Rand Simulation: Democratization is Fine — Up to a Point

We trace how focused simulation wins over all-in-one platforms, then try to find out more about the design of an Olympic helmet  — with no luck. Rand Simulation experts use LS-DYNA and validate the results to cut risk and time. We close with a frank take on what AI can and cannot do for complex physics and where humans must stay in the loop.

• ANSYS ability to simulate crash behaviour, fragmentation
• GPU solvers move design feedback to seconds
• How designers and analysts collaborate without overlap
• Services, training, and keeping skills sharp
• Aerodynamic shell vs impact liner complexity
• LS-DYNA for explicit dynamics and complex foams
• DOEs to find worst cases across speed and temperature
• 3D-printed lattices promise and current limits
• Safety trade-offs comfort, mass, and adoption
• Where AI assists setup vs where experts decide

Transcript

Meet The Guests And Set The Stage

Roopinder 0:19

Hello, and welcome to FoDES, the future of design and engineering software podcast. My name is Roopinder Tara. On the show, we will have guests that will discuss tools and technology that engineers will find interesting and useful. So today we have joining us, we have RAND Simulation people and uh experts on simulation. Jason Pfeiffer is here with us. Jason leads the advanced simulation at Rand. And we have also Turner Jennings. Turner is an application engineer, has a doctorate in uh simulation, and he's uh I see here he's in the Boston area. So I gotta ask, how is it up there?

Turner 1:03

It's uh it's snowy out. We got probably uh a foot and a half of snow last night, so everyone's still digging out their cars outside the window here.

Roopinder 1:13

I saw some of the shots, man. I'm so glad I moved away from the east coast.

unknown 1:18

Right.

Jason 1:19

Roopinder, that that burly beard Turner has is coming in handy right about it.

Turner 1:23

Yeah, it's a it's a functional accessory.

Roopinder 1:26

It is, it is. You gotta keep every every part warm. So that's that's great. Yeah, I'm glad you're still uh okay. Did you have to shovel out? I saw some of the drifts are like piled up in front of front doors, and people are shoveling their way out.

Turner 1:40

Yeah, yeah, it was a workout for sure.

Roopinder 1:44

Oh man. So Jason, where are you? You're in, you're in a warm area, you're in.

Jason 1:50

Uh well, I'd like to say warmer. I mean, more warmer than Boston, but I'm in Charlottesville, Virginia. I've been here, well, at least in Virginia, for almost 50 years, and this has hands down been the coldest winter uh I've ever experienced. So there's something going around. It's unfortunately pretty chilly here, also.

Roopinder 2:09

Yeah, oh yeah, I guess it is. Yeah, so warmth is a relative thing. Yeah. I think you're were you originally with the um Blue Ridge people? Yeah.

Speaker 3 2:19

Yeah, yeah,

Careers, Blue Ridge, And Autodesk Era

Jason 2:20

yeah. Not sure where how you knew that, but uh either way, good on you. Yeah, so I was the global VP of engineering uh at Blue Ridge Numerics. I started there in 02, uh, worked my way up to lead the global engineering team. Uh, I created a uh a services arm of that business starting in 2008. And then when we sold that to Autodesk, and I think it was March of 2011, uh, I took some of the capital as an executive there and hung my own shingle, created my own little CFD consulting company. It was called Informative Design Partners, which, you know, about a year and a half later, I sold to Rand. And that's how, you know, I personally joined Rand, uh, what is that almost 12 years ago now? And uh we operated a small CFD and structural consulting business unit under Imagine It, uh, the big business unit here at Rand Corporate. And then in 2018, uh started getting courted by ANSYS. They were looking to grow their partner network in North America specifically. And then we legally, repender, had to create uh a separate business unit because Autodesk, of course, has some of their own simulation tools. And that's when RAN Simulation became an official business unit, its own separate PL. So we've been doing this, it is our eighth fiscal year, and uh yeah, just just having a ton of fun, man. We're uh we've in in eight short years reached the highest. I'll brag a little bit, but reached the highest uh partner level in the ANSYS ecosystem called Apex. There's only four partners globally out of about 250 that are at that level. Uh, and we've done it in eight short years. So yeah, really proud of that and having a blast.

Roopinder 4:07

That's great. Yeah, what a what a history to have in CAE. Blue Ridge Numerics, if I'm I'm trying to remember, they were were they acquired by Autodesk. Autodesk, Autodesk, that's right. Autodesk. And I want to I want to go into that a little bit because Autodesk acquiring Blue Ridge Numerics along with a several other CAE companies, including uh NASTRAN.

Why ANSYS And Building Rand Simulation

Jason 4:32

Yeah, they bought the NEI version of NASTRAN. They they bought uh another, I forget what it was called off the top of my head, but another uh kind of FEA Light tool before they bought Blue Ridge. Of course, they bought MoldFlow. They they went on like a billion dollar simulation buying spree between the early 2000s to about 2015.

