Hello I'm working on heat transfer "fluid solid" with phase change for solid, I use CFX-5.7.0 In my geometry I have a Fluid-Solid Interface, and when i verify the wall heat transfer coefficient (h) in the result file i find it very high. Even if I raffine mesh with inflation technique in fluid-solid interface. And as consequence of this high (h), I have a wrong result sepcially a temperature field in the solid. 1) How can I resolve this problem 2) I use in CFX-Pre in the interface automatic connection instead of GGI. Does it have an influence? Thank you

Hi,

The solver works out the h value based on the fluid properties and flow conditions at the interface. Therefore you have no direct control over the heat transfer at the interface. A few comments about it:

1) The heat transfer coefficient (h) shown in CFX-Post is based on the fluid temperature being the near wall temperature. You will note that this temperature is usually a lot closer to the wall temperature than the fluid far field temperature. Hence the reported h value is higher than you would get for the same heat transfer using the fluid far field temperature. Most of the literature uses the far-field temperature.

2) If you want a h value comparable to the values in the literature (that is based on far-field temperatures) I suggest getting the convective heat flux at the wall and working out your own h value based on what you think is an appropriate far-field temperature. This number should then compare favourably with literature values.

3) If the h value is still way off then I suggest you have an error in your model. Are your fluid properties correct? Are you resolving the thermal boundary layer accurately enough? Whether you have a 1to1 or GGI interface will not matter.

Therefore, assuming point 3 is under control, that means your simulation should be correct. The high values of h are only because of the different basis of the value.

Regards, Glenn

Use this formula to calcuate h in post (manually):

h=q_wall/(Tw-T_ref)

Don't use the the post 'h' which does not give any accurate results.

Or export q_wall and Tw and do manually in the excell.

MNHK

Thank you for your answers Best regards