0:00 Trying. There we go. All right. Well, I guess I’ll start now. My name is Matt and if you’ve ever listened to one of my talks before, you know that I’m really, really bad at keeping track of time, unfortunately. So, this time my slides actually have a timer. So, if you notice that the color of the slide starts to turn coralally color, then someone just like, you know, raise their hand or something.

0:27 And it’s supposed to be my own internal cue to like shut up and keep going. So, you know, one of the things that I’ve seen definitely the theme of the conference is a lot of AI. So, I’ve got three things which maybe will make you excited, maybe not. I’m not really going to talk about any AI although you know, this will be a little bit more of a dessert, right, for all of the very heavy stuff.

0:47 This will by far be the less words that you’ll have to read on the screen. So, hopefully that’s enjoyable for you as well. And then similar to Andreas’s presentation before, this is going to be geometry ccentric versus sort of simulation solving, which we’ve seen a lot of and don’t want necessarily want to repeat the theme. How many I don’t know how many of you guys ever look at things that happen in nature or, you know, just day-to-day life and are ever impressed by like, you know, the structure of something natural or how something works.

1:39 In particular, I’ve always been fascinated by impactabsorbing applications because in general for for human engineering problems, it is something that if you look around in nature at what exists out there has been sort of solved in various intricacies all the way from the the micro to the macro. I have talked extensively in the past about the mantis shrimp, which is one of my absolute favorite animals just in general.

2:08 Super awesome sea creature. Plus, they’re really pretty, so look them up if you don’t know anything about them. And they sort of contain one of the the coolest, fastest operating biological mechanisms that exists in the world. And as such, the micro structure of their clubs is something really fascinating. On the flip side, a lot more macro scaled up is the the ram and the rams horn. And there’s actually quite a few papers out there about the rams horns.

2:33 Similar to the way my my brain thinks about a lot of this stuff every single time is less about you know obviously it’s solving a problem right but it’s about how we synthesize it and and make it available as a toolkit for engineers to use so 3400 newtons is essentially you know it’s like that’s like car crash impact right there right this is not a a small amount of force and this is something that they endure essentially you know multiple times a day weeks on end and I You know, if you look at the actual impact of the force, sometimes our brains can’t particularly compute what 3,400 ntons is.

3:09 But it’s essentially that’s like a human 10x concussion, you know, repeatedly over and over again, right? So immense amount of force. So again, looking at the natural, you know, these sort of events happening, the the first thing that always pops into my head is like, how on earth is this even remotely possible? Because the the structure of a ram’s brain is the exact same as a human’s.

3:33 It’s just the protection around it, right? That’s fascinating to me. So, as somebody who’s modeled 3D structures for the last decade plus, specifically for 3D printed devices that go anywhere from somebody’s body to, you know, a wearable or, you know, some sort of prosthetic or something. This sort of thing is fascinating and how we get that that sort of energy return. And if you look inside the structure of the horn, a lot of the times people look at the nice curved horns and they’re like, “That’s really pretty.” But it’s actually the horn core, the bone itself that is sort of torsionally carrying the load through the curve of the horn all the way down.

4:08 And we start to understand that it’s, you know, when we think about impact absorption, crumple zones on a car, you think a lot about like how much displacement can the structure go through, right? And still either m maintain intact or go through some sort of plastic deformation. Whereas the magic inside of the horncore is really more about attenuating the force rather than absorbing the force. And this is something when you look at like a force attenuation across curvature an isotropy or you know not regular loading characteristic is not the exception in biology.

4:42 It’s actually the standard. So if you look at almost any loadbearing mechanism it’s not just singularly anisotropic either. Look at a tree and the way that the the structures and fibers of a tree go throughout the branches. It is specifically isotropic. The exact load solved for every single point simultaneously right through natural iteration either through the growth of the structure over time over many many many years in the case of a tree or biologically adapted over time which you know I I like to say regardless of your viewpoint on intelligent design or whether it’s an adaptation over millions of years either way it’s the same thing it’s an adapted to a particular case right no one particular structure that solves a problem inside of biology is also singular it doesn’t just hold a load.

5:31 It may hold a load and also need mass transport, fluid flow, a temperature solution. So the things that we try to solve for as engineers inside of, you know, our real world problems, we can sort of look at nature. And so again, the the budget for the bone inside of the horn is not kind of it’s not magic. It’s the same bone as you you have inside of your body.

5:51 It’s just tuned differently. Sort of sheets, not struts. And these sort of sheets have a very specific way of carrying and attenuating that load throughout the force. So I really you know again was fascinated by this and with many of my problems in the past immediately I’m like what’s the proper analog mathematically to be able to solve this problem. So again, same material budget, 20 25% volume fraction, the exact same chemical composition of the bone that we have, but yet somehow it’s managing to hold that sort of impact.

