Hi, I try to simulate a cooling tower, but I don't know how to set the bc to this model work. I use gambit to generate the model, heigth = 133 m and radius = 60 m. I generate 2d model, the X axis is the symetric axis, the gravity is set on X axis, I use pressure inlet at the inlet with gauge 0, and outflow at outlet, I put an source term o energy 160MW in interior of tower to represent the heat echanger (the tower is a natural convection). I have created 3 fluids in gambit, one to represent the air before the heat exchanger, one to represent the air in heat exchanger, and in fluent I set the source terms of energy in this fluid, and another fluid to represent air after the heat exchanger. But the solution becomes only difusive, there's no velocity profile at inlet or outlet. I think the reason is the density of air, I use constant. For this problem what kind of model to density I may use???

Density must be some function of temperature.

What function do you sugest? I tried with ideal-gas, but these model could't be applied with the outflow BC at outlet!

You set pressure boundary condition, but no pressure gauge. So there is no force to drive the air move.

I think the gauge pressure is zero, the inlet is taken at atmospheric pressure, is this correct?? The force to drive the air move is the energy source term inside the tower or not??

Thanks.

Used ideal gas or define density vs. temperature points (piecewise linear). You can switch the outflow BC to a pressure outlet instead.

Thanks Evan,

But if I switch the outflow to pressure outlet I must define the outlet temperature and I don't know this value.

I will try with outflow and define density as function of temperature points (piecewise linear).

[mttg]

Can we see the relative humidity distribution inside and outside of the cooling tower using Fluent?

[Laci]

Originally Posted by Roberto
;115439

Thanks Evan,

But if I switch the outflow to pressure outlet I must define the outlet temperature and I don't know this value.

I will try with outflow and define density as function of temperature points (piecewise linear).

Why don't you use the boussinesq for density and set the 1/T value?