Structured grid generation is admittedly cumbersome and time consuming, however, the time spent optimizing the mesh is well rewarded.

The best structured solvers are faster, more accurate, more robust, and use less memory than the best unstructured codes. Also, because structured codes have been around for so long, they tend to have the most physical models.

Furthermore, post-processing tasks are far easier on structured grids. Topological planes within the grid are ideal locations for generating qualitative plots and calculating quantitative values such as mass flow.

A few of these benefits will be overcome by unstructured methods through further development, but many will not because they are fundamental obstacles to the unstructured method. For example, unstructured methods will always require more memory than structured methods because of the need to store the connectivity information.

Having said all that, are you using structured methods primarily? Will you ever switch to unstructured methods? Should we even be working on unstructured methods? Would the time not be better spent reducing the cumbersome aspects of structured methods?

These questions are seemingly answered by CFD developers (both commercial and academic) who spend the bulk of their resources developing unstructured methods.