I am currently contributing in the testing for a mesh adapter release. Mesh adaptation is a computational feature that can be of significant value in many CFD cases. The initial grid is refined based on an initial solution and an improvement in accuracy can be observed at the regions that are computationally more demanding. The application though is not as simple as it seems, as mesh adaptation relies on smart algorithms that should be able to detect the regions that new grid points should be added (and regions that grid points should be removed!).
Below a case that was examined which regarded an adaptation around shock waves formed in a supersonic duct.