I am solving conjugate heat transfer problem in Ansys Fluent. The system is like a cylinder with 4 layers of solid (one over other) and fluid inside it. The the inside surfaces of the solid has radiation. As one of the layer is too thin, I am getting very poor skewness for this particular layer. When I am solving, I am getting "Temperature limited to 1.000000e+00 in x cells of zone 1 in domain 1"
This results in very unrealistic temperature distribution in the solid.
The problem is transient. I am using S2S model for radiation and K-e for turbulence. I am using adaptive time stepping.
Please suggest any solution.
Check your bondary conditions and patch your temperature in all field fluid when inntlized calculations..
@ Lalmi
My boundary conditions are
1) Velocity inlet of air at room temp
2) pressure outlet at 1 atm
3) internal heat generation in the solid body
4) convection and radiation to atm from outer boundary
I think that your boundary conditions are right , you can check the values of the parameters when you input it in the fourth condition convection and radiation .Check also you grid and patch the temperature and initialized your problem .thanks .
it is better to reduce relaxation factor (residuals) to avoid these error
Hi,
if you report about "bad skewness" for the very thin material layer, than I start to get courios, whether you nevertheless try to mesh this region with tetrahedral cells? This would result in what you report about and obviously lead to solver divergence.
So if this should be the case, than I would urgently recommend to either use the ANSYS Meshing Sweep method for the generation of a hexahedral mesh in this thin material layers. Or right away start using ICEM/CFD to generate for this cylindrical shells a good quality hexahedral mesh. This would definitely help to obtain a physically reasonable result. But maybe my impression is wrong and the reason for your observation is entirely different.
Regards,
Th. Frank.
Thanks Ravindra D. Jilte
Reducing Under relaxation factor has worked.
Thanks Thomas Frank
I shall try meshing the way you suggested to check the accuracy of the solution.
URF=1.e-15 is not very useful. It essentially means, that you artificially freeze the flow in the state of your flow initialization. Nothing is allowed to change.
Chandae Jeong
Can you explain your problem in detail?
I mean the physics.
Hi Hardik,
It`s good to hear that change in URF has solved your problem.
I am trying to solve similar problem hollow solid box(per say) with air inside and outside of box. I am adding surface heat flux on external box surface.
I am interested in observing temperature distribution through box and fluid domain. So what solver and gradient is suitable. Should I also consider adaptive time stepping?
Thanks in advance
Hi Jitendra Seregar
Adaptive time stepping is always helpful because it lets you use smaller time steps during rapid changes in the system and larger steps, once system is "almost"stable. And thus reducing overall simulation time.
You can use any solver like Fluent, CFX, OpenFoam for this problem.
Hope this helps.
- Badges
- Science topic