CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Simulating Heat Transfer (http://www.cfd-online.com/Forums/cfx/109400-simulating-heat-transfer.html)

Spookz November 16, 2012 11:40

Simulating Heat Transfer
 
Hi everybody,

I'm new to this forum and I hope you can help me. :)

Concerning my thesis i have got to simulate convective heat transfer inside an autoclave as well as the heat transfer on a Tool placed inside the autoclave.
I am still learning to use ANSYS CFX at the moment. So I want to do some simple test simulations about convective heat transfer from fluid to a solid square inside a duct. I want to simulate the heating of the square from an initial temperature of 300 K, while to sourrounding fluid has a temperature of 1000 K.
The fluid domain is meshed with tetrahedrons and inflation layers near the square. The Solid is meshed with a structured mesh.

The fluid is air at 25°C, Reference pressure 1 atm, Turbulence Model SST, while the square is of aluminium. I have set both, fluid and solid Heat Transfer to Thermal energy.
I have set a domain Interface between solid and fluid and have set heat transfer to conservative interface flux with a contact resistance of 120 W/mēK.

The condition at the inlet of the duct are 3 m/s Normal Speed and a temperature of 1000 K. Outlet is Relative Pressure 0 Pa.

First i did a steady state simulation with the above parameters and afterwards a transient simulation with initial values from the steady state simulation. The timesteps are 2 min and total time is 30 min.

My first results do not seem to be correct. The steady state simulation shows a uniform temperature profile in the square AND the fluid of 1000K. The transient simulation doenst change over time and has the same temperature profile in every timestep.
Acutally I am expecting a time dependent heating of the square.

1. Do i have to set the temperature at the inlet to 1000 or 300 K in the steady state simulation?

2. Do I have to choose smaller timesteps?

3. What am I doing wrong? :)


Hope you get my problem. If not, please ask me.

Thanks for help.

Greetings.

mat_cfd November 17, 2012 03:09

I suppose you have to use coupled boundary condition between for Fluid sloid interface. It is taken by default even if u dont define any interface and flux conservation is satisfied there. Inlet at 1000k in Fluid domain is fine, Initialize solid domain with 300 K.

ghorrocks November 17, 2012 20:09

Why do you have a contact resistance between solid and fluid domains? What does this represent?

Quote:

The steady state simulation shows a uniform temperature profile in the square AND the fluid of 1000K.
Of course - the solid has no heat generation so its steady state temperature is the same as the surroundings. Basic stuff.

Quote:

Acutally I am expecting a time dependent heating of the square.
Then do a transient simulation with an initial condition as you specified (300K object, 1000K autoclave).

Here is the general FAQ on accuracy: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

Spookz November 19, 2012 12:08

Thank you for answering.

The problem was that there has been a uniform temperature profile in transient simulation.

I did the setup without contact resistance and just a transient simulation now. This gives better results now.

Do I have to adjust the solid and the fluid mesh for a better result(Contact Sizing)?

ghorrocks November 19, 2012 17:52

Quote:

I did the setup without contact resistance and just a transient simulation now. This gives better results now.
Fluid in contact with a solid has no contact resistance. The interface boundary takes care of the conduction/convection. But note unless you have radiation activated it ignores the radiation.

Quote:

Do I have to adjust the solid and the fluid mesh for a better result(Contact Sizing)?
You always have to check your mesh for accuracy. If you don't then how do you know how accurate your simulation is?


All times are GMT -4. The time now is 14:04.