As a PhD student of Marine Engineering with about 7 years’ experience in Computational Geometry and Mesh Generation, 3 years of which have been spent to Parametric Design and Optimization, I am so glad to find you guys who did such a good job around Parametric Design.
But it seems there are some points which have been neglected, as far as I understood from the video clip on “https://www.youtube.com/watch?v=8x75TJUmoDM”. For example, the propeller has been parameterized by some 2D profiles (NURBS curves) which make it possible for the user to modify them by moving their control points. In fact, I believe, providing some control points without defining any constraints to control their movements, as well as, skipping their effects on each other may cause lots of non-applicable geometries to be generated.
I should tell you that I have developed a C++ program to design ship hulls, propellers and rudders by parameterization approach. I’ve actually applied my own method in propeller designing module of this program, and there are no such issues. In other word, my program provides a design space rather than a tool just offering the ability of moving the control points. By supporting this design space defined by 26 parameters, my program is able to use, say Genetic Algorithm, for searching the optimum answer without produce any non-applicable models.
To share these with you, I tried to sign up in your forum, but surprisingly I saw that my country name is not in your registry menu and I had to register with another country name! which made me feel very bad.
FYI, I’m from IRAN and I hope this happened by mistake and will be resolved.
By the way, I have plenty of ideas to improve your software. Enclosed you can see some of my program's results.
Prop-1b-SK0.bmp 167.67KB 153 downloads
Propeller.bmp 245.99KB 167 downloads
And here are parameters: