I am attempting to create a mesh for an entire aircraft and am looking at trying to combine both structured and unstructured grids into it. I am looking to use the unstructured grids around the intersections that also have complex curvature. Has anyone had any luck either creating a structured grid at any intersection; like between a wing and the fuselage of the aircraft; or has anyone successfully built a grid that involved structured and unstructured meshes?

Hi Shane,

I am currently simulating DLR-F4 wing-body using several hybrid meshes of different densities. The technique I used is exactly the one you are talking about: employ unstructured surface mesh on almost all surfaces and structured in the most difficult regions (wing-body junction, leading and trailing edges). Then apply B-L mesh with 6-10 layers (y+ in wall functions range, over 30 and under 100), and fill the rest of the domain with unstructured tetrahedral mesh. Minimum mesh size for quantitative results is around 2 mill. elements, but for good results you must go over 5 millions (with at least 1 mill. elements in the B-L region only!).

So the answer is: yes, you can make a quite good hybrid grid for aircraft-type geometries, but you must be very careful about mesh quality especially in the B-L region! Drag is particulary sensitive to near-wall mesh quality.

All the best, Razvan

we also use meshes generated in parts. This is my workflow :
[*] make the the meshes in parts such that the mesh on the sharing faces (with other blocks) is exactly same.[*] use tmerge to merge these meshes into one.[*] in fluent, there is a option Grid->Fuse, use this option to fuse the sharing faces of the mesh.[*] save the case[*] import the mesh from this saved case[*] export the final mesh to be used in simulation.

hope this helps , you can avoid the gambit import and export again, for me i do this because this way the mesh is written as a single block, and i prefer that.