natural convection in a air-cooled tower
 

I want to know the velocity and pressure fields in the air-cooled tower, how should I set the model? I have simulated it in a closed domain, but the tower is too big, about 100m in height and 50m in radius. So it's impossible to make the mesh fine. If simulating it in an opening domain, how should I set the information of inlet and outlet?

Re: natural convection in a air-cooled tower
 

A couple of questoins to consider with a domain this size.

1) Can you use parsol : it lets you get better refinement out of you grid structure.

2) Can you make use of any symetry in the model and only look at half/quarter of the shape being investigated.

3) Look at adjusting the F Array to accept more cells. Look in POLIS for an example.

I frequently look at spaces this large and have no problems. Just rember that in a model that is 100m long then it is not likely that you are able to achieve 1mm accuracy. The same goes for pressures and velocities, the bigger the model, the bigger the cell, the less able the CFD is able to define flow in the domin.

Note: on some of the larger domain it can take 48hrs of run time on a pretty faster p4 for the model to converge.

Hope this helps,

Leon