Creo Parametric Community Challenge 2 - Isogrid on a Curved Surface

[DaveMartin]

This month’s challenge is based on a modeling problem I’ve seen throughout my decades in aerospace. An isogrid is a repeating rib structure that adds strength with low additional mass. Although it may appear simple, modeling it in CAD can be difficult, especially when you need to place it on a non-planar surface.

Public domain image courtesy of NASA and Wikimedia Commons.

Your mission, if you choose to accept it, is as follows:

• Beginner to intermediate: Model an isogrid on a flat surface.
• Expert: create an isogrid on a cylindrical, conical, or radome surface. The isogrid can have a size / spacing of your choice and can be placed on the interior or exterior surface. (If placing on the interior, I highly recommend creating a cross section to make it visible.)
• Optional: add fillets to the interior “vertical” edges.

For this challenge, you can choose to ignore the center circular post and hole as depicted in the pictures. You can focus on the vertical sides of the isogrid triangles. (Of course, you can always challenge yourself by including those circular posts and holes.)

How would you do this in Creo Parametric? Would you use a profile rib, trajectory rib, extruded protrusion, extruded cut, datum curves, surfaces, pattern, toroidal bend, or some other feature? (Most likely, it would be a combination of two or more of these.) The sides of the triangles in the isogrid can either be connected or they can be disjointed (as in the image below).

An optional Creo Parametric 2.0 model has been attached with a radome part model as a starting point if you choose to use it. As creating the isogrid all the way to the tip can be challenging to say the least, a suppressed Solidify feature has been included that removes it.

Public domain image courtesy of Wikimedia Commons.

Please submit a part or zip file of your submission as a reply to this post and indicate which version of Creo Parametric was used to create it. Also, please include an image in your reply to this post. This challenge will close on Friday, November 6^th.

Example of a radome on the nose of an aircraft. Image courtesy of Kentaro Iemoto and Wikimedia Commons shared under CC BY-SA 2.0.

Additional examples of isogrids on curved surfaces can be found here and here.

Find the PTC Creo Community Challenge Guidelines here!

[tbraxton]

For the purposes of this exercise, I have a question about the requirements. Assuming pocket milling is used to remove material from a curved part, how must the tool path be managed when cutting the pockets? Is it acceptable to use a surface normal reference at the centroid of each pocket or must the tool trajectory be normal to the curved surface at all points along its path while milling?

This is an example of using the surface normal at the centroid of each orange 3 sided pocket on a curved surface. This topology will not be congruent to one where the cutting axis is normal to all points on the curved surface.

[DaveMartin]

One of the reasons isogrids are not very common is that they are incredibly expensive to manufacture. For the sake of this challenge, concentrate on the geometry and not the manufacturing method. You do not need to assume this will be manufactured via CNC machining. When evaluating the submissions, I will look at the resulting geometry and what features you chose to use. I will not be taking manufacturability into account.

[tbraxton]

Isogrid applied to the interior of a domed shell.

[DaveMartin]

Wow. Looks fantastic.

Can you share your model file with the community?

[tbraxton]

Creo 7 model enclosed. Resume all features after opening the file and prepare to wait (1 min 50 sec regeneration time on my 11th gen i9 laptop). Fully regenerated this model zipped exceeds the 47MB file size limit for upload here. I suppressed the patterns to reduce the file size.

Notes:

There is a geometry check in the model that is embedded within a trajectory rib internal round. If anyone can offer a fix for that I would like to see it.

The patterns are structured to regenerate "quickly".  Quick is relative to other potential construction techniques. It will regenerate faster than if the auto round function were used on the pockets for example. If there are techniques that will regenerate faster than what is in this model, please post them.

[DaveMartin]

Holy cow, you are truly a master! Honestly, it's been about 15 years since I've seen or even discussed using the old school technique of suppressing features or using Insert Mode for sharing files with people to cut down on file size. (It eliminates the geometry definition from being stored in the file.) That's what I love about these challenges: we get to see techniques that aren't taught in classes or best practices.

I was literally clapping at home while I was reading your post. Huge kudos. Thanks for sharing. And you are absolutely right, this kind of geometry can result in huge file sizes and long regeneration time. That's why an ostensibly simple problem can be so challenging.

[tbraxton]

I spent some time thinking about how to exploit the pattern functionality with respect to decreasing the regeneration time. The part is also axisymmetric, so you can exploit this as well to reduce the feature count. I also used multibody functionality in both parts used to build this version.

I was able to split the geometry into two models and get a significant reduction in the regen time required build the geometry (27 seconds vs 110). This would be very useful during the design/development of the part geometry as design changes are implemented within Creo.

[tbraxton]

Isogrid on O.D. of cylinder. Used removal of material to create the pockets on this one.

Creo 7 model posted.

[Fiddix]

I found the exercise harder than it looks!

here my 2 first try.

Thx to Dave for this challenge .

made with creo 4

.

[Aaronm87]

Fiddix, Can you explain how you used copy geom? I'm trying to replicate how you generated the part, but your copy geom window is different when I try to do it. I cannot select the same references that you selected. 

[Fiddix]

sorry for my very late answer !

I dont do anything particular:
I just make a box selection of all the geometry and then make a "repeat geometry"

I make you 2 screen shot to illustrate.

[jvadalia-2]

Hi Dave,

Thank for this challenge.

