FLUENT - How to define your own turbulent thermal conductivity, C_K_T
The UDF macro DEFINE_PROPERTY allows you to only modify laminar material properties.
When you do a turbulent simulation, FLUENT will always compute a turbulent thermal conductivity and add it to the (UDF-defined) laminar.
But there's a workaround to this (with some limitation..)
The turbulent thermal conductivity is calculated within FLUENT as follows:
#define C_K_T(c,t) (C_MU_T(c,t) * C_CP(c,t) / M_keprt)
The effective thermal conductivity is then defined by...
#define C_K_EFF(c,t) (C_K_L(c,t) + C_K_T(c,t))
As you can see above, C_K_T is calculated using the turbulent Prandtl number "M_keprt". This can be modified in the Define---Models---Viscsous panel.
(In FLUENT 5, you can modify it using...
(rpsetvar 'keprt 1.0)
... -- note that you will probably have to change something else using the GUI (and then change back to the original value) so that FLUENT performs an internal update of all cached RP variables.)
If you set the turbulent Prandtl number to a very high value, C_K_T will become infinitely small. You can then add your own calculation of the turbulent heat conductivity in your DEFINE_PROPERTY UDF that is used to define C_K_L (the laminar heat conductivity).
Setting the turbulent Prandtl number to a very high value will have (at least) the following SIDE-EFFECT(s) in FLUENT6.0.13:
- The turbulence production due to buoyancy effects will disappear from FLUENT's calculation. This is only effective when Gravity is turned on and the density varies with temperature (or pressure, or what ever you defined through a UDF -- holds also for boussinesq approximation!). Reason: The buoyancy turbulence production term is calculated as "....../Pr_t".
- No more side effects found by now...
For the premixed combustion model, the turbulent Prandtl number is set from the RP variable "species/sct" -- this is the turbulent Schmidt number for turbulent species diffusion. Modifying this has most probably far too many (and large) side effects..