I am running simulations (with a commercial CFD package) to determine the wind-driven natural ventilation of various building configurations under different wind indicences. My goal is to achieve steady state solutions.

I get a very good convergence, i.e. a reduction by 5 orders of magnitude for all varibles appart from turbulent dissipation which drops by 3 orders.

However, for some configurations I get a an unsteady regime developing in the room, with the main vortex moving from one sidewall to the other.

I know that I could run transient sims, then figure out the period of oscillation, then time average over that. However, my sims contain about 2 million cells, so to do that would take alot of time on my pentium4 32bit windows desktop.

So my question is can I accept the first steady state solution, given the good convergence?

Hi Tom,

What CFD package are you using? We might be able to help with this depending on your answer.