CFX-5.5 simulating air free convection
 

Dear all, I'm working on CFX-5.5 to simulate the air free convection in solar heated spaces. I really appreciate if anyone give me some suggestions to solve the following problems. First, The "Global length", shown in the Out File, is used to calculate Rayleigh number but that value is not fit with the equivalent vertical length (i.e. actual heated space length times Sin£c, £cis the inclined angle of the heated space). So I'd like to know how to calculate the "Global length" in CFX-5.5? Secondly, Is it reasonable to use k-epsilon turbulence model & Scalable turbulence wall function to calculate the heat transfer coefficient of the turbulent free convection flow (i.e. Rayleigh number is larger than 1.0E+09). Is it a possible to input the empirical correlation to calculate the heat transfer coefficient of the turbulent free convection flow? I'd like to know the better approach of turbulent models to simulate the heat transfer coefficient in free convection. Finally, In CFX-Post, the temperature distribution plotted over the Global range option are obviously lower than those values plotted over Local range option. In general, which one option will be used to evaluate the mass and energy conservation in solar heated spaces?