Srf match challenge unsolved

hello!

ive been trying to srf match 2 edges. ill upload the model for alternatives! please toggle on the zebras.

ill mark in red the area im interested in. see pic below

notice that to model the rear of the car, i need tangency or G1 continuity. curvature would not be possible due to the geometry of the wheel arch.

file model;

rhinoSample.3dm (1.3 MB)

please feel free to suggest any alternative.

thank you!

I think @Rhino_Bulgaria gave you a very good solution to this on the previous topic. just extend the top surface and trimm it.

1 Like

Here you go (again). Simply build new crease surfaces above the one that I extended and split. :slight_smile:
Split this surface.3dm (274.4 KB)

thank you for your reply!

your alternative projects very sleek zebras but leaves a gap between the creased edges. ive been trying to close this gap, but it seems that the trimmed edge is way to complex that its impossible to close. even when increasing the absolute tolerance in the preference section Units (options command) thus it is mathematically impossible for the software to adjust those cv’s from the trimmed srf in order to give the desired result, theres no way…

the reason of this file is to model a component as if it was kind of ready for production condition that means no gaps, tight tolerances and best surface quality possible.

the file i uploaded is not the same since ive made minor changes in the layout of surfaces.

i did more or less the same that what you’ve stated but without trimming. instead, by CV massaging. and is worked to a point…

my problem now is that when i run the srf match command it doesnt work accurately. the results it gives are very poor in that particular area. my question would be:

is there any way to obtain continuity between srf edges manually? how?

thanks anyway!

please would you be so kind to check the model and try it by yourselves? run the srf match!

try this.

rhinoSample.3dm (2.2 MB)

Hello - if you increase the U direction degree of the surfaces to 5 and MatchSrf, you can get a better looking zebra, but I’d say approaching this by crowding control points up by the sharp bend

is not a good approach- the way I read this, the sharp bend has nothing whatsoever to do with the curvature if the large flat surfaces; by going about it this way you are forcing the outline of the shape to influence the design shape - see how the zebra is more wiggly as the points crowd up. As I suggested in my first response, the other day, and as others have pointed out, the way to do this is to trim that trailing edge into a clean, over-built surface that makes up that side of the fender. Build your smaller surfaces at the edges to that big surface, not the other way around,

-Pascal

1 Like

@oliver.jans this is the idea as I understand

rhinoSample (re).3dm (1.4 MB)

finally someone! congrats! challenge solved. :hear_no_evil: :100:

thanks pascal! ill check the curvature there. i think that the excess of cv’s generates that srf anomaly…

I thought that the goal is to achieve what @pascal mentioned above - a quality main surface that determines the additional border surfaces. Your input border surfaces and bottom curve are such that they make it impossible to achieve good main surface given the restrictions it’s forced to follow. If it was mine project, I would be focused on having a flawless main shape and would worry less about some small details that my have a few more control points. After all, it’s the quality of the shape that matters in manufacturing.

Here is a quick comparison between my proposed solution with a single surface that don’t follow your input surfaces, and what @vikthor did while trying to follow your input surfaces (not trying to compete with anyone, just making a suggestion why it’s important to keep the main surface smoother):







My surface could have been even better, but I followed the input wheel arch shape that’s slightly wrong, this is why it affects the zebra stripes at the bottom in the way you see it in the last 4 images. In reality, car body panels are made the reverse way: You first build the main (big) surfaces, then add the extra details.

3 Likes

@vikthor, @pascal, @Rhino_Bulgaria, all your ideas have been welcomed! interesting what you’ve mentioned there @Rhino_Bulgaria I quote

it’s the quality of the shape that matters in manufacturing.

and

body panels are made the reverse way: You first build the main (big) surfaces, then add the extra details.

Thank you @Rhino_Bulgaria for sharing your knowledge. I’ll probably follow your instructions. please would you be so kind to upload that file? :yum:

Oliver

ufff I did not stop to carefully review all that
here is the file, according to details mentioned by . without However you should change the surfaces in red
rhinoSample(re2).3dm (1.8 MB)

One thing to remember in such cases is that you can’t build a rounded shape along nearly straight border surfaces or curves, and expect the former to be still rounded while being straight at the border. You either have a flat surface with flat border, or a rounded surface with rounded border.

