I am runing a case about the wind passing throw some buildings. The solution does not converge. I use RKE to run my case and set the convergence criteria of the residual be 10^-3.

Due to the limit of the storage of my computer, the length of downstream is about 4 time of city size so that I can save it and export as .msh file. I have setted the exit plan be the pressure outlet boundary. In the boundary (the exit), I find that there are velocity gradient alone the boundary plan. Are the length of downstream so short that my case cannot converge?

I have found from Fluent's manaul that the boundary should set a place where there are no gradient. Actually, I am not understand the meanings of this. So what does that means?

Finally, Do there any other method or criteria so that we can decide the solution is converge even if the residual do not converge?

Thanks very much!!

Try using the pressure outlet B/C.

Oops. You are using the PO B/C. Don't worry about velocity gradients existing at the outlet. THE PO B/C takes care of that. Only if your velocity profile is NOT fully developed should you bother about the exit B/C, and that too only if you are using the OUTFLOW B/C as it is based on a fully developed flow assumption.

Hi,Maoasis It may be due to the bounary condition. Most cases about wind environment use velocity (inlet) and outflow (outlet). The downstream length is about 10~12 times of the characteristic builidng height.The details can be found from website: COST ACTION C14: Urban Wind Engineering and Buildings Aerodynamics.

Good luck Justin