[fshak92]

In the two pictures you can see the same simulations but with the difference in Gravity.

The temperature distribution(Can be seen in plane section) in the simulation with gravity is so strange and unrealistic. the mesh quality and iterations are fine and enough(about 500 iteration with good residual)
-Segregated
-Ideal Gas(incompressible)
-SteadyState
-Laminar
-Air
-One of them with gravity,the other one with no gravity.
There is heated body(26'C) which surrounded by a block with 6surfaces(21'C).

Is there anobody know the reason behind the unrealistic distribution with gravity? because i cannot ignore gravity in my simulation...
Thank you in advance

[fshak92]

**According to the gradient of temperature,using boussinesq model with constant density is ALSO valid, but i'd like to know why using ideal gas cannot give me a realistic distribution of temperature.

[fshak92]

**Also by using the boussinesq model with beta=0.003 and reference temperature of 23 (even with a very finer mesh!) for both Segregated and Coupled solver the residual of my momentum equations goes further than 1.

[fshak92]

Has somebody any idea about the mentioned issues?(i used about 19milion cells and i think the problem is not associated to meshing)

[siara817]

Hi Omid,
I think in the case of simulation without gravity, the temperature cascade is mainly because of radiation. While in the other case i.e. when you add Gravity, the convection is added to system.

how is your residual? how far has been converged?
Good luck

[fshak92]

Originally Posted by siara817

Hi Omid,
I think in the case of simulation without gravity, the temperature cascade is mainly because of radiation. While in the other case i.e. when you add Gravity, the convection is added to system.

how is your residual? how far has been converged?
Good luck

Thank you for your reply to this post and another new post about this simulation.
I think the problem was occured because of my mesh quality and mostly number of iteration.I increased the mesh quality and the number of iteration and the strange result for gravity has been solved.
but the accuracy of results is not reliable which i wrote a new post about it...
(just for the people with same problem:
http://www.cfd-online.com/Forums/sta...imulation.html