Fire (heat transfer) modeling in ABAQUS

[cgbridges]

I'm attempting to look at heat transfer in structural members from a fire load using ABAQUS/Standard. I want to be able to input gas temperatures in the cavity (which I have found I can make using *cavity) and watch the effect on the enclosed member over time.

What I'm trying to learn is:
a) if this is possible (using gas temp as an input),
b) how ABAQUS handles heat transfer, especially through a medium like air
c) what kind of assumptions I would have to make, and what more standard ABAQUS heat transfer problems consider.

Any information on how ABAQUS handles heat transfer problems would be appreciated (I've only used it previously for strictly structural applications).

 

[corus]

Cavities are used to allow radiation between facing surfaces. If the cavity has an opening then radiation to the ambient (with the appropriate view factor) is also allowed. In your case you have radiation to an ambient (the gas) which is in the enclosure. If the gas obscures the surfaces then cavity conditions wouldn't be appropriate. I'd review radiation and see what boundary conditions are appropriate to your case.

corus

 

[cgbridges]

corus -

Thanks for the thoughts. What I'm really trying to understand is how ABAQUS handles the interaction between a gas and a surface, in terms of temperature. There's radiation and convection between the two, dependent on absorptivity, emissivity, etc and the difference in the temperatures, and I'm wondering if ABAQUS considers this or not.

About the best idea I've had so far in terms of modeling the interaction would be to use two plates, one being the "gas" and the other being the structural member. The temperature gradient across the space between (and the final surface temperature of the member) would give me an idea of the difference in gas temperature and surface temperature in a fire situation.

Any thoughts or alternate suggestions?

 

[corus]

You could have radiation to an ambient (the gas), and convection using the product of the gas emissivity and the surface emissivity. What coefficient you'd use for convection I wouldn't know. The gas temperature would be the bulk ambient temperature and hence from the calculated surface temperature you'd have the difference. I'd look to see what others have done in similar situations and where they have had good correlation with measured results.

corus

 

[torpen]

Hello,

You don't need to model heat transfer with continuum elements.
You can define a cavity thermal radiation and a thermal boundary condition for the convective part.
For the thermal radiation part you have to specify the temperature of the flames and the emissivity of the walls of the cavity.
For the convective part of the heat flux (product of the coefficient and the difference of temperature between the flames and the walls) the data are the temperature of the flames (which is known) and the heat transfer coefficient. The value of this coefficient depends on the nature of the fluid (air), its temperature and its velocity and of the geometry of the cavity through the Nusselt number.
You can find correlations for convective heat transfer in books.

Abaqus considers a nonparticipating medium.


regards

Torpen