[eling]

Hello guys,
I encountered a problem when simulating the heat transfer of a turbine. Please help me out.
Problem description is here:
The rotor is made of two marerials, lower part is rubber which is a heat source, upper part is steel. There are two fluid zones and two solid zones, ie the lower rotor solid zone, the upper rotor solid zone, the rotor fluid zone and the stator fluid zone. The rotor rotates in a constant speed wrt the stator. I use MRF for the ratation. Since the flow is periodic, I split a quater of all the zones for simulation and use periodic bcs which inludes faces of both fluid zones and solid zones (the rotor is cut). I link the adjacent faces of the two fluid zones with interface without coupled wall check on. And for connection of the fluid-solid zones and solid-solid zones, I use interface with coupled wall checked on. The lower rotor solid zone is heat source. Inlet and outlet are pressure inlet and outlet.
After the calculation is converged, I found the heat flux of interface btw the two fluid zones is zero. And the heat flux of inlet and outlet is very small compared to the heat source. This is impossible bcz the energy is not conservative. However, the residual of energy is below 1e-6. I am confused about it.
I don't know if I have given a sufficient description. Sorry I cann't upload the result bcz I am not in office.
Your help is highly appreciated.
Eling

[AlexanderZ]

Does fluid flow look good?

Fluids have different material properties? Why do you use interface on the boundary, not interior? In case of interface why don't you use coupled setting?

Best regards

[eling]

Originally Posted by AlexanderZ

Does fluid flow look good?

Fluids have different material properties? Why do you use interface on the boundary, not interior? In case of interface why don't you use coupled setting?

Best regards

Dear Alexander,
Thank you very for the reply. The fluid zone material is water and the flow field seems reasonable but not the heat flux across the interface.
I use interface btw the two fluid zone bcz I want to separate the rotor and the stator for MRF method, and not coupled wall interface bcz there is no real wall btw the rotor fluid zone and stator fluid zone.
Do you have any suggestions on my solution to this problem? Thx.
Eling

[Fole]

I do not consider myself the best person for assisting you in this, but in CHT that situation occurs (sorry, I remember a thread in which this is discussed but I don't remember it right now).

Residuals can be low with a field not being the actual solution (if they were = 0, then it would be the solution, but even below the low limits, some fields also fits the criteria).

You have to iterate and monitor for energy conservation or temperature stabilization. I think you should see how energy is more well-behaved as iterations proceed.

Cheers,
Fole.

[eling]

Originally Posted by Fole

I do not consider myself the best person for assisting you in this, but in CHT that situation occurs (sorry, I remember a thread in which this is discussed but I don't remember it right now).

Residuals can be low with a field not being the actual solution (if they were = 0, then it would be the solution, but even below the low limits, some fields also fits the criteria).

You have to iterate and monitor for energy conservation or temperature stabilization. I think you should see how energy is more well-behaved as iterations proceed.

Cheers,
Fole.

Hi Fole,
The energy residual is very low at the beginning say 1e-7. As iteration proceeds not very long, it enlarges and stay at say level 1e-6, which seems converged. But the result shows there are spots at the wall where the temperature is lower than environment. This is impossible since there is heat source in the domian. Any suggestions? Thanks.
Eling

[AlexanderZ]

Hello,

try to plot contour of temperature and switch of Node Values (in Contours menu)

Do you still have "problem" elements?

Best regards

[eling]

Originally Posted by AlexanderZ

Hello,

try to plot contour of temperature and switch of Node Values (in Contours menu)

Do you still have "problem" elements?

Best regards

Hi Alexander,
Yes, some cells are still below environment temperature. I have a picture, but I don't know why it cannot be attached. These cells are mostly located at the downwind periodic face (solid heat source domain) and there are spots on the face with very high temperature.
I tried not to use boundary layers(prism) in the solid domain and refine the mesh on the face, but it doesn't work. The solid heat source domian is adjacent to both solid domain and fluid domain. I suspect if fluent could deal with the interfaces of two solid domains with different materials.
Besides, the solid heat source I give is 3000W/m3. However, when I check the total heat tranfer rate of the faces surrounding the solid heat source domain in the flux menu, the result is much higher than 3000W/m3*the volume. I don't know why not equal. Any suggestions? Thank you.
Ivan

[litzj]

Although I am not good at heat problem

when you made interface zone, is there any shadow zone? for interface

If it is, you should care about that shadow b.c

[eling]

Originally Posted by litzj

Although I am not good at heat problem

when you made interface zone, is there any shadow zone? for interface

If it is, you should care about that shadow b.c

Yes, there is shadow wall. How should I care about it or any settings for it?
Thanks,
Eling

[litzj]

Originally Posted by eling

Yes, there is shadow wall. How should I care about it or any settings for it?
Thanks,
Eling

this might be help you

http://www.afs.enea.it/project/neptu...ug/node250.htm

find Thermal Conditions for Two-Sided Walls part