SOS ! SU2_GPC gradient projection with other geometry parameterization
 

Hi Stanford Guys

I am working on combining the SU2 with an open source geometry parameterization suite GeoMACH, developed by MDO lab at University of Michigan.

So, I have to project the gradient of the objective function respect to the surface grid to the direction of the gradient of the surface grid respect to the design variables within GeoMACH.
Now I am diving into the SU2_GPC code, but I do not understand it very well.

1. in SU2_GPC.cpp, there are variables such as Normal, VarCoord, Sensitivity, deps, delta_eps and dalpha , what do they stand for separately ?

2. I noticed that the Sensitivity variable is a scalar instead of a 3D vector, why is that, cause I thought the gradient or sensitivity should be a vector instead of scalar.

3. Finally , how can I get the derivative of the objective function respect to the very surface grid of the geometry to be designed?

Thank you guys
Looking forward to response.

So, are the Sensitivity variables normal to the geometry shape which is to be optimized ?

Yep,

The mathematical formulation is based in the fact that, if you don't have geometrical singularities, just with normal displacements of the surface you can achieve any new geometry (a tangential displacement is nothing more than a normal one plus a re-parametrization).

Best,
Francisco