[ghost82]

Goodmorning all!
I'm studying a problem of heat transfer in 2 adjacent rooms (air inside) separated by an insulation wall (domain dimension: heigh=9 m, width=36 m).
Left room has an avarage temperature of +5 ˚C, right room is at -22 ˚C.
Inside there is air and Boussinesq approximation was used.
All the walls are adiabatic, except for the floors which have a temperature equal to the mean room temperature and for the insulation walls which are coupled walls.
My case is 2D.
I tried different models:
- laminar
- turbulent, standard k-e with scalable wall function
- Transition, SST (y+<=1)

I have to calculate heat transfer coefficient for the left room.

The parameter for the choice of laminar or turbulent model is the Rayleigh number (for vertical wall it should be turbulent for Ra>10^8).
In my case, for the left room I have Ra=9,8*10^10, so it should be turbulent.

Results:
- laminar simulation: the problem is not steady, in the left room I have 2 macro vortices and smaller vortices which are created and dissipated.
Area average temperature of the room: 278,126 K
Area average temperature of the insulated wall: 276,447 K
Heat flux, from left room through the insulated wall: 2,534 W/m2

h=1,510 W/m2/K

- Turbulent and transition: the problem is steady; no smaller vortices are created.

k-e model:
Area average temperature of the room: 278,133 K
Area average temperature of the insulated wall: 277,813 K
Heat flux, from left room through the insulated wall: 2,301 W/m2

h=7,199 W/m2/K

SST model:
Area average temperature of the room: 278,133 K
Area average temperature of the insulated wall: 277,829 K
Heat flux, from left room through the insulated wall: 2,530 W/m2

h=8,310 W/m2/K

Literature correlations gives low value of this heat transfer coefficient...what do you think?
I'm attaching some pictures of velocity distribution.

Thank you,

Daniele

[Bionico]

Ciao Daniele,
Could you please show the setup for this simulation (solution methods, solver,...) ?
I usually work on radiators in room: I suggest using RNG-k-epsilon turbulence model (even if the my usual temperatures are different)

Regards

[ghost82]

Originally Posted by Bionico

Ciao Daniele,
Could you please show the setup for this simulation (solution methods, solver,...) ?
I usually work on radiators in room: I suggest using RNG-k-epsilon turbulence model (even if the my usual temperatures are different)

Regards

Ciao Flavio,
sorry for delay in my reply..
I deleted these simulations some months ago as the main simulation was a big warehouse to study air circulation with different fan configurations.
I made the above preliminary simulations to try to find a heat transfer coefficient to input as boundary condition in the main simulation.
However, since I obtained very different results I decided to directly input the transferred heat, calculated with classical simple equations.

As I remeber, for these preliminary simulations I adopted no particular setting, I retained the default ones, changing only methods to second order upwind.

PS: yes, for the main simulation I used RNG k-epsilon model with scalable wall functions.
Also I couldn't obtain a stable solution for the stationary solver and I had to switch to the transient one, obtaining good results.