CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Convergence Problem (http://www.cfd-online.com/Forums/fluent/30105-convergence-problem.html)

Duc August 1, 2002 07:05

Convergence Problem
 
Dear All,

I am experience convergence problems with my engergy and continuity. No matter how short time step I set, the residuals of energy and continuity will be reduce less that 1.5e-3. After 150 iterations, they seem to level to a steady value. Does anyone have any adice of what to do? Many thanks in advance.

Regards,

Duc

Laika August 1, 2002 08:17

Re: Convergence Problem
 
They are leveling out to a constant value around 1.5e-3 and you're talking about convergence-problems???

--don't thrust only on residuals; monitor something else --what about your mass-balance? Is it ok? same for heat-balance --is your mesh of good quality?

What ere you simulating? You're talking about time-steps and 150 iterations? Do you mean you have 150 iterations for each time-step?

Laika, still orbiting

Duc August 1, 2002 08:35

Re: Convergence Problem
 
Thanks Laika,

The convergence criteria was set to 1E-3 at default. In 2D, I am simulating a normally oscillating wall parallel to a stationary wall with a thin, long air gap, two ends are open to ambient. The frequency of the oscillation was 2kHz, small amplitude of 5.7micons. I divided 1 cycle to 100 time steps, and solving 1 time step takes over 150 iterations. I know it should be far lower, like 20. But soultion wont converge even if I increase the time step to 400 per cycle.

I was thinking the set-up was wrong and that I should be applying compressible mesh, but this function is not available without license.

Any further help or advice would be very much appreciated.

Regards,

Duc

Laika August 1, 2002 11:28

Re: Convergence Problem
 
looks like a challangin problem...

How have you specified the oscillating wall?

Laika, still orbiting

Duc August 1, 2002 12:04

Re: Convergence Problem
 
The oscillating wall was specified using a unsteady velocity UDF. The problem is I am not specifying any flow and both open ends between the walls can be inlets or outlets. I am quite new to Fluent and CFD, so if the set-up is incorrect, then I won't be surpised.

Regards,

Duc

Alamgir August 11, 2002 21:04

Re: Convergence Problem
 
I think you can have a look over under relaxation factor, if u half the all values, it can be worked.

Alamgir


All times are GMT -4. The time now is 02:08.