Stuck? 
support@mecway.com
support@mecway.com
Mecway
FEA
Unexpected Cooling during Heating
Hi,
I’m working on a model to predict temperatures of a reinforced concrete beam subjected to a standard fire. I have tried a few geometries, a pretty standard T-beam to a simple cube. I’ve attached the cube example for simplicity.
Three (bottom and 2 sides) faces of the cube are subjected to convection and radiation. For the moment these are with constant convection and radiation coefficients. And subjected to a table containing the time-based temperature values from a standard fire curve (in this case ASTM E119). I’ve set the two other sides to have a constant heat flux of 0, to simulate a semi-infinite beam. And finally the top is subjected to convection with a constant ambient temperature of 20C, open to the air without being subjected to the fire. I’ve set the initial conditions to both the nodes and the elements to a temperature of 20C.
When I look at the temperature distribution profile in the beam it shows me the types of isotherms I would expect after a certain time, however there is one small glitch. The temperatures initially drop for about the first 7 minutes. Considering my initial conditions are at 20C (in this one i'm pretty sure that i have the nodes set to 20C, however i have tried the elements also), my unexposed surface is 20C and the fire starts at 20C and quickly heats up I see no motivation for this initial cooling. This is the most pronounced when the fire exposed surfaces are subjected to both the convection and radiation. However it is present when subjected to simply convection and radiation independently. Radiation seems to have the most impact.
Could you look through the attached to see if I have made a mistake setting things up?
Thanks,
Tom
I’m working on a model to predict temperatures of a reinforced concrete beam subjected to a standard fire. I have tried a few geometries, a pretty standard T-beam to a simple cube. I’ve attached the cube example for simplicity.
Three (bottom and 2 sides) faces of the cube are subjected to convection and radiation. For the moment these are with constant convection and radiation coefficients. And subjected to a table containing the time-based temperature values from a standard fire curve (in this case ASTM E119). I’ve set the two other sides to have a constant heat flux of 0, to simulate a semi-infinite beam. And finally the top is subjected to convection with a constant ambient temperature of 20C, open to the air without being subjected to the fire. I’ve set the initial conditions to both the nodes and the elements to a temperature of 20C.
When I look at the temperature distribution profile in the beam it shows me the types of isotherms I would expect after a certain time, however there is one small glitch. The temperatures initially drop for about the first 7 minutes. Considering my initial conditions are at 20C (in this one i'm pretty sure that i have the nodes set to 20C, however i have tried the elements also), my unexposed surface is 20C and the fire starts at 20C and quickly heats up I see no motivation for this initial cooling. This is the most pronounced when the fire exposed surfaces are subjected to both the convection and radiation. However it is present when subjected to simply convection and radiation independently. Radiation seems to have the most impact.
Could you look through the attached to see if I have made a mistake setting things up?
Thanks,
Tom
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Comments
Unrelated to the error, you don't need to specify a 0 heat flux because that's the default boundary condition when nothing else is specified. It's redundant but harmless.
Another tip, I prefer to set initial temperatures on elements rather than nodes because then you can refine the mesh and it'll still be applied to the new elements, but not for new nodes.