I am trying to check the accuracy of the adaptive solver . I have implemented a 2D lid driven cavity problem but my observation is that the mesh is basically re - refined in the same mesh elements and more specifically in the center of recirculation and in the position of the lid. I mean that the adaptivity is working but there are no new marked cells for refining.Is this normal ? Is there any parameters for marking strategy or adaptive fraction?
I am trying to test the adaptivity on 3D mesh in a Beltrami flow test problem. What should the goal be for the error indicators? When i implement a goal of u^2*dx it seems to work but when i reduce the tolerance, the adaptive iterations never end meaning basically that the error estimate is never lower than the tolerance. Is there any solution to this?
Thanks in advance!!