surfaca curvature analysis

[g.lopedote]

Hello everyone,
I am a new user and I begin by complimenting the quality of the work done on this forum, really exceptional.
hoping not to make mistakes I ask you my doubt.
I'm working on a surface and I'd like to check its quality. Therefore, taking advantage of the means available in the workbenches gsd and freestyle I came across the command "surface curvature analysis" (surgical curvature analysis). In particular, I used the option "flex area" (inflection area) to highlight any unwanted flexes. the surface seems to be rather "clean" if not that, using the curvature analysis to porcupine, they appear flexi.
I read the caia documentation, but I couldn't fully understand the potential of the command "surface curvature analysis", so I ask you a hand on the subject.
in the attached image, there is the surface in question, analyzed with the command "surface curvature analysis" and the option "flex area". in the green area should not be flexibly, but by doing a porcupine analysis, these appear.
I also opted for the command "flex lines" (inflection line) that traces curves obtained by joining the curved points nothing. as shown in figure (second image), there is only one line of flexo on horseback between the blue and the green area (exactly as I expected), but none in the area where the porcupine shows me the flexes.
So the doubt remains, why do I get the flexes with porcupine analysis?
which command allows me to better verify the quality of my surface?

greetings

 

[gianni55]

Hi.

I see that no one answers, so I try to give you some suggestions (but not the answers you ask, if you place the surface in step perhaps I can understand more, I add that those commands I know little and that it is not easy to explain them "distance")

on the fact that the documentation is not sufficiently clear you are right, but I suppose that by dealing with concepts of mathematics and analytical geometry, they are in fact essential and remain to what one has studied (and it is not my case).

I believe that the curvature of a curve (and the search for the flexo) responds to a specific geometric condition, while the curvature of a surface considers others that are not tied to the cutting plane but to the normal compared to the point and to the averages (but here for me the speech is too difficult), I would say as an example: try to create 3 sections with splines without any flex, trying to build a saddle or donkey back surface and then see what happens.

First tip:
creates sections with splines and generates a surface, then changing the splines and using the analyses should be easier to understand the operation.
according to advice:
search the rhinoceros help online, the corresponding entries have some extra information.

two words on the command surface curvature analysis - flexo- area, which reports text:
"This analysis allows to identify the orientation of the curvature:
in green: the areas where the maximum and minimum curvatures present the same orientation.
in blue: the areas where the maximum and minimum curvatures have opposite orientation. "
then in your case report the opposite orientation areas.
(to better understand the concept, try to draw a "toro" and analyze it, this while not presenting sections with flexo, will be analyzed with the opposite orientation areas)

the other command you used: flex lines, track you curves, but if I don't remember badly, emit the result according to the direction of the manipulator (make the test), then either master it or take you off the road.

the porcupine curvature analysis, through the graph, highlights whether the curve has a flexo, if it is in continuity of tangency and curvature.

the goodness of a surface is aimed at what you want to achieve, sometimes the flexo is necessary, other times it is to avoid as the plague, depends on the field of application, the constructive needs, the needs of style etc.

 

[g.lopedote]

thanks for the answer
You gave me some great starting points. actually the online guide of rhinoceros is more satisfying than that of catia.
I don't hide that some evidence I've already done, but your ideas have enlightened me. I'm trying to do some more tests.
more than solving the specific case would be very useful to learn how to fully master those commands, I understand what they actually allow you to view.
about the "flex line" command, there are two options: compass plane and parametric.
greetings

 

[gianni55]

in the automotive sector there is specialization related to the creation of class surfaces a (b etc.) that wrong or right is a little elite of surface modelers... from them you can learn something more.

Hi.

 

[Lettore espatriato]

Good evening.
I share everything that has been said for years.
I add, for what is my experience, that it would be good to leave, where possible, from "clean" theorists/wireframes. For clean I mean that they do what you need. It could also be a prominent theorist.
I hope tomorrow I can add some extra help if it can serve.
for example see a profile with stretch view to highlight the trends .

 

[g.lopedote]

Good evening.
I share everything that has been said for years.
I add, for what is my experience, that it would be good to leave, where possible, from "clean" theorists/wireframes. For clean I mean that they do what you need. It could also be a prominent theorist.
I hope tomorrow I can add some extra help if it can serve.
for example see a profile with stretch view to highlight the trends .

the starting profiles I used are quite clean, or as you say, as clean as necessary for my application.
I understand better how the caia commands work. having understood what I visualize to monitor, with some small modification, I also cleaned my surface.
p.s. material for any type of study in this field is well accepted
greetings

 

[Lettore espatriato]

Good evening.
I would like to point out that my experience is only practical, so if some industry experts have to add or correct or criticize what I am saying in each sector is welcome.
1) I try to avoid caia splines for style profiles . controls well as points of passage but as the curve in the middle is not always predictable . Sometimes you have spikes in the analysis curves . Usually to define the theoreticians I start from 3d curves or bezier that are those with the external control polygon.
2) What do I need? If it is style, then you must define "velociously" volumes to evaluate the effect, then perhaps it is better to use the im and subdivision surfaces . Parts from a sphere or another elementary geometry are and by hand the plasms until you reach your result . This especially if you have a zone where you do not have theorists or do not know how they will come .
If you already have the main views with theoreticians or are reversed, then I leave from the 3d curves with fewer possible poles (usually at least 3), you put in a main view, move the poles according to that view by putting the normal manipulator to the view, if necessary increase the degree trying to have the poles closer where the curvature is greater and less where the curve is more relaxed. What? depends.... side and top, or side and front .
For example, the surface you posted has two "guides" (normal ones to porcupine analysis) . Have you verified if seen of "side" have a "coherent" pattern i.e. if they "enter" gradually or if they "close" continuously?
In cases of curves that coincide in a view, you can take a defined, copy it and move only the changing poles.
If you reverse from stl sometimes you can start directly with bezier surfaces and check theoreticians with multiple surfaces intersections.
I hope I have helped and not confused ideas.

 

[g.lopedote]

thanks for the answer and for sharing with me your know-how.
I add that my main problem was the lack of understanding of the means made available by catia. In my opinion, catia is not very clear on this subject. Now I am much more, and therefore I add that I have solved most of the problems of "understanding".
basic curves, guides and profiles have already been cleaned, trying to remain faithful to the mother surface.
Thanks to all of you, now I know more than before.