[Carlos_PdL]

Good Morning,

First post, so I'm not really sure how I have toi post to make everything understandable.

I'm trying to simulate the flow of some heat exchangers in a semi-enclosed region. Basically I have 3 walls and open ceiling. I have an industrial transformator which is generating heat. To get rid of this heat it has several heat exchangers with fans that blows the hot air out the open wall.




My simulation consists of a mesh with the transformator and the 3 walls surrounding it enclosed within a cube of air. The fans are modelled as a constant change in pressure (using the internal fan BC from Fluent) and the heat exchangers are modelled using the temperature of the oil and a heat transfer coefficient. The external BCs are all Pressure Outlets at 0 Pa modified pressure and letting Fluent deal with where there is air flowing in or out. Convection is modelled wih boussinesq aproximation.



Steady state calculation does not converge (which is expected since the aiflow meassured in reality is unstable).

Transient simulation diverges due to floating point exceptions after little time.

I'm trying to improve the model, but I'm not really sure how.

My first idea would be making a larger cube and getting rid of the external BCs, since the Pressure Outlets are probably the cause of the divergence.

I think changing them to walls far away, could help the convergence for steady state and transient simulation. I'm guessing for steady state I'd have to define some kind of heat excahnge on the walls of the cube (convective BC maybe).

I'm open to suggestions on how to better model this flow, or possible causes of my divergence.

Thank you in advance

[helgaella99]

Increasing the size of your computational domain (the cube) can help minimize the impact of boundary conditions on the Drift Hunters 2 flow and thermal fields. This can reduce the artificial influence that the pressure outlet boundaries might be having on your solution, potentially helping with both steady-state and transient convergence.