CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Converge but monitored parameter not constant!! (http://www.cfd-online.com/Forums/fluent/32198-converge-but-monitored-parameter-not-constant.html)

kanok September 24, 2003 07:27

Converge but monitored parameter not constant!!
 
Hi

I've got a problem. I monitor the residual and heat transfer coefficient for convergence. When reach convergence, Iexpect both of them are constant.

I had run for long time... The residual is constant and converge already but the "monitored h" is not constant, still incresing.

Is the result close to the convergence result? What should I do to get the best result, more iteration?

Thank

Jonas Larsson September 24, 2003 15:48

Re: Converge but monitored parameter not constant!
 
Good observation - the heat transfer converges much slower than other properties. The explanation is that the thermal field and the thermal boundary layers first need a fairly well developed velocity field before they can start converging and the residuals are quite insensitive to the details in the temperature field.... You should monitor your heat transfer coefficient and run until it is constant, even though your residuals show that things are converged long before h stops changing. If you have very well resolved walls this can mean that you have to run much much longer (I'd guess from 2 to 10 times as long)than you would have done if you weren't looking for heat transfer.

kanok September 24, 2003 23:53

Re: Converge but monitored parameter not constant!
 
thank you very much... I have some confusion - - the value of h that I observed dynamically from monitor and the value of h from Report/surface integral.. panel are difference (23.49 and 25.7, respectively). Both of them based on area-wieghted average.

What wrong I did?


All times are GMT -4. The time now is 09:49.