Learning Forum

[KORAY2255]

I conduct a conjugate heat transfer analysis in a diesel engine to see the heat transfer phenomena around the block and cylinder heads. In this analysis solid domain is solved in ANSYS Mechanical - steady state thermal by using the near wall temperatures and related HTCs exported from FLUENT where the fluid domain is solved by using the wall temperatures as boundary conditions exported data ANSYS Mechanical via two way System Coupling. About calculating/extraction of local heat transfer coefficients in Fluent I have read all the posts shared in this forum and elsewhere but I couldn't get a satisfactory explanation. I know to introduce a reference temperature to use Fluent's surface heat transfer coefficient but it is not possible to represent the bulk temperature as a single value in such an huge domain. And if it was a straight pipe it would be better to crate as much as iso-surface and get wall and bulk temperature and calculate the local htcs. But it is also not practical in this problem since it is a complex geometry. So I am very confused how to define the local htcs, maybe an UDF but how? Or, are there any practical and correct way?nThank you in advanc.n

[Rob]

I'd cheat and model the solids in Fluent too. Then transfer the temperature to Mech to see if it bends/stretches/twists. Much easier (as a Chemical Engineers solids are only there to heat up, cool down or stir things) and means you're transferring real data as opposed to a derived value (HTC). n

[KORAY2255]

Thank you for your comment. To clarify:nThe reasons why I do like that is : Too much computational cost occurs when I include the solid into Fluent and it may be impractical since the timescale between fluid and solid is large. And I am using Mech just to get wall temperatures not for another purpose like bending/stretching/twisting. nSo, to conclude I need to extract the local HTCs from Fluent. Which methods do you suggest?n

[KR]

Hello,nI agree with Rob. It seems like you are solving a steady state problem (I'm mainly basing it on the fact that you are transferring data to Ansys Steady state Thermal, please correct me if I'm wrong). You are better off with solving everything is Fluent and transfer the final set of temperature to understand the final deformation due to these thermal stresses.nRegarding your other question on local HTCs - yes, you are correct. For such problems, you might want to use the local HTC as well as the wall adjacent temperatures as well. This might help you obtain the net flux on the Fluid side. When you apply this flux, you should be able to solve for the final thermal field in Ansys Mechanical and subsequently solve the deformation. Please see if this approach works for you.nThank you.nKarthikn

[Rob]

To add, if you take the htc etc from Fluent's wet wall to Mech for the solid you're also assuming an external boundary condition for Fluent that may not be correct (eg wall thickness) so you'd ideally want to transfer data back and forth a few times. n

[KORAY2255]

Yes it is a steady-state analysis. Thank you for your suggestion on using heat flux values. This is applicable to my problem. But it would be better to see the local HTCs to make a comparison with literature. As I see there is not a pracitcal and always-aplicable way to get these HTCs. Maybe in the future releases it could be inclluded similar to another commercial software adapted as y+ specified heat transfer coefficient. Here is a link:nArrayDo you think that it can be also applied in Fluent via an UDF?.

[Rob]

You mean like this option? nn

[KORAY2255]

Yes definitely what I metioned. Wow. In my release (19.3) it doesn't appear. I think this is the final release (20.2), Isn't it?n

[Rob]

It's the current release, 2020R2. n

[KORAY2255]

Thank you, I think it will very helpful..n