[jack1980]

Hi,

I'm trying to get adjoint based sensitivities for my design parameters, however, I would like to use self-defined shape deformation basis functions.

To do so, I deform the mesh "by hand", then run the direct solve and adjoint solve. I thought I should then be able to compute the sensitivities from the information in surface_adjoint.csv. This is my first attempt for a deformation on the top of the wing (I should probably take into account that 'senstop' is normal to the surface later):

data = np.loadtxt( 'surface_adjoint.csv' , skiprows = 1 , delimiter = ',' )
xtop = data[102:-1,6]
senstop = data[102:-1,1]
return np.trapz( senstop*basisfunction(xtop) , x = xtop )

This seems to give the sensitivity, up to a (negative) factor. What am I doing wrong here?

Thank you for your help!