Implementation of algebraic turbulent heat flux modells
 

Hello,
I am currently trying to implement the Daly & Harlow and the Younis et. al. modell for the turbulent heat flux in STAR CCM+. I want to use them in combination with the standart cubic k-epsilon turbulence model.
In STAR CCM+ the turbulent heat flux is calculated with a turbulent Prandtl Number by default.
To improve this calculation, I want to implement both models which I stated above.
Therefore I first set the turbulent Pr-Number to a very high number (10E10) to supress the default term for the turbulent heat flux.
Then I implemented the Daly & Harlow modell as a vector field function and chooses its divergence as a volumetric heat source in the energy equation.
Unfortunately the calculations do not converge with these modells acting as enegry source terms. Unnatural high and low temperatures appear in the field and I can't gain a converged solution.
I tried underrelaxing the energy and momentum equations, but that did not really help. Further, I tried to calculate the first 1000 iterations without the new source term and then switching it on slowly (by a blending function) but that did not help either.

I enclosed a short PDF to illustrate my problem better and to show the relevant equations.

Did anybody of you tried to do something similar and was succesfull ?

I would really appreciate any help.
Thanks.

Best regards,
Thomas

I maybe terribly mistaken, but I believe Star-CCM+ has had a few newer forms of turbulent heat fluxes implemented in recent versions. Not sure if you can use it with the cubic std k-e, but it maybe what you're looking at. At the time I last looked, it was only available with single phase.

Yes, there is a modell that works with the std k-e but you can only use it for the case of constant density.
Unfortunately I can't asume constant desity in my case.

I've had the same frustration when I found out it wouldn't work with the multiphase modeling. Best of luck!