What's new
What's new

Generative design- Is this on everyones radar?

Trboatworks

Diamond
Joined
Oct 23, 2010
Location
Maryland- USA
I seem to be the last one they tell anything to.
Is everyone up to speed on what is coming down the pike for design development.

I have in my modest way have spent my life solving mechanics problems designing things that do work or are parts of structures.
I have always enjoyed the challenge.
But I always thought the whole craft was getting real world experience so one refines the skills with materials and technique to meet the days problem with a robust solution.
It appears overnight I have been obsoleted...


"What are the advantages of generative design?"

"Generative design has several great benefits:"

"Breaking down creativity barriers
Firstly, the designer’s creativity is no longer a limiting factor. The software can come up with design proposals that engineers have not, or barely, thought of, breaking down the restrictions of human creativity."
Generative Design: Design the Future | Innovation Blog


I'm no longer the limiting factor....
Well I have always been the limiting factor- I am sorta slow and distractible but when on task can bring to bear the experience and a native talent for the work.
It sounds like "talent" just got cut out of the loop and my one skill strong enough to offset my weaknesses has been replaced with a couple of bucks of cloud computing time.


Ok- I am just bitter as the world is truly moving on.

And it moved on some time ago apparently:

Autodesk and NASA's Jet Propulsion Laboratory partner to build interplanetary lander - 3D Printing Industry

I guess management is just being polite and hoping I retire before they have to push me out the door...
 
That makes me laugh but I don't know if me finding that funny means you are in fact offering humor...

I honestly think this is going to change the world.
I have long thought that the future of architecture is in membrane/skeletal/ligature structures.
This design approach is going to revolutionize material use and design optimization in everything from car chassis to bridges they drive over and the buildings they drive to..

This is another case where Sci-fi was way ahead of the curve.
I have noticed for a long time that the aliens are always presented with this organic insect like design aesthetic.

I have joked with friends about "Bug architecture" when I have a particularly daunting design problem.
I can feel in my bones that design optimizations are there if I can find it and lament for the advantage insects have in their millions of design iterations over the millennia.
I have always sensed that the answers are going to be in these organic shapes but have long struggled to get in front of the forces problems.
Hence my lament about "maths" being absent to design well.
 
Your initial post has the flaw of not really describing what you are talking about.

After reading it over a couple times, I said WTF and clicked on the link...something I seldom do. After reading the article, I realized it, too, doesn't really describe what it's talking about. It's layers upon layers of beating-around-the-bush bullshit. Then, I realized that must be what generative design is - layers of bullshit all laminated together to make something.

It sorta sounds like someone who has never swung a hammer has precisely invented, on paper, a way to drive a nail. But nails are real, and a hammer sure is a fast and simple way to drive one.

I smell lots of meetings, coffee in styrofoam cups, and quarterly awards handed out to people sitting on those metal/plastic chairs that are linked together in conference rooms.
 
Your initial post has the flaw of not really describing what you are talking about.

After reading it over a couple times, I said WTF and clicked on the link...something I seldom do. After reading the article, I realized it, too, doesn't really describe what it's talking about. It's layers upon layers of beating-around-the-bush bullshit. Then, I realized that must be what generative design is - layers of bullshit all laminated together to make something.

It sorta sounds like someone who has never swung a hammer has precisely invented, on paper, a way to drive a nail. But nails are real, and a hammer sure is a fast and simple way to drive one.

I smell lots of meetings, coffee in styrofoam cups, and quarterly awards handed out to people sitting on those metal/plastic chairs that are linked together in conference rooms.

Yup, these "programs" have been around for many years.
They "modularize" a assembly, and then allow sales people to simply click some boxes
to "configure" a product for a customer.
I actually had to deal with a supplier that does this, the drawings were crap, the locations of things were wrong, and the actual machine didn't look like the drawing, and
didn't preform as purchased.

A failure all around, but they eliminated the engineering department.....

Just like CAD salesmen that had preloaded commands to impress a sales demonstration.
 
I am more on the engineering and design side these days.

I think most of the promo pics and examples you see, whether it's a quadcopter or a lunar lander, or some fugly custom car wheels, are all just show pieces. Don't give them any more credence than you would Hermle machining a big aluminum bull, or someone machining a basketball goal and net... no one is likely machining parts like that, it's just a flex and attractive piece to get attention.

Generative design is pretty handy, though. I've been using it here and there, but it's more like a interim stage in the design. Obv salesmen are gonna tell you it's the world's best Easy Button (TM) but it's really just another analysis tool that does some of the shaping for you, instead of an FEA telling you where your existing shape might suck.

It's also only as good as the guy who set up the engineering model.

But as for manufacture, the information pruned from a generative design can just as easily be applied to a part made easy for 2.5d machining as it could for true 3 axis, 5 axis, or turning.

