Repository navigation
The 2D case's (1-cell thick meshes) adjoint equation can't converge #2644
Description
I've encountered an issue while using the SU2 continuous adjoint method:
When running the OAT15A 2D case (with 1-cell thick meshes, the two faces perpendicular to the thickness direction are both set as symmetry planes, please see OAT15A_mesh.jpg), the direct equations can converge, but the adjoint equation residuals remain unchanged. The adjoint velocity is shown in OAT15A_adjoint.jpg. When running other 3D cases or pure 2D cases (without thickness), both the direct and adjoint equations can converge.
Does anyone know why this happens? Is it because SU2's continuous adjoint calculation inherently does not support 2D cases with 1-cell thick meshes?
The mesh and configuration files for this case have been attached in the appendix. Try running the adjoint solver only for the first optimization iteration. I haven't added any parameter settings for subsequent optimization iterations in the .cfg file.
Activity