I think there are lot of different ways for this feature creation.

I have used on the way in non-planar, non-cylindrical, non-conical surface.

Please find below image

Part is attached. Give your feedback.

[jvadalia-2]

Hi,

Creo 9 has been used to create this part.

Regards,

Jignesh.

[HamsterNL]

Your "isogrid" is different from the rest 🙂 You have intersections with 8 or 4 "beams" and the length of your "beams" are not all equal (you have short and long beams).

But it's a nice grid! 😄

[tbraxton]

Isogrid applied to exterior of a cone. Design goal is to leverage symmetry to keep regen time down. Creo 7 model.

[Olivier]

Not realistic, but I have some fun to draw.
I have few opportunity to use the toroidal bend.
Thank you for the challenge.

Olivier 

[kdirth]

Here are my 2 submissions.

I started with fairly simple open shape:

Then I decided to try the radome supplied by @DaveMartin: 

 

I would say that my models were created by brute force, repeating the same operation over and over.  However, I did use a lot of patterns and automated the repeated operations with temporary mapkeys.

Below is a link to the Creo 7 files.  They are both too big for the 47 MB limit for the community board.

CHALLENGE-2

[kdirth]

The 3D printer was in need of some "exercise" as it had been sitting idle for almost a week, so I printed a scale model about 2.5" tall.

[Pettersson]

Fun challenge! I did mine in Creo 7 with some Spinal Bends and Multibody. It got a bit thin towards the bottom, which could be fixed, but I couldn't be bothered. 😄

[EF_10127429]

I'm new with Creo but have used SolidWorks for 10+ years. Here's what I came up with. I created the following and the made a pattern from it. Outer ribs are half the thickness of inner ribs so that all ribs are uniform when the pattern is applied. Using Creo 9.0.4.0

[EF_10127429]

Attaching model here

[DaveMartin]

Please let me know if you would like me to use your username or your real name in the blog post summary. If you would like me to use your real name, you can either message me or reply here. Thanks!

[jvadalia-2]

Hi Dave,

You can use my real name "Jignesh Vadalia".

Thanks & Regards.

[DaveMartin]

Thanks, Jignesh.

[Pettersson]

I tried a different approach. This time on an easier surface, just somewhat deformed, meaning I could use Flatten Quilt Deformation, which is a really fun tool. My approach this time was to use surfaces as far as possible and only thicken them towards the end. the advantage of this method is that it keeps a constant thickness on the ribs. If you deform the solid geometry, the thickness will vary, but if you deform the quilts and then thicken them, you get a constant thickness, which might be desirable. It was a bit finicky, though, and I had to play around with the accuracy to make Creo do all the thickening. Even so, there's actually a small rib missing, but it's close to an edge, so I don't think it would matter for the strength.

This part is in Creo 9, which is what I happened to have open when I started playing. Apologies to anyone on a lower version!

[Pettersson]

I, uh, need to stop this. But I had another idea for the radome and played around some more. This version also has uniform wall thickness, since the walls are made with sweeps. It also gave me some additional control, allowing me to change the pattern as it approached the nose, keeping the triangles from getting too squished together. I'm quite pleased with the pattern; I think it looks lovely. From above, it looks like a flower.

Part is made in Creo 9

[iDZignit]

I like your approach, on a varied curvature surface such as the radome it will be difficult to maintain rib length.

However, the means to create the ( physical ) ribs will be a separate challenge.

[Pettersson]

I, uh, made two more. Yes, I have a lot of time on my hands at the moment. And it was fun to play around with this challenge. As I kept working I got new ideas!

This first one has horizontal ribs, and then a zigzag curve that gets patterned. The zigzag curve goes up and down between two planes, and by using a table pattern to each time move up the bottom plane each instance, the pattern gets inverted, creating the isogrid pattern. I'm quite satisfied with this solution, and the table pattern means you get a lot of control of exactly how the pattern looks. Each row can be moved individually. It did require a pretty low accuracy setting, and trying to round off the corners made it hugely computationally heavy, so I skipped that part. It does have quite a few geometry checks, but these are all for edges that get cut off by the solidify features, so I don't think they should cause downstream issues. I didn't do any "finishing" work at the tip of this one.

The second one is an easier part, since it's a simple cylinder. I read @iDZignit's comments about manufacturability and I wanted to see if I could make one that's easily millable. When I teach Creo, I always caution the students against the temptation to try and make as short a model tree as possible. This time I fell for that temptation myself, and my part consists of one revolve and a pattern of extrude features. The extrudes have a somewhat tricky sketch with some sketch relations that move the triangle up and down as it moves around the cylinder (necessary to keep the rows straight). The result is an isogrid consisting of straight extrudes with planar walls, simple radii and flat bottoms. Of course, that means the wall thickness varies, getting thinner towards the bottom, which might not be ideal, but it should be very simple to manufacture, which was my goal here. Make the triangles smaller in relation to the diameter and the problem should be smaller. Or just add a taper to the extrudes. You could also do the extrudes "up to surface" with an offset and you'd have a bottom that follows the cylinder curvature as well. But I kind of liked the very simple nature of this pattern.

Both parts made in Creo 7.

[DaveMartin]

Everyone, the second Creo challenge will close this Friday, November 3rd. (In my original post, I had accidentally written November 6th.) If you want to participate, please post your solution this week. Thanks!