[lukas']

Hello everybody,

currently I am working on my thesis for which I did some Fluent simulations. My studies focused on the temperatur-profile in an installations space. The heat-source is in the middle and the components are symmetric. In the results the temperature-profile is not 100%-symmetric.

Can anyone tell me possible reasons and maybe a source that I can cite for my thesis?

Thanks in advance!

Best wishes!

Lukas

[AlexanderZ]

how could you know, that it's not symmetric? how did you check it?

[lukas']

I expect that because of theoretical thoughts.
In my view there is reason why it should not be symmetric.

[karachun]

lukas' You should bring to us more information - physical description, computational (fluent) description, also it will be good if you give us your thoughts why flow is unsymmetric.
At this point we are talking about "black box". For black box I can give two general assumptions.
1. Your mesh is too coarse.
2. Flow have transient nature.

[lukas']

Originally Posted by karachun

lukas' You should bring to us more information - physical description, computational (fluent) description, also it will be good if you give us your thoughts why flow is unsymmetric.
At this point we are talking about "black box".

There is no external flow. It is a heater with a constant temperature in the middle of an installation space. It is surrounded by cylindral components. All the components are symmetric. I used the SST-model and S2S-model for radiation.
I don't see a reason why the temperature of a part of a component should have a different temperature that the same part that is on the other side of the heater with the same distance away. We talk about ~10°C at temperatures between 20°C and 500°C. So the diffenrence is not really big. I hope this helps to understand my problem in a proper way.
What I actually want to know if the difference might come from the meshing or the solution algorithm,... I checked mesh quality - it should not be the problem.
Sorry if my information was not complete yet but I'm a beginner in CFD... I hope this helps now.
Thanks for your answers!

[AlexanderZ]

from my experience, convection problems sometimes shows this asymmetric solutions.

you may try:
1. use other turbulent model
2. decrease residual levels
3. increase order of spatial discretization (in solution methods)
4. initialize the flow symmetrically using PATCH tool (which may be tricky), easy approach is to initialize case with small flow velocity value from bottom to top, and calculate
5. ofc mesh could be a problem, try to make new mesh, most likely you will get new result
6. you may switch from steady-state to transient (but it is computationally expensive)