Do you have any idea about the influence of inflation layer and structured/unstructured mesh? I found there is some effect of those on the final solution, based on the calculation with the whole same settings, excepting the mesh.
The above figure shows the axial velocity of pipe bend exit. The black lines and blue lines represent unstructured and structured mesh respectively. Clearly, different shape can be observed.
Also I compared the results with different inflation layers setting. All of them have the same first layer thickness and body size. I only changed the number of layers or growth ratio.