[UTA_ARC]

I have a simulation about hot-film probe and has difficulty on accurate mass flow rate data from my result.

The initial mesh is a 3D unstructured mesh neglecting the hot-film. Then in post-process, I created a half cylinder surface in hot-film location to calculate mass flow rate over it. Comparing velocity & density profile, I'm sure the mass flow rate it gives me is wrong. The reason is most likely that the unstructured mesh in hot-film location is not specified and the node/element center may not sufficiently involved in my sampling surface as the film is super small.

To fix this problem, I now try to involve the film in meshing phase. I think I cannot treat the film as solid and its surface as wall because then there will be no mass into it, the mass flow rate will be 0.

What I did is to create the film as fluid and combine with other fluid domain, like a multiblock method. The mesh in film location is then specified and should give accurate data in post-processing as all nodes will be involved in my sampling surface.

However, now the biggest issue is the mesh quality. My case involves inflation layer and I have to create body of influence around film to make the mesh nice. But still, the best I got has issue on skewness(I intend to keep the max. below 0.92 but it didn't) which is not good for solver. The bad cells come from inflation layer to unstructured mesh at the corner.

I attached a few pic to help demonstrate my question. Can anyone give me some advice on achieving my goal? Improvement on current approach or a different way? Thanks!

P.S. I can only think of other ways like (1) treat film as solid, but use 0 specified shear stress on its surface (2) treat film as a porous solid. But have no idea if it would work.
film geo.PNG

film mesh1.PNG

film mesh 2.PNG

skewness.PNG