[Segermans]

Hi everyone,

Currently I am working on the aerodynamic design of a multi-element front wing for a race car. For this, I am looking for some advise regarding an optimization study using the ANSYS DesignXplorer tools. The setup of the project is as following:

- Created 2D parametric models in NX of 1 dual-element wing and 2 triple-element wing configurations. The following parameters apply:
1. Angle of attack mainplane
2. Angle of attack Flap 1
3. Angle of attack Flap 2
4. Overlap Flap 1 - Mainplane
5. Overlap Flap 2 - Flap 1
6. Gap Flap 1 - Mainplane
7. Gap Flap 1 - Flap 2
8. Scale flaps (the size of the flaps is equal, and the size of the mainplane is dependant on the size of the flaps. I.e. larger flaps mean smaller mainplane and vice versa.)

- Meshing and setup using ANSYS fluent with output parameters Cl & Cd. The final goal is a maximum value for Cl and a as high as possible Cl/Cd-ratio.

Now comes the tricky part where I would like to ask you guys for advise.

I ran a Design of Experiments using Central Composite Design which generated 81 design points. Next, I tried to generate a response surface using Genetic Aggregation. However, i found out the response surface has got a very bad fit. This consequently led to wrong results with large errors in the optimization stage, where I made use of MOGA. I did some additional research and at this moment I am generating a 'Kriging' response surface using 'automatic refinement' with a maximum of 60 additional refinement points and a maximum predicted relative error of 5%.

Since I really am running out of time, I was hoping to get some advise on this forum so I can quickly get on to the next stage: 3D optimisation. So, my question is: what would be the right path to take when optimizing such a problem?

Attached is a picture indicating the parameters.

Thanks in advance! Kind regards,


Seger