Page 1 of 1

Boundary Condition

Posted: Mon Dec 17, 2018 7:02 am
by wenhao
In the boundary conditions section, for the exchange coefficient, is it means the thermal conductivity?
Currently, i am doing ventilation problem, for the composite wall material, I calculate the total heat resistance by using formula R=1/hA+L/kx then convert to overall heat transfer coefficient, U. Hence, i fill in the U value in the exchange coefficient column. Please correct me if I am wrong. Thank you.

Re: Boundary Condition

Posted: Tue Dec 18, 2018 3:40 pm
by Luciano Garelli
Hello,

When solving a thermal problem and you select as boundary condition "exchange coefficiente" you have to impose an external or reference temperature and the heat exchange coefficient (SI units W/(m2*K)), in order to get a flux based on the internal temperature
h = q/Delta T.

The thermal conductivity (SI units W/(m*K)) is a material property and is defined as the ratio between the heat flux and temperature gradient vectors
k = -|q/ grad(T)|.

You can check the user documentation sections 6.4.1 and 6.4.2, "Coding boundary condictions"

Regards

Luciano

Re: Boundary Condition

Posted: Fri Dec 21, 2018 3:00 pm
by wenhao
Hi, thank you for your reply. If the boundary condition (wall) is made up of composite material, can I treat the exchange coefficient as overall heat transfer coefficient, U as it also has same unit W/m2.K?

Re: Boundary Condition

Posted: Fri Dec 21, 2018 3:11 pm
by wenhao
I not sure about the exchange coefficient means convective heat transfer coefficient, h or can be treated as overall heat transfer coefficient, U?

Re: Boundary Condition

Posted: Fri Dec 21, 2018 7:09 pm
by Luciano Garelli
Hello,

I think that the thermal conductivity in composite materials will be anisotropic, so I think that you will need to couple saturne with syrthes or use CS internal coupling in order to model the thermal conductivity correctly, but if you have an isotropic thermal conductivity and do you want to model the thermal resistance through the composite material (not the diffusion process), I think that you can use

h_eff = h_ext + k/t

Regards,

Luciano