Roopinder 4:53

Yeah, and suddenly they started saying uh we're a CAE company now. So that's and I had to laugh. I'm sorry, I had to laugh because uh I always thought you know, I knew they were a CAD company, and but they were trying really, really hard, spending a lot of money. I'm gonna read a lot into why you didn't go with them. And I'm thinking you wanted to say hardcore simulation and not be become a well, yeah.

Democratization Debate And Designer Tools

Jason 5:21

Listen, nothing against Autodesk. I mean, Imagine it, as you know, is the world, our sister division is the world's largest reseller of Autodesk. So even though my group has its own separate PL, I mean, I'm still dependent on the success, you know, that Imagine, and therefore Autodesk has. Um, and I think overall it's a it's a great company. You know, it's they got a lot of diverse technologies. I um, you know, I think it's a challenge, uh, and this kind of gets into why I decided to, you know, go with ANSYS. I think it's a challenge to split your development across that many areas. I'm a big believer, repender, that focus equals effectiveness. Uh so like in the ANSYS world, yeah, they have a plethora of simulation tools, but that's all they do with simulation. You know, they're pouring half a billion dollars or more every single year into simulation and simulation only. So it's no wonder they're the 800-pound gorilla, and these tools are known for speed, accuracy, robustness, and ease of use. You know, if you're Autodesk or PTC or Dassault, you're splitting roughly the same amount of development dollars across simulation, CAD, data management, PLM, and they're all you know fine tools, but it kind of is that you know, uh jack of all trades, master of none sort of thing. So I kind of saw the writing on the wall when we sold to Autodesk that they were gonna try to democratize it a little too much. That's what Autodesk, and it's a noble task, you know, but I just felt like simulation in general, regardless of who the OEM was, simulation, even I think to this day, it's not possible to truly democratize it yet. It just still has too many nuances. Now, maybe AI changes that, you know what I mean? But but back then, remember, this was 12 years ago, simulation was still too nuanced, and you couldn't just give it to a fifth grader and expect them to go run, you know, some million-celled model with all kinds of complex physics. So I just, and it was the right decision. Um, because when I started my little consulting company, naturally CF design, that was the tool back then that we sold. You know, that's what I knew. And it was a great tool, but I just had this sneaking suspicion it was going to be democratized to a point where I couldn't continue to run a successful, profitable consulting business. So we we had to switch to something different.

Roopinder 7:48

Right. Attorna, I haven't forgotten about you. I'll promise I'll be right back. I'll tell you how this started. Peter, your your ace PR guy, he was uh saying, oh, we designed a we helped design a helmet for the Olympics. I thought, well, that's pretty timely. I gotta get it, I gotta find out more about that. So we're gonna get to that and all your helmet research, but I can't resist but tell I talk about democratization because I have been hearing that word, that awful word for so long. I mean, I don't even think it was a word before Autodesk bought simulation companies. But all of a sudden I heard it galore. And I actually sat in on one presentation by Autodesk at their Lake Oswego. And this was right after they bought a whole bunch of uh, I think after they bought Blue Ridge numerics, maybe, or after ANSYS, I think it was after they, I'm sorry, after they bought NASTRAN, and they did a whole stress analysis for us just to show how they could do stress analysis. Or just uh, you could just be a designer. Uh anybody could do this. You just hit a few buttons and there you go. And they showed us the and then they showed us the you know the contour plot of the stresses, and I and I asked them, I said, What uh what uh what failure theory did you use? Von Mises is for that because it was aluminum part. And and they they were like the blank, I got a blank look. Like I think they I they didn't say this, but in my mind I I imagine them saying, Von Mises, how do you spell that?

Jason 9:21

Right. Well, well, listen, I I will say, Roopinder, I mean the the day is coming. Uh I mean, you know, democratization of simulation, it is coming. Uh, and and I think AI will empower that and accelerate the timeline for it to become a reality. But I tell you, though, there were many reasons I decided to partner with ANSYS and create ran simulation. But one of the things for me, now mind you, this was back in 2018. At the time, you know, it was, I think they called it ANSYS Discovery Live or something. ANSYS is the closest to really democratizing and shifting left. That's the big word, right? Shifting simulation left to the designer level where earlier stage where the uh cost of change is lowest, chance for innovation is greatest. They've cracked the code more than anybody else in that it's running on the GPU. So this is really what sold me with ANSYS. They took a model that you know would take like 24 hours, like external aerodynamic CFD, 24 hours to solve on a CPU, and they did it literally in 60 seconds.

Speaker 2 10:26

Yeah.

