CAD softare - auto pack/nesting of 3D models in 3D volume 2

[HilaryE]

Hi All,

I am looking for a CAD software which would have a feature of auto packing 3D objects in a specific 3D volume.

More specifically, I have a 3D CAD box 100mmx100mmx100mm. I have also 3D CAD model of a product i.e. 13mmx13mmx9mm (with rounded edges). I need to fill 3D CAD box 100mmx100mmx100mm with randomly arranged 3D CAD models of a product.

I remember that such auto pack/fill (of products) options are available in software used to program CNC laser or plasma cutting machines (for 2D sheets of metal), for CNC punching machines, etc.
I wondered whether there exist corresponding features in some software for 3D objects. In such a 3D software, I could define some boundary conditions for packing (such as max and min spacing between objects) and get them automatically arranged in requested virtual volume (box).

Can you help me with this guys, please.
I would be very grateful for any help.

Thank you.
Regards,
Robert

 

[MikeHalloran]

I have used a very few of the 2D packing programs, and was generally disappointed by severe mismatch between what I wanted done and the actual capability of the program.

I think there are empirical functions in chemical engineers' books to calculate how many truckloads of a given shape you need to fill a given tower with 'packing' devices. Maybe you can extend or adapt those.

Working from my own observations, I doubt that you'll get optimal packing if you insist on random anything.
For the specific problem you pose, you could use a spreadsheet to work out how many widgets you can get in with orthogonal arrangements, and the biggest improvements you will find, will come from adusting the overall package size to some multiple of the ordinal dimensions of the product, e.g., 13 x 9 x 13 mm x XX x YY x ZZ, or by making the product envelope such that the products intermesh with each other, i.e., nest.



Mike Halloran
Corinth, NY, USA

 

[GregLocock]

How do you deal with bridging?

Perhaps you could bound the problem, using a manually packed layer as the upper bound as Mike suggests (by observation about 98%), and say container volume/volume of a spheroid that encloses the part * the best packing density of spheres as a sort of lower bound. I know from reading around this subject before that the definitive answers are few.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[HilaryE]

@MikeHalloran, GregLocock, avscorreia,

Thank you for your kind responses.
Indeed, the considered scenario requires randomness of a product positioning (13mmx13mmx9mm with rounded edges). I should receive samples of that product soon (I have been waiting for them for a long time). I will insert these samples (by allowing them to fall freely) into a small container. I will count a number of products accommodated by a container (of know volume). This will do.

Thank you for all your ideas how to deal with my problem. I am grateful.

Regards

 

[MikeHalloran]

You might want to wander around YouTube for a bit.
There is an entire industry devoted to conveying, orienting, and inserting widgets of all sorts of shapes.
Start at 'vibratory conveyors', which typically wobble a big bowl of random parts around, and induce them to climb a peripheral helical ramp, where some remarkably simple custom tooling can either orient the parts as desired, or allow the ones that are not already oriented to fall back into the bowl.

From there, you could use custom translators or robots to pick up the oriented parts from a fixed location, and insert them into boxes is a specific pattern.
... or just let the stuff fall from the ramp, and count it photoelectriclly.

Good luck and have fun.


Mike Halloran
Corinth, NY, USA