Towards an Output-based Re-meshing for Turbomachinery Applications

The truncation error estimation methodology using topologically inconsistent fine and coarse meshes was presented. Estimated truncation error was weighted with an adjoint solution to obtain a robust output-based adaptation sensor. Re-meshing using Boxer and output-based sensor field was successfully applied to the simple cube test case showing almost an order of magnitude cost function error reduction as compared to the uniformly refined grid. More application examples are required including a more realistic turbulent cases e.g. turbine stator in order to investigate how useful is the methodology in practice. The key challenge for viscous flows is related to the treatment of boundary layer when performing re-meshing with Boxe