Learning Forum

[gioeleg]

Dear all,nI am simulating air diffusion inside an EDM machine (in order to do a thermostabilization) and I would like to separate the Fluid and the Thermal simulation.nTherefore I thought to simulate the air first, and then from the overall diffusion and velocity profile of the air near the walls, calculate a convective heat transfer coefficient and then give this one to the thermal simulation.nMy problem is: how can I evaluate the velocity profile near the walls? Clearly the velocity at the walls is zero, and the internal geometry of the machine in complex, so creating plane at some near distance from the wall and then plot the velocity there is not doable.nThank you very much in advance.nGioelen

[DrAmine]

Have a look into the heat transfer coefficient description and you will then understand that you require that coefficient + a reference temperature to get applied for your thermal analysis run. I genereally recommend the Yplus based HTC as well as Yplus based reference temperature.n

[gioeleg]

Hi DrAmine,nThanks for your answer and sorry for replying so late.nI am not so sure if I well understood your advice. Maybe I did not explain myself clearly.nWhat I would like to do is a Fluent CFD simulation without any energy equation first. Then from this I would like to be able to obtain some heat transfer coefficient, in order to do a thermal simulation with Mechanical. Is this possible/reasonable? Does it make sense what I would like to do?nOr (in order not to mesh both fluid and solid domain) should I enable energy equation and take the wall function htc? If so, could I export this wall function htc profile to a thermal sim on mechanical?nAgain, thank you very much in advance.nGioelen

[Rob]

If you have a correlation for velocity and HTC it's not a bad idea. nJust to point out, if you're doing a static structural too then compute the temperature in the solid in FLuent then transfer the temperature in the solid region to Mechanical. Saves all the uncertainty regarding HTC and whether the part heats/cools differently as the fluid temperature evolves. n

[DrAmine]

Yes you can do that: Velocity Information->Reynolds Numbers + Prandtl number-> Heat Transfer Correlation (using an analogy) but I recommend to do that in one step! (as the velocity scale is not uniform and for that reason we do CFD) and one has in general a two way coupling between flow and thermal effects (buoyancy, material properties changing with temperature...). First quick estimation is sufficient.n

[gioeleg]

Thank you both for your replies!nDrAmine, what do you mean with the workflow Velocity Information->Reynolds Numbers + Prandtl number-> Heat Transfer Correlatio?nCan you please explain further?nThanks.