[Hkn]

Hi everyone,
I am trying a cooling analysis which a hot solid part is passed through a cold air (like fan) on like conveyor band it's all of in air domain. I have tried many analyzes since 3 4 months maybe more time. I need the actual heat transfer and the display of the final temperature of the solid (Contor in solid part). I have been run moving solid body at rigid body dynamic. But I can't get solution to transfer temperature in solid body. I tryed system coupling (transient thermal->transient structual and cfx) but its not give permission temperature fields. I tryed again solution together transient thermal->transient structual part and I transferred cfx but I get error read when solution. I'm thinking of transferring the results of the cfx to the transient thermal, but this time there will be no solid moving.
Do you have suggestions, thank you.

[evcelica]

Why do you need the structural part of the analysis?
Can't you just use the solid pieces reference frame, and consider air motion relative to that? Pics of what you are trying to do would be helpful.

[Hkn]

Thank you evcelica your reply.
I defined the motion there using FSI and experimented by combining it with thermal analysis. I tryed this my last trials because I can't move solid domain with solid motion tab , interface is ok its moving, in cfx.

[IMG] https://ibb.co/g27G7K [/IMG]

[evcelica]

And the X ends are just openings? To ambient space?
Easiest way to do this would be to keep all solids stationary, and just change the position of the inlet through time.
Make the inlet box longer in the direction of travel, and use a source term instead of an inlet.

FluidVelocityIn = if(abs(XFanCenter-x)<=(FanWidth/2),FlowInVelocity,0[m/s])
where xFanCenter would be a function of time, moving in the X direction. Now fluid would come out of the inlet, which would be moving relative to the bar and convener.

Or you could also do the solid movement in CFX alone, no structural needed, using rigid body motion, (with mesh deformation), and your mesh would stretch.

Or you could move the whole domain, and your fan inlet would again be a source term that is now stationary. Or you could have a fan box that doesn't move, and use a domain interface between the two. No mesh stretching/remeshing would be needed here. You could also move the fan box as well, same thing, just different reference frames.

Or you could do several steady state models with the solid at different positions along the path, and calculate the heat transfer at each position, curve fit that, then just do the math to solve for the transient temperature. Then you could use that and easily adjust the conveyor speed to whatever you wish.

Any of these methods should work, it's up to you which is easiest, and most appropriate for your assumptions. There are probably other ways to do this as well, these are just some quick suggestions I thought may be plausible.

[Hkn]

Thank you very much, I am set analysis set picture other all wall is adiabatic.
Also, I want to asking several questions.

In your post for you could move the whole domain I thinked whole the air domain

you could have a fan box that doesn't move, and use a domain interface between the two, I didn't understand clearly, can you explain that little more ?

I set up two domains in CFX which solid and fluid. So rigid body could not connect to the solid domain and did not move at solid body in results. Although solid movement box is active.


https://image.ibb.co/kmP9CK/tempsnip.png
https://image.ibb.co/nEL7XK/Fluid_Flow_CFX.png

[Hkn]

I used total time 10 s and time step 0.001

[ghorrocks]

If the motion is known in advance this is not a rigid body or a FSI simulation. It is a moving mesh simulation. The heat transfer can be handled by CFX as a conjugate heat transfer model. Note that translating solid heat transfer bodies have only been supported in CFX since version 18 or so. If you have an earlier version this will be much harder.