CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Natural convection in a vertical cylinder (http://www.cfd-online.com/Forums/fluent/46030-natural-convection-vertical-cylinder.html)

 Virginie September 14, 2007 05:38

Natural convection in a vertical cylinder

Hello,

I want to model heat transfer (buyoancy-induced flow, diffusion) in a vertical cylinder with a high aspect ratio (L>>r). First I choose a 2D configuration (X-Y). The boundary conditions are : X = r et X = -r (lateral edges): wall conditions, non heat flux (insulated) Y = 0 (bottom): wall condition, fixed temperatute at 55°C Y = L (top): wall condition, fixed temperature at 15°C The interior is filled with water, I use the boussinesq approximation for density calculation. Gravitationnal acceleration is set 0 in x direction and -9.81 in Y direction. I do a steady simulation. Initial conditions are a 0 gauge pressure and a 35 °C temperature in all the domain. When I run the model, it is solving the energy equation but seems not to solve the flow equation, or not to take into account the gravity. Indeed, when I view the results, the static pressure remain 0 in the model (absolute pressure equal operating pressure), density is not modified too, only a temperature gradient is calculated by thermal diffusion. I wonder why I have not any hydrostatic pressure gradient in Y direction, and how can be the density uniform whereas there is a temperature gradient, the results make me think that gravitationnal acceleration is not working.Thanks for your help.

Virginie

 Joe September 14, 2007 06:18

Re: Natural convection in a vertical cylinder

Gravity is not turned on as disregarding any thermal density fluctuation modeling you should still get a hydrostatic pressure distribution. Fix that and your other problems will likely be solved.

 Jun Cai September 18, 2007 00:40

Re: Natural convection in a vertical cylinder

I think there maybe two reasons: 1. you didnot set the thermal expansion coefficient in "material" panel; 2. The property values you set resulted in a very low Ra(Rayleigh number). if Ra is very low, the buoyancy-drived flow will not take place. the heat tranfer is thermal conduction, not convective heat transfer. for details, you can see the fluent help.

 All times are GMT -4. The time now is 17:27.