[AndreP]

In my current workflow I use RANS+LES to run unsteady simulations.

Recently I started using AdjointOptimisationFOAM to help me with ideas for optimizing the shape (vehicle). My current procedure for adjoint is running between 5 to 10 iterations on a local part of the vehicle and hopefully be able to observe a clear trend of mesh movement that it's advised. I then morph this in the initial geometry, export and run RANS+LES with the normal workflow.

Problem is that this ends up being time consuming and a lot of times, since adjointOptimisationFoam is running RANS, the LES result is underwhelm.

Would anyone have any pointers or ideas, on how would I go about running a 10 iterations loop adjointOptimisationFOAM, and at the end, pickup the morphed mesh, and directly launch it in my normal workflow RANS+LES?

I replicated this manually by downloading constant, then merging the "points" generated in the 10th iteration folder and launching the case like that, but since I'm using a cluster, would be very usefull to be able to do this automatically and at once!