CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   transient converging, but not steady (http://www.cfd-online.com/Forums/fluent/90845-transient-converging-but-not-steady.html)

PHS- July 21, 2011 18:21

transient converging, but not steady
 
I everybody!

I am currently working on a CFD-simulation of a Solar Tower Plant (that produce renewable solar electricity: the sun heats up the air which rises in the chimney through turbines)

The transient simulation is converging (temperature, pressure, and velocity remains stable after a few time step)
A steady simulation should also been converging but it's not! (no matter what turbulence model I try)

Has anyone already seen such a problem, and could give me some advice?

nikhil July 22, 2011 14:14

Are you trying to run the steady state with the initial/boundary conditions that are equivalent to the last flow time from your transient simulation?

If so, you should be getting the same results. If you are running the steady simulation with different BC/IC, then you may get different solution

PHS- July 22, 2011 14:29

Thanks for your answer!

I tried to run the transient simulation until I get convergence, and then to turn on the steady calculation without initialising (since the solution has already converged, steady calculation should work).

But when I do that, it's diverging, for example the temperature was stable for 2000 time step at 322K, and when I turn on the steady calculation, it's oscillating very fast between 300K and 360k, and is not converging at all (even after 30000 iterations...)

Micael July 24, 2011 12:58

Did you try to decrease the under-relaxation factor of energy equation? From 1 to 0.8 as an example?

PHS- July 24, 2011 20:59

yep, but that leads to high discontinuities in temperatures: the more I decrease it, the more convergence I get, but more discontinuities appears...

(by discontinuity I mean high change in temperature on the elbow, I can "see" the meshing through the temperature contours, instead of having a smooth transition.)

jsm July 25, 2011 14:25

Hi,

I think this phenomenon is transient in nature. If you have good results in transient solver, then why are you switching back to steady state solver?

If you want quick results, then increase the time step size without decreasing convergence level.


All times are GMT -4. The time now is 18:41.