:25		#3
ritwikdas18 New Member Ritwik Das Join Date: May 2015 Posts: 4 Rep Power: 11	1.My residuals are absolute. 2.I at this point of time don't have much idea of what to monitor in my problems since for me the major concern is volume fraction of solid(to check dispersion of solids) which is already present in the residual. 3.Also Fluent says to use a time step which takes 20 iterations / step to converge. However when I use a time step size of 1e-4 then it converges in 2-3 iterations. I think this is normal. However when I use a time-step size of 1e-3 the continuity residual keeps oscillating and hence it never converges not even in 100 iterations.Is the above normal? Thanks for your time.