Like shitty castings, impossible forgings, parts that require a much-too-complicated injection mold... there are gonna be dumbass parts but we'll still make money selling the tooling or services to make them. Then there will be plenty of people actually using generative design the way it's best used for, and they'll go on to do ever better things. c'est la vie.

Just another vector for machinists and engineers to talk shit to each other in the same old fashion :D
 
The problem is my call on how influential this nascent tech will be to our world of structural design challenges will take some time to play out.
Let’s say an update every five years.
While that time line plays out I am going to start developing models and running them through this design process for some problems I am working on.

Oh- post just above.
I in no way see this design approach limited to any single manufacturing method.
And the current application stream isn’t limited to drones.
Airbus, space flight- let’s see how this plays out.
 
Last edited:
I hold that the last three posts are wrong.
Problem is my call on how influential this nascent tech will be to our world of structural design challenges will take some time to play out.
Let’s say an update every five years.
While that time line plays out I am going to start developing models and running them through this design process for some problems I am working on.

Oh- post just above.
I in no way see this design approach limited to any single manufacturing method.

I hold that both YOU and the Linked documents don't explain it very well.
"Garbage In, Garbage Out"....:nutter:
 
It attempts to 'squeeze the last drop' out of manufacturing and design. A noble concept, but one that in the real world is usually more trouble than it's worth.

If you are making a billion of a certain item, like an Iphone, it might make sense. But if you're making only a hundred, probably not. It also overlooks that not every item produced must be made to perfection....many times, something that is close enough is good enough.


It might be true that a chair that looks exactly like the exoskeleton of a beetle represents the height of design, materials, and construction....but a chair that looks like a flat surface with 4 sticks extending down will hold your butt 16" off the floor pretty well.
 
Our most compelling design problems in the society are not chairs.
But like I said- let’s see how this plays out.


And yes- the applications are not for the sort of things a guy can knack together out of old orange crates.

I am seeing some of the old us vs the engineers popping up here.
I believe you are missing the point.
Everything I have ever designed goes through rapid prototyping as I lay out the task.
First in my mind as I run the first few dozen iterations them on a napkin or in CAD as the case might be given if I need to communicate with clients or complexities demand it.

Generative design takes that process one step further and challenges the design further.
Is it required for my work- obviously no.
But for design challenges to society this will change the world.
Electrochemical engineering in advanced battery structures is the the very first thing that springs to mind.


Don’t see the importance of this process.
I am betting you are wrong.
 
I think it's pretty revolutionary. I think codgers and naysayers will look at the "beetle skeleton chairs" and poo-poo it as a marketing ploy, children's toy, etc. The same was done about additive manufacturing / 3d printing and... by some troglodytes, still is.

But generative designs is simply something that's been dreamt of for decades, that's simply recently possible because Moore's Law has been in play long enough.

It's a valuable engineering tool that approaches things differently and enables solutions that were previously impractical to calculate for, for the vast majority.

I am particularly interested in the helpful tools that can quickly iterate through designs with manufacturing limitations in mind. Being able to get designs based upon good engineering data, known geometric limitations of manufacturing processes, and then iterate through solutions until something works? That's pretty damn amazing. It's incredibly difficult and time consuming to constantly iterate through a design, trying to balance the needs of tooling/workholding, manufacturing processes, weight limitations, material cost factors, heat propagation... and then try to bring any one of those into acceptable bounds without cancelling the validity of any other factors?

This is why computers are so great. They do the things our monkey synapses just can't do so well simultaneously. Brains aren't so good at multi-threaded processing... or maybe that's just mine :)

But yes, if you try to look at a chair or a bracket and think "how the hell does Generative Design make this better?" it's just not the right question. There's no reason to run calcs or FEA on most brackets, either, unless there are documentation/procedural reasons to. So to take something that no one even bothers to do the math on, and shoehorn G.D. into it... nah, not applicable.
 
I agree that is valuable in a number of ways, albeit pretty impractical for most stuff due to manufacturing/cost limitations.
  • With respect to stress, material usage and placement, it's never a bad thing to know how efficient a design can be to accomplish a given task
  • As soon manufacturing techniques catches up it will be more useful. At this point it makes cool looking models that would be expensive to accomplish (outside of 3D printed parts)
  • I haven't messed with Fusion 360's implementation lately but they have some tools that allow the user to define the manufacturing strategy which is pretty cool, allowing you to get a bit closer to the optimized model's efficient material usage and placement.
  • The FEA behind the generative model still needs to be set up properly to generate an optimized model. Setting up FEA for complex loading structures is not always simple and there has to be some (sometimes large) safety factors to account for various assumptions made while setting up constraints and loads. A potential downside of generative is that you are removing "superfluous" material based on the loading/constraints which may not be superfluous at all. Of course proper testing of prototypes take care of these assumptions.
 
