Hi, experts

my model involve the buoyancy force under the natural convection,piso, the density is f(T), when star is running, there is a warning013 "maximum specified corrector stages 20 reached before convergence criterion is satisfied" in star.info. and i find the resp=2.53338e-1. so i try to modify under relax and tolerance. but it still in. do you anyone have met this problem. is it serious? how can i do?

thank you in advance!

jane

hi, In transient analysis, is there any method to deal with this warning or -ve density error, apart from reducing the delta time. Any inputs are appreciated, since i am suffering from -ve density error in transient analysis.

thank you, novice

but my model is not transient analysis, but stead analysis, because my model involves natural convection, so the temperature and momentum equation effet each other, only way is using piso. so in this case, how can i do?

have a good day! jane

Do you running your simulation in double precision? It is recommended by such problems. And did you try to increase the number of iterations?

I stand under correction here, but can PISO be used for steady state analyses? I was under the impression that it can only be used for transient cases. SMPISO (sp?) is the version of PISO used for steady state. But I have never had any trouble running bouyancy cases using SIMPLE.

There are a number of reasons why a run may start running into PISO corrector limits. But probably the most common if these is the case is approaching steady state. PISO works on time gradients, and if nothing is changing in time, there is nothing left to solve, so the solver starts to hit the corrector limits. If you suspect you may have steady state, start a new steady state run with the previous flow field as initial condition, and see what happens.

That said, it is true that you need to use double precision for bouyancy driven cases. The reason is that particularly at the beginning of the run, the velocities and forces are so low they can easily disappear into round-off error.

In analysis control, under eqn behavior and solver params, u can change the number of sweeps to a higher value and afterwards check if this could eliminate this warning message. There may be other things involved though. See pages 1-15 and 1-17 of user guide for clearer info.

You can also try to look for problem areas in your model, turning on output of highest residuals. May be your mesh needs additional refinement or it is too scewed somewhere or etc

I was going through the archived forum and saw this one which may be interesting on this negative density thingy:

http://www.cfd-online.com/Forum/star...0.cgi?read=619

Just in case.