[Marcus021]

Hello,
I'm trying to optimize the fairing shape of a space launch vehicle. The objective function is DRAG. The shape of the fairing in my mesh is modelled through a spline function. I'm wondering what is the most appropriate DV_KIND to be used for such a problem.
Thanks in advance.

[bigfootedrockmidget]

Could you use for instance the FADO approach of this tutorial:
https://su2code.github.io/tutorials/Species_Transport/


Using FFD boxes is a pretty flexible approach and the unconstrained free-form deformation gives you a good idea of how to optimize the shape.

[Marcus021]

That is what I went for and I'm currently using, indeed it seems to work well enough for my application. It's good to have confirmation on this in any case, so thank you very much.

[bigfootedrockmidget]

Good to hear that is approach works for you. If you have some nice pictures that you can share with the community, please do so, I am always interested in seeing how SU2 is being used :-)

[Marcus021]

With pleasure! I have attached a picture of the mesh used for the axisymmetric RANS simulation as well as a detail of the fairing and an example of the box deformation that I'm currently using.

As much as I'm glad to see the FFD_box is generally working, sometimes the optimizer will perform a deformation that will compromise the result of the simulation and the optimisation will either stop or simply stall. I wasn't able to fix this even by limiting the maximum displacements of the deformation points unfortunately, so I'm wondering if there's a setting that will allow the optimizer to simply discard a solution if it diverges or stalls and continue searching for others, instead of simply returning an error. As far as I'm aware is does not seem to be doing that and it's taking a lot of time to get meaningful results.