I have an exhaust fan in physical model which pumps the air out of the system. In fluent i could not decide whether it should have a negative or positive pressure jump. To make it act as an exhaust (not the opposite, an intake) which one should i choose? (In documentation it says "You must be careful to model the exhaust fan so that a pressure rise occurs for forward flow through the fan". )

Hey: In my mind just check it two times. Once with positive and next with negative. If you had flow out it means you have applied negative and in next it is positive. I had some experience but I'm not sure, maybe you can apply other boundary conditions instead of this type. I'm pretty sure you can. Thanks/Farid

I built a small grid to make some tests. A small rectangle with an entrance, a mid boundary and an exhaust. I used laminar flow eq.

For an intake and an exhaust with positive values both, (mid assigned as interior) there is a net flow from inlet to outlet. I noticed that the average velocity is close to the value below;

According to Bernoulli eq. p/rho + (v^2)/2=c and if v1=0 then 2*(p2-p1)/rho=v^2 and v=sqrt(2*(p2-p1)/rho). The pressure difference between two points decides the magnitude of the velocity and it changes with friction (laminar and turbulent different).

If we come back to my case the average velocity (by fluent) is close to the value i estimated with bernoulli.

Further, i put a fan boundary in the middle, which has the same direction with the flow. I gave a constant pressure jump. I found an average velocity with fluent analysis. Guess now, how i estimated the velocity. I added the pressure jump value to my total pressure difference and found the velocity very close to the average velocity. I concluded that the velocity is found using "net" pressure difference value. (If you have a fan in reverse direction you will substract it.)

Besides, i assigned velocity inlet to entrance, fan to mid and outflow to the exit. My observations are such that the velocity boundary is dominant and it is not affected. If you say 10 m/s, whatever you give for fan pressure does not change the amount of flowrate decided with 10m/s.

If you want your flowrate to be affected by pressure, then all your boundaries should be pressure type boundaries.

These are my observations but i can not assure that they are 100% right.