Aerodynamic Shape Optimisation Using Parametric CAD and Discrete Adjoint
Round 1
Reviewer 1 Report
This paper describes an optimization framework created using an open-source CAD system and CFD solver. In this paper the direct use of CAD models for optimization by developing a CAD system application programming interface, and creating a link between CAD-MESH-CFD analysis are proposed. And the optimization framework is applied on a rectangular wing and a three section high-lift aerofoil configuration within FreeCAD. The results show that the CAD parameterization can be reliably used to obtain efficient optimums. It is very interesting topic for optimization of wing configuration in CFD process, and it will be useful for other CAE to determine optimum geometries. There are a few questions about this method and conclusion.
(1) In this optimization process in this method, the baseline and the perturbated geometries are used. In this example, the scale ratio between baseline and perturbated is very small value. But how is the limitation of this ratio for general geometries for other CAE applications. Please clarify this limitation for applying this method to other CAEs.
If you have already checked this limitation by the approximation process or computational trials, please explain it in the main text.
(2)In this introduction, "providing a gateway for the creation of a fully automatic CAD-Based optimization framework…" is described. In addition, "It is shown the methodology is robust for a range of different types of CAD based geometry parameterization" is described.
In this paper, do you recommended to use the combination of FreeCAD and SU2 for a fully automatic system, or suggest that other combination of Open-source CAD and Open-source CFD can be applicable?
Author Response
Dear Reviewer,
Thank you for your constructive comments. Please consider our changes to address your concerns. Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
- The overall text needs a thorough grammar revision and spell-checking. Please make sure to do so before submitting a correction
- The introduction needs improvement. It is not clear what the novelty is regarding your "design velocity". It appears to improve your own previous work, but it is not clear how this is different to what is done in the literature. You should compare your work to other published work along the same lines (the group from Prof. Müller at Queen Mary University in London comes to mind) and stress the benefits of your proposal or at least position your work with regards to theirs and others in the literature.
Optimization methodology: you do not describe how you obtain the adjoint surface sensitivities with a discrete-adjoint-based methodology, for which the sensitivities are naturally volumetric.
First test case:
This is a turbulent case. Which turbulence model do you use? Does the discrete adjoint fully differentiate the turbulence model?
Table 1: last row, is CL OK? Also, giving the mesh sizes is not very illuminating, particularly for a 3D viscous case. What is the distance to the wall for the first nodes off the wall?
Author Response
Dear Reviewer,
Thank you for your constructive comments. Please consider our changes to address your concerns. Please see the attachment.
Author Response File: Author Response.pdf
