I have been comparing the results obtained with a mesh generated in TGrid and one in Gambit for an aircraft fuselage and for some reason there are large differences in the predicted drag co-efficient, as well as the number of cells needed in the domain. For both volume meshes an identical surface mesh (formed in Gambit) was used. Around the fuselage a 5 cell boundary layer was generated with a growth rate of 1.2 and an aspect ratio of 5. The initial cell size is such that the y+ value ranges between 15 and 250 so this is in the range of the standard wall functions. For the Gambit mesh the max skewness was 0.785 and for the TGrid one 0.75. The first point to note was that the volume mesh from Gambit comprised 2.9MM cells whilst that from TGrid was 1.89 MM. For this comparison use was made of tet cells. The prediction of the viscous drag from both meshes is the same whereas the pressure force (Z direction) differs markedly. For both simulations the reference values were identical and also in the X and Y directions the pressure force was the same, the only difference was in the Z direction. Has anyone seen something similar?

You have used the size function in gambith? Size functions may change number of cells.

I did but then again I used a refinement level of 9 in TGrid that according to the Fluent best practive for external aerodynamics should also give a growth rate through the domain of 1.2, which is the same as the sizing fuction in Gambit.

I think this is not same but near same. You problem lie in mesh dependensy. You must refine mesh.

Thank you for that. I am trying local refinement in TGrid around the fuselage body. Fingers crossed.

You are fuzelage with wings or mithout? If yes then mesh must be 10 or even more millions.

Currently without wings and tail area. They will be added shortly. Wanted to define a reliable meshing strategy before adding these items.