[feldy77]

Dear forum,
I am really confused about how I should compute heat flux through the static boundary in the presence of turbulent incompressible flow if I use the basic k-e model. According to the heat transfer theory
Q''=(k_lam+k_turb)(dT/dn), where k_lam is a some constant value of a certain fluid and k_turb=cp*miu_t/Pr_t. It looks like everything is clear : cp is known, the commonly used value of Pr_t is 0.85, miu_t is calculated from k-e model. The twist is that the boundary value of miu_t (niu_t for incompressible flow) is set to zero (I am talking about buouantBoussinesqPimpleFoam tutorial for hotRoom). The same is done for the boundary value of miu_t in Fluent. Then it is completely unclear what is the difference between calculating of heat flux for laminar and turbulent flows.
I will be very grateful if somebody clarifies this issue,
many thanks, Yuri

[feldy77]

please anybody can say his opinion,
I really need your help

[kaifu]

k_turb

Originally Posted by feldy77

Dear forum,
I am really confused about how I should compute heat flux through the static boundary in the presence of turbulent incompressible flow if I use the basic k-e model. According to the heat transfer theory
Q''=(k_lam+k_turb)(dT/dn), where k_lam is a some constant value of a certain fluid and k_turb=cp*miu_t/Pr_t. It looks like everything is clear : cp is known, the commonly used value of Pr_t is 0.85, miu_t is calculated from k-e model. The twist is that the boundary value of miu_t (niu_t for incompressible flow) is set to zero (I am talking about buouantBoussinesqPimpleFoam tutorial for hotRoom). The same is done for the boundary value of miu_t in Fluent. Then it is completely unclear what is the difference between calculating of heat flux for laminar and turbulent flows.
I will be very grateful if somebody clarifies this issue,
many thanks, Yuri

I dont see any problems inside. As you said, Q''=(k_lam+k_turb)(dT/dn). if k_turb=0 then Q''=k_lam dT/dn.

And the temperature relies on the turbulent model you included although you will have the same gradient at the wall. I am sure temperature in the bulk is totally different between laminar and turbulent flows due to the different heat transfer coeff in the bulk.

[feldy77]

Kai,
Thank you for your reply. Maybe you can explain me then, why the openFoam utility wallHeatTransfer utilizes alafaeff whcich consists of turbulent amd laminar components and not jus uses the laminar component only
Many thanks, yuri

[kaifu]

Originally Posted by feldy77

Kai,
Thank you for your reply. Maybe you can explain me then, why the openFoam utility wallHeatTransfer utilizes alafaeff whcich consists of turbulent amd laminar components and not jus uses the laminar component only
Many thanks, yuri

It is all dependent on the expectation you have. The wall viscosity should be modelled actually. I can only say that the treatment of zero wall viscosity is rather crude. If you would like to model wall viscosity, plz refer to the file "wallViscosity.H" in the solver of bubbleFoam.

[Goutam]

Originally Posted by feldy77

Dear forum,
I am really confused about how I should compute heat flux through the static boundary in the presence of turbulent incompressible flow if I use the basic k-e model. According to the heat transfer theory
Q''=(k_lam+k_turb)(dT/dn), where k_lam is a some constant value of a certain fluid and k_turb=cp*miu_t/Pr_t. It looks like everything is clear : cp is known, the commonly used value of Pr_t is 0.85, miu_t is calculated from k-e model. The twist is that the boundary value of miu_t (niu_t for incompressible flow) is set to zero (I am talking about buouantBoussinesqPimpleFoam tutorial for hotRoom). The same is done for the boundary value of miu_t in Fluent. Then it is completely unclear what is the difference between calculating of heat flux for laminar and turbulent flows.
I will be very grateful if somebody clarifies this issue,
many thanks, Yuri

Dear Yuri, did you able to solve this problem? I am facing same problem.
Looking forward to see your reply.

[niaz]

Dear Foamers
for k-epsilon model, you must use wall-function for nut.
and without it, the solution is not acceptable.