[Mohamed Abdulazim]

Hello everybody,
I am trying to solve a heat transfer problem(conjugate heat transfer )by the fluent, which is a heat transfer enhancement in a rectangular channel with baffles. I left the computer solving the problem for about 12 hours and it performed about 3396 iterations. I need to know whether I have reached to the convergence and the results are reliable or not. The residuals for the equations reached to the values as they appear in the captured photo of the screen and the value of the net mass transfer rate is 3.921359e-08 . Also, I used monitors for variables such as mass-weighted average at the outlet temperature, surface integral for lower plate surface heat transfer coefficient and the sum of all heat fluxes from all entities, but I am confused about the monitor for the surface heat transfer coefficient as the curve fluctuated up and down as shown in the figure. Is this enough for convergence knowing that the URFs are the default values?
Thanks in advance.

[vinerm]

The simulation appears to have converged. However, you are not monitoring the right quantities. For heat flux, always measure either area integral or area-average. Temperature monitor appears to be good. Mass imbalance should be compared against all the inlets and outlets; it should be smaller than 1% of the smallest mass inlet or outlet.

[Mohamed Abdulazim]

Thank you very much, Vinerm, for your helpful response. You mentioned that for heat Flux I should use surface integral or area -weighted average,what about boundaries for setting this monitor? Does choosing all boundaries make different values than choosing surfaces boundaries only?I mentioned in the above discussion that area -weighted average of heat transfer coefficient fluctuated,Is that monitor useless in convergence criteria?

[vinerm]

Multiple boundaries should be chosen only if they are of similar type, such as, all with positive flux or all with negative. It is not the value but the stability that matters more. If the monitors are stable, then work is done. There will always be fluctuations but if those are within 2-5% of the absolute values, then it good. Since this percentage is based on absolute value, therefore, it becomes important that similar types of boundaries are chosen together and correct type of averaging is done. Most of the time, it is good to use mass-weighted or area-weighted averages. Integrals should be chosen for volumetric quantities or fluxes.

[Mohamed Abdulazim]

Originally Posted by vinerm

Multiple boundaries should be chosen only if they are of similar type, such as, all with positive flux or all with negative. It is not the value but the stability that matters more. If the monitors are stable, then work is done. There will always be fluctuations but if those are within 2-5% of the absolute values, then it good. Since this percentage is based on absolute value, therefore, it becomes important that similar types of boundaries are chosen together and correct type of averaging is done. Most of the time, it is good to use mass-weighted or area-weighted averages. Integrals should be chosen for volumetric quantities or fluxes.

Dear,Vinerm.
I performed another simulation for the same problem but with a higher velocity. I have tried to reduce residuals lower than the shown results but there is no hope to decrease them, although I have changed the URFs many times. I attached the captured screens of monitors. Are the shown results sufficient for convergence? I need your advice.
Thanks a lot.

[vinerm]

The results look good. Though you do not have good convergence for continuity, which could be due to mesh or numerics, fields are more or less stable. If the mesh quality is good, try running with first order discretization for first 200-300 iterations and then switch all to second order to improve residuals. If it still does not work, try changing the gradient scheme or pressure-velocity coupling to coupled.