Why the solution process stops?
 

Hi,

I'm running pretty complicated case with chemical reaction and heat transfer. For some reason the solution stucks at the iteration no. 19. The enthalpy residual is huge from the start. Its order is 10^17. What could be the reason for this?

Thanks in advance,

Julie

Re: Why the solution process stops?
 

I wish I knew which version of Star you are using. I had great discrepancies with the initialization of my runs using v3150 and v3150a, whereas v3100 never had that problem. The code would then diverge within 10 to 20 steps. If you have the possibility, try to run v3100 first, and continue with 3150 or 3150a with swicth 60 on. What could also help is to have the switch1 on as well (zero order pressure interpolation). Good luck!

Fred

Re: Why the solution process stops?
 

Hi Fred,

I'm running the v3150a.

What is the meaning of switch 60 and where can I learn more about the prostar switches?

Thank you for help,

Julie

Re: Why the solution process stops?
 

Are you running the case in parallell? Is it featuring droplets? In that case, there is a known time-out bug which seems to trouble the Linux version most. You can tell if this is what is troubling you by running the case several times, you should be suspicious if the case always stops iterating, but at different time steps/iterations. For this bug there is a bug fix patch, that you should be able to retrieve from your support office.

As for finding out about the switches, ask your support office, but beware they are different depending on which STAR version you use.

Good luck, Carl Fredrik

Re: Why the solution process stops?
 

A few more infos (sorry it took me so long to answer by the way): switch 1 (in v3150a) turns on the use of zero order pressure interpolation. Not very accurate, but you can choose to only use it at the beginning of your runs, when the flow behavior is always a bit hectic (to say the least...). There is no official list of these switches. We gathered some understanding of them with time: for swicth60:

SWITCH 60 Y: Default Gaussian procedure switched off (least sqaures procedure

will be used unless local coordinate procedure selected by

SWITCH 33, 35 or 61) Of course, if your code diverges abnormally, you can underrelax your euqation parameters a bit more, and you can steer that from the .ctrl file at every step having switch 50 on (say, at the beginning, you have 0.1 for pressure, and later 0.12). Hope it helps! Email me at frederic.souliez@bmw.de if you need faster help!

Fred