I have a decoupled Defrost problem that diverges in the thermal transient analysis. The steady state is calculated at the starting temperature of the transient analysis. I had subroutines that I suspected as the source of the problem, but I arranged the same mesh in an adiabatic calculation and still diverges. Could be the turbulence model(I have used k-e highRe and cubic) or high values of k or epsilon the source of the divergence? I think the tstep shouldn`t be the problem even though a rough tstep would give poor results. I would appreciate some hints or help from someone who has background or experience on this type of problem.

HOW I CAN EXTRACT THE ASCCI DATA VOR U,V U'and V' from starCD thanks

Jordi,

What kind of Courant numbers are you getting?

You may have play with your time step (decrease) to keep the Courant number in check.

When you say it diverges what do you mean? Will the model run at all? Can you get it to write a post file? What type of error messages are you getting?

Ron

Ron,

Sorry for the dealy in my response, I had holidays and I couldn`t review my e-mail.

I don`t get courant numbers, because my velocity field is decoupled from the thermal transient analysis.

THe error messages I get deal with the increasing of the enthalpy residual. (Could an increase of the turbulent dissipation be a "heat source" in my energy equation? I am using k-e/High Reynolds number model)

The ...unimportant... messages of the .info file are of the type...

*** WARNING #065 *** OUT OF BOUNDS VALUE -3.3711E+05 FOR SCALAR VARIABLE ENTH AT CELL 1780258. CORRECTIVE ACTION TAKEN

*** WARNING #011 *** NEGATIVE TURB. K.E.AT CELL 1483459 -1.9211E-02 CORRECTIVE ACTION TAKEN

*** WARNING #013 *** MAXIMUM SPECIFIED CORRECTOR STAGES 20 REACHED BEFORE CONVERGENCE CRITERION IS SATISFIED.

<<<<ENTHALPY EQUATION NOT CONVERGED>>>>

REST =nan

... to give some examples

... the model will run for a few time steps, but the decreasing of the time step will not modify substantially the behavior of the calculation, it still diverges.

Thank you a LOT for your help.

Jordi

Have a look at the pressure at the monitoring cell. If it is still zero after the second time step you can stop the calculation. Then at least at steady calculations it sometimes helps to run the case without pressure initialisation and very strong underrelaxation.

If this does not help ask the support.