To better illustrate what I mean, check this example. A rounded shape that has straight border (while looking at it sideways) will never have a straight border while looking at it from another angle.

Build the other small surfaces around that main one.3dm (1.8 MB)

By the way, what’s the Rolls Royce model from the 60s that you are trying to recreate in 3d? I did a search in Google and it was unable to find a Rolls Royce with such a fender.

@Rhino_Bulgaria, @vikthor always a pleasure!

@vikthor I find your file models very hard to download lit. around 38 min left. and let’s say that by minute 3 the download freezes…

Maybe you could export it as an older file, let’s say V4 3dm. Anyways, your ideas are very clear but it would be better to for me to visualize your model. Please follow the instructions I listed above or try sending your model via email if you don’t mind: oliver.jans@icloud.com. that should do… :smile:

@Rhino_Bulgaria, regarding the model I’ve sent you a couple of days ago… I must confess that my RR '60 fender interpretation was seriously mistaken. Anyways I’ve uploaded an uploaded file version :stuck_out_tongue_closed_eyes: duh of the new fender but there’s still a lot of work to do. :sneezing_face:

Look Look… I’ve made fender comparison before and after:

original RR fender from 1960. Well indeed it’s a toy car, but accurate :expressionless:

After;

before;

Something to point out is that my model is very amateurish… very simple, easy-to-read surfaces. If you check the original car it’s way much more (let’s say: curvier) almost natural or organic geometry.

It seems that I got both of you very intereseted in this topic… I would like to explore more with both of you guys! let’s make this model happen to test our skills.

I believe that the main idea of this forum is to learn from others… I could share some of my file models in exchange of yours (of course if you don’t mind). I have thousands of hull models exported as NURBS srfs for rhino to read. As you may know, hull modeling has always been a challenge. If I could just send you these 3dm. files it would be great for your portfolio. :yum:

Best Regards,

Oliver

Semi-Displacement Hull Form SDHF V6 3dm.file ScreenShot


Conventional Hull Form V2 3dm.file ScreenShot

I got plenty of them!!!

What is the original source of the hull models?

I model every curve (stations, buttocks and waterlines) in 2 naval Architecture softwares:

I use DelftShip for preliminary geometry definition and then run MaxSurf for precise geometry definition, analysis, lines fairing, etc.

you interested in a specific hull model? name it and i’ll see if I can model it. (it will take several weeks)

These softwares usually come with a library of standard hulls to make the modeling easier. But it’s worth mentioning that these hulls are not, let’s say in a ready for production condition. They do requiere surface improvement and hours of dedication to multiple specific tasks.
Naval architects usually run multiple softwares each committed to a specific task. MaxSurf is a software widely used around the world and it’s currently catalogued as the best because it covers lots of tasks.

check out this file model IGES.file provided by library hulls.

hullWknuckleLine.igs (409.9 KB)

Best Regards,

Oliver

If unable to download previous file, try this file model…

libraryExample3dmFile.3dm (2.0 MB)

If you toggle on the zebras you’ll probably notice that they look messed up… this is partially because when exporting a hull model into IGES file model there are several conditions or aspects to consider. 2 of those many aspects are that surfaces are going to be:
rebuilt with degree 3 (smoothest degree possible.) - Good aspect :blush:
and,

the continuity or tangency to fails in some areas. - Bad aspect :rage:

@davidcockey, what naval architects usually do to export hull models accurately is for example, they export the hull model as CFD meshes to accurate computational fluid dynamics analysis, or they can simply export the cross sectionals (stations and waterlines) and then run a special software capable of reading and understanding these intersections. Such they is Orca 3D (plug-in for Rhino) available at Food 4 Rhino web site in marine section.

link provided:

https://www.food4rhino.com/browse?lang=en&sort_by=ds_changed&f[0]=im_field_unified_type%3A773&f[1]=im_field_platform_app%3A722&f[2]=im_field_term_reference_category%3A701

Other method widely used around the world is exporting the coordinates of the intersections of the hull into an excel spreadsheet. But unfortunately, you won’t be able to 3D visualize your model :sneezing_face:
This method proves to be the best since working in excel has always been relatively cheap and easy to operate.

Best Regards,

Oliver