I have played a bit with Fusion generative design, and with minimal parameters, Fusion pukes out something that looks like it was created from some alien crystals- all weird and impossible to machine, but easily printed. Also very ugly to my eye


Sent from my iPhone using Tapatalk Pro
 
In the extreme “output” the designs are questionable.
But I see this being easily directed towards balanced outcomes.
All of the work has a somewhat unfamiliar look to it but to my mind perfectly acceptable.
But I have been working years now with a structural aesthetic which embraces ‘bug architecture’ so I am a receptive audience...lol

8C70A76A-4490-471C-8050-D00FCBB01B3B.jpeg

StackPath

Pioneering bionic 3D printing - Company - Airbus

I am going in with 360 and see how much tweaking is possible.
 
Holy Crap guys!

I have not had a drop to drink (yet), but can honestly say that all of what was so far said in this thread sounds like a socialite's deeply considered essay
on one of Salvatore Dali's creation.

IOW, the whole discussion flew way way over my head.

One thing did however catch my eye from the linked article:

Generative design can explore new solutions that even experienced designers might not have considered, while improving design quality and performance. Because the designs are nearly impossible to manufacture using traditional methods, additive manufacturing techniques like 3D printing are critical to generative design’s success.

( thinkin' that I'm glad to be of the age where I will likely not have to deal with what might be the outcome ...)
 
I see it as a cool but dangerous tool of design. The plasma cutter has done more harm to metal design (art + some extent structural) than most anything else. It is not the lack of what it can do, but people not thinking about its output and overall aesthetic.
I did play with fusions a tiny when they had it on the low tier plan - now its 100$ per pop! I foresee a google (or open source) thing coming soon, which will make 'everyone' a structural designer. Like sketchup makes everyone a cad and design genius.
In designing I like the organic shapes it leans to, but more importantly it would give a chair designer the vision to remove material (visual weight) as opposed to our backyard engineering of just adding more material to make something stronger. Having real time feed back and nudges on were I can drastically very weight or even skimp to 1/8" steel on a shelf would be another tool.
 
Holy Crap guys!

I have not had a drop to drink (yet), but can honestly say that all of what was so far said in this thread sounds like a socialite's deeply considered essay
on one of Salvatore Dali's creation.

IOW, the whole discussion flew way way over my head.

One thing did however catch my eye from the linked article:

"Because the designs are nearly impossible to manufacture using traditional methods, additive manufacturing techniques like 3D printing are critical to generative design’s success."

( thinkin' that I'm glad to be of the age where I will likely not have to deal with what might be the outcome ...)


That is my fault as I provided that piece as it knit into my sense of humor on being obsoleted by this new approach.

I really do not agree with the assumptions made there- for one I do not in any manner believe that this approach is limited to 3D printing.
I believe what is potentially revolutionary here is the simple iterations done in rapid succession testing established constraints.

It speaks directly to my long held believe that an evolutionary analogy is the exact barrier to my own design work.
I simply cannot build and effectively test enough interactions to approach true optimization.
Granted I am good enough to produce good work as like most I understand my materials well enough that I don't get failures.

To my mind generative design is revolutionary- full stop- a game changer.

I am of the believe it will be adopted in everything from car chassis to bridges.
I think it will find use in subtle application such as optimizing chemical processing structures.
It may (and forgive me for bringing it up..) crack some intractable problems such as fusion containment.

Now I will know that this subject is just a continuation of long established computer modeling of structures.
I guess (and it is a non experts guess) it that we are starting to see a combination of several disciplines which will define the future of many manufacturing and engineering approaches.

For me- In speaking with Autodesk they are offering that the tech is rapidly being embraced and used.
They suggest refinements are occurring and will continue to be so.
I would love to see some really practical constraints- design iterations limited to stock tube and cut plate for instance.

Ok I'll stop.
Honestly though this knocked my doors off as I have faced so many times that plan page which has to be filled with a design which works and pinning down the best one has long been my nemesis..
What is not revolutionary is that engineering is an evolving field and naturally has long been so.
Design iterations take a damn long time though- I see this as a breakthrough innovation which will rapidly accelerate our approaches and the products produced.
I do feel the 'look' of the manufactured world will shift as a result.

Welcome to a new age- I saw it first.
Bug architecture...

We can note that Airbus puts the topic safely into the far future.
"Our 2050 jet liner"
Nice to talk about but....

05294380-6C84-47E9-9FE9-28DD51F6F1A3.jpg
 
Over here we have aname for such outfits ;- Wanking Bullshit Artists.

That moment when you realize the world has moved past you.
Decades ago....

Llis

Or rather I should say the engineering world is still trying to catch up with the state of art established billions of years ago...

D75D6916-3D5E-49F2-AD9C-B4DF05CA2B0C.jpg

Humans have been cobbling things together with somewhat poorly developed engineering practice for a couple of thousands of years.
Similar constraints have produced structures in nature for billions of years.

We are at best still playing catchup in many ways..
The referenced tech will accelerate that process.
 








 
Back
Top