Wall y+, heat transfer coefficient
 

Hello

I am investigation heat transfer from a flow through a pipe with solid. The fluid and the solid are initially at 900K and the flow is at 900K.
I have a mesh for the fluid, and one for the solid (1mm thickness).
The fluid mesh has 5 fine elements next to the solid.
My flow velocity is about 20m/s, and the fluid is air.
I am using k-epsilon model

For the heat transfer of the solid to the environment, I am using 'convection' as thermal specification, with 10 W/m2K for heat transfer coefficients and 300K for ambient temperature.
For the interface fluid/solid, I let the default specification of CCM+

At the end of the computation, I am checking the 'local heat transfer coefficient' at the interface fluid/solid (I have read in the manual that the local definition is used there).
I am getting quite high value for the coefficient (around 220 W/m2K)
When I am checking the local reference temperature, the temperature range between 900 and 840K (frm inlet to outlet of the pipe), corresponding to the last layer of fluid at the vicinity of the solid.

My first question is:
1/ Does the reference temperature should be the one at the vicinity of the solid? I thought this temperature should be the bulk one? (which is then roughly 900K)?

I have then investigated the y+, giving a value of around 5 (taken at the fluid/solid interface).
My second question is thus:
2/ I am wondering if I should coarsen the mesh next to the solid, in order to achieve a larger y+ and get a reference temperature nearer to the bulk one.
Does it sounds ok to someone? (I will investigate it, but I wanted to have some external point of view on that)

3/ Finally, is it possible to force CCM+ to compute the local heat transfer coefficients using the bulk (or mass flow average) temperature?

Thanks a lot for your tips on this mater
Best regards