natural convection
 

hello everybody! I'd like to simulate the vertical flow in a uniformly hetaed parallel plate channel. I expect to see a vertical upright flow. Instead I got (CFX4.3) an inflow in middle of the channel, probably due to the boundary conditions (PRESS). Could anyone tell me, please, what boundary conditions I could use in the simulation ? Thak you in advance BY Franco

Re: natural convection
 

Hi Franco,

The heated parallel plates should be walls, set to the specified temperature. If you set them as pressure patches then strange things would happen (eg inflow in the middle of the channel). The top and bottom boundary conditions will vary depending on exactly what the flow is doing - they could be walls, pressure or mass flow depending on the set-up.

If you want me to be more specific you will have to describe your set-up more fully.

Regards, Glenn

Re: natural convection
 

I set the heated parallel plates as walls with the specified temperature. At the top boundary I get an outflow (along the heated walls) and an inflow in the middle of the channel, in order to balance the mass I suppose. Indeed there is no mass flowing inside the channel, even though that's should happen in the real world. I tried with outlet and press boundary conditions but without relevant changes. The channel width is 2 cm and 50 C degress is the temperature of the heated wall. Thank you for your kindly attention. Franco