How can I set up a problem of natural convection full bouyant ? My density should vary on by temperature. I have introduced a CEL function wich specify this law. But the solver cannot read my CEL. Why ?????

that are really interesting details on solving a problem

Details:

- CFX 10.0

- The fluid is created by Owner: so the properties have the value that I want (in order to have Prandtl number = 1 - 10 - 100 - ...)

- The Problem is when I set the equation of state:

Equaetion of state = value

Density = law

% law is an appropriate CEL expression just created %

Density depends on = temperature

- Buoyancy properties =

Thermal expansion factor = 0.01 [K^-1]

% But the Manual tell that the thermal expansion factor is used only for the Boussinesq problem, and so for the full Buoyant model is not used. So I don't know if I have to set this parameter or not %

- Another parameter: CEL expression: I introduce this CEL expression

law = 1000[kg m^-3]*(1+0.01[K^-1]*(100[K]-T))

% this is the relationship between the density and the temperatura %%

--> in this way, the PreProcessor tell me to introduce:

- a Temperature reference value

- a density reference value

% It is so strange introdurre both of them becouse:

- for the Boussinesq problem I have to introduce only the Temperature reference value

- for a Full Buoyant problem, for example of an ideal gas, I have to introduce only the density reference value: and the solver goes on !!

--> Where is the problem ??????

law = 1000[kg m^-3]*(1+0.01[K^-1]*(100[K]-T)) When we guess T=300 K your density will be become negative!!!

I would copy a material, which is near to wat you want to a ne material and set up a bouyancy simulation with that. when it works, i would start to change the properties step by step to my wishes. So you have a change to detect the failures and correct them.

Yes, at 300 K the density could become negative, but in my problem 700 C < T < 1000 C, so the density will be always positive. Now I found the correct statement in CFX. For full buoyancy model you should leave the coefficient of thermal expansion "beta" (becouse this coefficient is for only Boussinesq problem) and you should impost a "Table generation" by click on the panel of the material model, that you creating. In this way Cfx works correct !!

Hi Tommy, The Boussinesq approximation is automatically set, if you use for example a fluid of "air at 25". But if you use "air ideal gas", then the full buoyant model is automatically used in CFX. You then need to set a reference density instead of a reference temperature. Hope this helps TobiasZ

thx, that's true

fine ;-) ;-) hope this solves your problem