[silviliril]

Hello everyone,

I am Working on conjugate heat transfer problem. The conduction and forced convection takes place from the surface of the air cooled engine FINS to the environmental air. The Fluid domain is made around the engine geometry.

Which one is the best turbulent model for solving forced conjugate heat transfer problem and why (standard k-e, or Enhanced wall treatment k-e, or SST, K-w) ??

The picture of the engine body and air domain is attached.

[scipy]

I'd go with the Realizable k-epsilon and Enhanced Wall Treatment. As far as the why goes: it's the most widely used in the industry and has been most validated/optimized. Since your geometry is pretty blunt (it's not an aerodynamically smooth wing or a streamlined body) RKE should give you good results. You may also want to try k-omega SST which would be aimed more at streamlined bodies where RKE fails to predict flow separation correctly.

As far as wall treatment goes, use EWT (for RKE, for k-omega it's the default wall treatment anyway), and try to play with different meshes (i.e. refinement in the normal to wall direction). EWT was made to handle y+~1 meshes, but it can also be used on wall function meshes (30<y+<300). If you want to get reasonable results for heat transfer though, try to make a mesh with a lower y+ value (as far as your resources will allow) and fit at least 10-15 prism in the boundary layer for the wall-function mesh and up to 30-40 for the y+~1 mesh.

[silviliril]

Thank you.. Your tutorial seems helpful for meshing.

[scipy]

If you've watched the tutorial, don't use the First Aspect Ratio method to grow your prisms. You need more layers and more delicate control over the height of the first layer, so either use the height of the first layer and then have a 1.2 growth factor and 10+ layers, or use First Aspect Ratio but increase it to 20 or more instead of 5 (it's going to compress the first layer further down).

Use the NASA grid spacing calculator to check how high your first prism layer should be to get the wanted y+ (both for 1 and for 30+).