[czaru]

Hi Fellows,

I am modeling PV panel heated by sun. Between panel and roof there is 5 cm air gap in order to cool it down by natural convection.

please refer to the picture for more information:



I am using solar load model and s2s for radiation between pv and wall (insulation).
Air properties (density: incompressible gas, operating density set in the operating condition: 1.17, air temp set on the pressure inlet 300K)
Operating pressure 101325 Pa

initially I was using laminar model and got results:



then I change for k-e model (enhanced wall treatment, full buoyancy effect) (on pressure inlet I set: 0.001 for turbulence scale and 1% for turbulence intensity):





what do you thing about that?
1. which solution is more probable
2. why we can not observe any air leaving the gap when modeling with turbulence model?
3. do you have any suggestion for this model in general?

Thanks for hepl.
Cesar

[EDRMedeso]

Hi!

In order to decide whether or not a turbulence model is appropriate for natural convection problems like this one, you should consider the Rayleigh number (Ra). Problems with high Ra call for a turbulence model. As Ra increases rapidly with respect to the characteristic length (Ra ~ L^3), a turbulence model might be necessary for long panel lengths. In that case, you should consider the k-omega SST model with transition.

However, with a pv-panel length equal to 3 m as in the current problem, this becomes a low-Ra problem, and turbulence modelling is neither necessary nor appropriate.

[kennethchaw]

Originally Posted by EDRMedeso

Hi!

In order to decide whether or not a turbulence model is appropriate for natural convection problems like this one, you should consider the Rayleigh number (Ra). Problems with high Ra call for a turbulence model. As Ra increases rapidly with respect to the characteristic length (Ra ~ L^3), a turbulence model might be necessary for long panel lengths. In that case, you should consider the k-omega SST model with transition.

However, with a pv-panel length equal to 3 m as in the current problem, this becomes a low-Ra problem, and turbulence modelling is neither necessary nor appropriate.

Then I wonder if a rectangular box with length of 7m? What is the range of Ra that telling me I should use turbulence model for natural convection problem or not ??

Thanks much!!