[PDH]

Hello, I'm new in CFD and I need some advisement or some light about if I'm in the right way.

I'm making a simulation of an oven. I want to manipulate the power of the heaters to achieve around 550-600 ºC in the central zone of the oven. I don't need extreme accuracy, with +-10 or 20 ºC is enough.

The geometry is simple, a box of 650x650x350 mm with a hole (340x340mm) at the bottom. The heaters are at two of the sides. The other sides are insulation. I have created a space around the hole to get away the ambient air and I placed BCs of outlet and inlet pressure in there.

Based on what I found in other threads about natural convection I made a pressure-based, transient study with the following parameters:

Energy: On
Viscous: Laminar (If my calculations are not wrong I have a maximum Re of around 720, corresponding to a max vel of 0.1 m/s)
Radiation: Surface to surface
Heat transfer coefficient: 5 W/m^2K (I'm not sure about this value. It's normal for natural convection but maybe at these temperatures is invalid)
Air density: Incompressible ideal gas
Scheme: Coupled
Pressure discretization: Body Force Weighted
Transient Formulation: Second order
HOTR Activated
Relaxation Factors reduced to 0.65 (momentum) and 0.8 (energy)
Step size: 0.5 s, with 10 iterations per step.

I'm not sure if I can trust the results. I have residuals of 10^-3 in continuity and 10^-4 in the other variables.

The results are these: https://imgur.com/ChYBz9O

Thanks for the help!

[vinerm]

For natural convection, it is not Re that is looked at but Grashof number. If Gr is greater than , then turbulence model must be enabled; most likely you need that. Secondly, use a URF higher than 0.9 for energy. Since you have some openings, i.e., inlet and outlet, you do not need to run a transient case. A steady-state case will work. Turbulence will have major effect on thermal diffusion and the thermal field would change drastically.

[PDH]

Originally Posted by vinerm

For natural convection, it is not Re that is looked at but Grashof number. If Gr is greater than , then turbulence model must be enabled; most likely you need that. Secondly, use a URF higher than 0.9 for energy. Since you have some openings, i.e., inlet and outlet, you do not need to run a transient case. A steady-state case will work. Turbulence will have major effect on thermal diffusion and the thermal field would change drastically.

Thank you vinerm. Your are right, my Gr is bigger than 10^6 (5x10^10 aprox).

For heat tranfer what is better, low y+ with k-omega or big y+ with k-epsilon?

[vinerm]

Every phenomenon happening close to a wall is called low phenomenon, such as, conjugate heat transfer, drag and lift, natural convection, etc. But you can use either or . Just use low

[PDH]

Finally, I used SST k-omega. Much better (and logical) results now. Thank you!

[vinerm]

Nice to know that you got good results.