CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   free convection inside tube (

Agis March 12, 2013 07:24

free convection inside tube
Hello!! i am new in CFX and a need your advise in a simulation i am trying to perform. My geometry consists of a vertical conical tube. All i want to do is to set different values for temperature at the top and the bottom boundary in the same time and as a result create recirculation inside tha tube. My working fluid is water. The geometry dimensions are:
  • 3 microns diameter at the bottom
  • 1 micron diameter at the top
  • 2.5 mm length.

Any advice on the boundary conditions I should use?

JuPa March 12, 2013 07:51

Hi! I suggest you visit my thread: as I have a similar problem setup.

Your units of length are microns! Are you sure about this?

As you have 3 microns diameter at the bottom and 1 micron diameter at the top your aspect ratio is not homogeneous. I'm not sure if Rayleigh Bernard theory will cover non homogeneous aspect ratios.

I suggest you read Fundamentals of Heat and Mass Transfer by Incropera and DeWitt, namely Chapter 9.
Also, read Physical Fluid Dynamics by Tritton chapters 4, 14, 16, 17 and 22. The book is quite short and those chapters can be read in a few hours of good reading.

ghorrocks March 12, 2013 18:17

As stated in the previous thread already quoted, the key non dimensional number is the Rayliegh number. I suspect the Ra number here is quite low meaning the flow is laminar and dominated by conduction.

The bounadry conditions you describe sound trivial to set up. Have you done the tutorials?

Agis March 13, 2013 04:09

I have solved the most of the tutorials. I am quite familiar on the free convection theory but i am still learning. Please give me at first some advice on the kind of boundary conditions you would use in an enclosure similar to this supposedly that the aspect ratio is low and the Rayleigh convection is valid. i think wall at different constant temperatures for the top and bottom plates and adiabatic wall for the rest boundary. the run could be both steady state and transient?

ghorrocks March 13, 2013 07:02

The boundary conditions must match what you are modelling. I have no idea what you are modelling so cannot help. If you describe what you are modelling we might be able to help you.

JuPa March 13, 2013 07:10

Please provide an annotated diagram. The annotations should state what your boundary conditions are.

All times are GMT -4. The time now is 18:53.