6:26 So what’s you know recently I would say you know in the last 5 years or so there have been some interesting mathematical papers on spinids and sort of the spinoidal decomposition mathematically to create bone surrogates. So essentially spinoids are nothing super fancy. If you take How many of you guys like trigonometry? Anybody? Oh, that’s more hands than I expected. So cosine waves are essentially you know minus1 to plus one.

6:55 Most people when they see cosine mathematically they think of a wave. Mathematically it also describes when the the function crosses zero minus1 +1 -1 +1. So if you model that in 3D you essentially get a series of planes in space. So the cosiness essentially if you sort of jitter them a little bit so you induce some sort of randomness but you can still maintain alignment you essentially sum those waves.

7:20 So there’s a superp position of the waves and then you get this really neat sheetlike structure that you can control the anisotropy by controlling the directionality of the cosine waves. So you essentially get those summed waves and you sort of get this velar bone architecture of the rams horn. I hate saying for free. That’s like the new buzz word, for free. But you know, essentially for free, right?

7:44 The problem is, how many of you guys have ever modeled 3D structures before? How many of you guys Oh, a lot. How many of you guys have ever tried to like conform them to a like by convex curvature? You know, for things like armor absorption or things like, you know, cushioning and shoes, the problem comes when you’re trying to warp unit cells to shapes, you end up having some issues.

8:12 And does anybody know why mathematically this is? I told my wife I was going to say this. She laughed at me but the technical term for this is the hairy ball theorem and it’s essentially if you have a ball and it’s totally covered with hair and you try to comb it you can’t comb all the hair in one direction you’ll always have a cowletic that’s why we have calicks on our head so most of you are familiar with a spherically mapped TPMS right you always have these defect points or something that’s a genus of you or it’s a different genus a different oiler characteristic doesn’t have those ical defects, but for things that are complex in shape, you’ll almost always have these issues with warping the structure.

8:54 So, one of the things that’s really cool about spinodoids or gaussian random waves is essentially that you don’t run into the same problem cuz you can sort of control where your rotational symmetry sits. So, and then here’s our our issue. So, basically oiler characteristic of two there are issues for certain shapes there are not. And why in particular this is important is because we’re we’re able to use a structure again mathematically inside of the body that that sort of doesn’t have these sort of stress rising characteristics through through its characterization which I think again is particularly fascinating.

9:29 So you know one of the things that I’ve worked on o over the years is essentially how you synthesize these sort of tools using these sorts of things. So the other characteristic again for impact absorption is whether or not the cellular structure is bending or stretch dominated. These sort of rules mechanically were defined decades ago called the sort of the the Gibson Ashby theorem is they sort of are construct how cellular materials behave.

9:56 So certain constants that are part of the calculation are highly dependent on whether you’re stretch or bending dominated. And so, you know, again, when you attenuate a force, you may be able to increase the high peak that you’re going to be able to impact. But if you absorb the force, you’re going to be able to sort of soften the load at the back end at the expense of probably some sort of deformation, cracking, plastically of your material.

10:23 And the other thing that I think is fascinating, again, looking at nature, is very rarely is a structure monolithic. We think sometimes in engineering that like we’ve got to solve for a modulus at least again in medical devices like I need a modulus of this. The problem is that in nature again there’s no one explicit singular modulus that you’re solving for. What the body prefers and what sort of these applications are is sort of a steered and guided way of accounting for the forces either over time or through whatever the the application that you’re trying to do and not necessarily just a singular 400 megap pascals that’s my number.

11:02 So in a similar manner again the the the first sort of finding that came from working on this sort of idea is that you know that sort of steering along the load path became really important and one thing that you know I think is pretty fascinating again from from a tool standpoint is understanding which sort of forces right impact forces sort of sum where they conflict. If you’ve got something like a shoe and you’ve got various pressure points as you’re sort of mitigating certain impacts, the idea is do we attenuate the forces, right?

11:38 Do we redirect the forces into areas of stiffer materials or do we work on sort of absorption of the of the impact and those two things drastically change the the energy absorption curvature. So in some cases you may be looking at a sort of entirely different regime of your structure depending on whether you’re trying to absorb or attenuate and then the various ways that you sort of see those forces interact with each other over time whether they’re in the same direction or whether they diverge whether they’re opposite is I think particularly fascinating because we sort of see the analoges in in again in nature all the time.

Nature doesn’t explicitly give us a singular solution. It often times gives us a range of solutions. Again I I I I talked about this 2 3 years ago at CDFM which was we think as engineers and scientists towards converging towards a solution right there must be a best solution to fit my problem or maybe there’s a handful whereas nature sort of treats things divergently. If it encounters a problem, you often see divergence in nature, not convergence to a singular solution.

12:51 Thus why we have many many species of of various things. So to go back to to here, you know, the the the idea that you we can sort of solve the problem look at sort of the the crushing of the architecture over time. Define a three-dimensional solution or an architecture that solves the problem through attenuation or absorption. Is something where you know again we lots of people have talked about this sort of reverse interpolation problem.

13:21 Give the solution that you’re looking for and hope that the model sort of solves what you’re you’re trying to do. And you know, in order for that to happen, you’ve got to be able to mathematically be able to seed those sort of things into, let’s say, a machine learning algorithm, which a lot of smarter people than me are currently working on. But in order to solve for those problems, you have to be able to know some of the the information.

13:44 So, one of the things again, this this sort of tool is kind of like research grade just a analysis tool. If any of you guys are interested in sort of poking around with it, if you’re working on these sorts of applications, just I’ll send it to you. It’s a a super lightweight tool and these sorts of of math things are sort of estimated right right now. And some of the constraints, you know, again, I like to I don’t like to pretend like I have everything in order, right?

14:13 There’s always going to be some issues. And so certain things that I’m working on with this problem is you know it’s a hard leap to go from like quasi static analysis to you know transient. I was actually talking to somebody earlier today about like can we do fast transient solves in the GPU yet like instantaneously. The short answer is not really. I mean maybe with physics informed neural networks eventually but being able again to solve those systems equations over time is is something that’s pretty challenging.

14:43 The Gibson Ashby constants are unccalibrated and this is problematic for lots of different reasons but one of them is if you’ve got a cubic unit cell you you can test it over time and understand exactly what the constants are. If you have a graded architecture, you know, something where you keep the unit cells the same and you just thicken the struts, then you can still approximate your constants pretty well.

15:08 But if you’ve got a steered load through three-dimensional space and you’re sort of approximating that through these sort of stochastic pseudo stoastic structures, it becomes a lot harder to understand how again you crush those and simulate those. So that’s one of the limitations right now to working through these sorts of problems. Grading is analytical. You know again I I’m more f I’m less fascinated with volutric grading.

15:32 I think that that’s interesting but there’s a lot of smart people working on those problems. I’m more interested in this sort of morphological grading where you’re actually steering absorbing forces over time. And then one of the the cheap math things that that this sort of tool is doing is it’s sort of making assumptions that if you’ve got, you know, an impact load Oh, five minutes. Okay. I thought you had a question.

15:57 I’m like, yes. But no questions. Sorry, Don. No questions. So if you’ve got a you know two two stiff ends and you sort of have an impact absorption middle what happens when you apply a force to one side the center portion is what’s going to deform first. So the the cheap math sort of trick in this sort of application is to sort of just you know compute layer by layer and then sort of sort those rather than again solving a time dependent strain rate through time.

16:32 So this is actually an extension for those of you that you know follow me at all or we’re connected on LinkedIn. I’ve been working on a series of open source tools. These are sort of brower browser native tools. Don’t require any download. Don’t require sign in you know will never cost anything. And the idea behind the tools is that they’re just simply architectural metamaterial exploration. So one of the fun things that I I really like are these spinid materials.

17:01 And the grain tool which is in the the design hub there. The sort of armor impact simulation tool was built fundamentally off of the math that I already had worked on over the last 18 months or so. And so again, if you know these these sorts of tools are are free. You can use them now, mess around with them, throw some lattice in a part. And they of course export only as 3MF because 3MF is the world’s best solution for triangle and non-trial geometry definition and you know again one of the things that I want to stress throughout this whole thing is you know in these sorts of applications these sorts of kind of problems that we’re solving you know I would say my my advice and sort of suggestion to to everybody is to, you know, we’ve got so many ways right now to interact with these new tools that are out there, right?

17:58 Machine learning tools, you know, I I am fond personally of of cloud, but there’s just so many out there. And I think one of the problems that I’ve seen, especially with some younger engineers that I helped to mentor, is it becomes this deacto, you know, I run into a problem and instantly like solve this for me. I think the issue is again with all of these toolkits that are out there now for us, we sometimes lose the wonder of like evaluating and solving these these things for ourselves.

18:26 And I think the problem with that is again not, you know, I’m not going to pontificate on like whether or not humans are getting dumber because of AI, but I think the biggest thing is understanding again being able to look at something like nature and really sort of ponder like what makes this thing so special and how can we sort of utilize what we have in nature.

18:45 For these sorts of applications because again that that this problem has already been solved somewhere out there and so again I would just encourage you guys to to take a look out there see what it is and be able to apply that to the problems that we face today. So thanks a lot.