GPU Breakthroughs With ANSYS Discovery

Jason 10:27

And it was spot on answers. And now, of course, that tool, Discovery, has evolved over the last eight years. So it's gotten infinitely better. I think one of ANSYS's bigger challenges now is everybody loves the discovery interface, and to the point where they want that probably to start becoming more prevalent and become the base interface for a lot of their technologies. So they're putting more solver functionality into it. That tool historically has been a lower entry price point. So if they're not careful, they're gonna put too much into it, and therefore their average deal size potentially could go down. So I think that's a a challenge, but they've done a wonderful job, wonderful job of you know, supporting the democratization goal of simulation and uh what they've done with discovery is nothing short of amazing, really.

Roopinder 11:18

Yeah, that is amazing. I have to say that when I saw that real time, real-time analysis simulation. I'll you know, it realise we have we have to say though that it was for lower, how should I say lower level simulation, easily understood simulation, if you can say that. Uh you know, stresses, uh deflections, it was pretty good. But when you get into complex stuff, right? Complex stuff. You get into like impact and you get into crash analysis, stuff we're gonna talk about, like helmet design, right? Then then like all bets are off.

Jason 11:52

I don't think you could no, but but I will say this, and then you know, we'll we'll migrate over to kind of the the helmet stuff, but I will say this uh again in support of what ANSYS is doing, which testament to their development staff there. A, it's doing more than just basic linear statics now on the on the structural side, they're putting more and more solver functionality in there. But unlike the example you you said before where people are like, Hey, I can't even spell von Mises. This is using the actual mechanical solver behind the scenes, it's using the native fluent solver behind the scenes, you know what I mean? So it's it's not just pretty pictures that you get in real time, it's high fidelity solutions as well. It's unbelievable.

Roopinder 12:33

Yeah, yeah. I have a PE and uh, you know, I have a pledge. I have a pledge to uphold the safety of what I do, of safety of the public and what I do. And uh, you know, I've always said I wouldn't want to, I would want an engineer, a responsible engineer, somebody who knows what they're doing to back up the results of a first-order simulation, right? That okay, go ahead, designers, go at it. You know, do your do your thing, but let me check what you've had before or validate it somehow, right? Somehow.

Analysts And Designers Working In Tandem

Jason 13:05

I'll give you one other example, and this kind of gets into the sports uh uh that we're gonna talk about. So Wilson Sporting Goods is one of our bigger ANSYS customers here at Rand Simulation. They have quite a few analysts, like you're talking about, driving the hardcore Fluence CFX, mechanical, LS Dyna. Matter of fact, they just used LS Dyna to innovate their latest Wilson irons, golf irons. It's that was a press release. There's all kinds of videos you can see online now. It's it's phenomenal. So great account for us using the traditional hardcore analyst tools like Dyna, but we're now working with about a hundred CAD designers to do the first order stuff. So we're not trying to replace the analyst. You know, the power for them and what their management is excited about is they can go in now with Discovery when they're early on in a development phase of, say, a brush roll or you know, a plenum or whatever the case may be, and do some quick what-ifs on, you know, stress, thermal, CFD, airflow, uh, and kind of optimize it to a certain extent and then send it over to the analyst, which is just going to make the analyst even more uh that role even more streamlined. They don't have to run as many hardcore 10, 12 hour solves. They've kind of worked through the kinks and got a decent starting point before they throw it over to the analyst. So companies are starting to figure out how to marry, if you will, simulations done by the designers and the analysts, such that it's saving them all time and producing better products uh, you know, more quickly and efficiently. So it's it's a fun time, frankly, to be in simulation.

Roopinder 14:53

Yeah, yeah. Yeah, from where you sit, I imagine. Like you guys are the experts, and hopefully they just these design companies or these manufacturers will call on specialists when they need them for the more advanced simulation. That that is a service that ran simulation provides, right?

Services, Training, And Real-World Value

Jason 15:08

The absolutely. I mean, at our core, I mean, don't forget, I started a CFD consulting company. So we consulting is a is a vital component of our of our business. Uh, quite frankly, I don't see how people you know represent any simulation OEM if they don't provide consulting services because that's the only way guys like Turner keep his skills sharp. Yes, as a PhD. Uh, you know, but if he's not using the tools that we're coaching our clients on how to use on a daily basis, eventually, you know, he's he's gonna uh you know plateau a little bit, so to speak. So it's a great way to solve real business problems, show your customer firsthand the value the tool is going to deliver. So there's no buyer's remorse, there's complete confidence. Then we can train them specifically around one of their own models because their attention spans and retention rates are naturally stronger on their own stuff. So yes, it's a it's a uh wonderful uh component of our of our business model for sure.

Roopinder 16:06

Yeah, yeah. And uh I just hope that they do call on experts when they do need them. That's my fear is that they won't, you know, and they'll try to put in put into a the market products that could fail and they don't know better. But anyhow, I'm gonna stop. I'm gonna stop preaching. Uh, but okay, now about this in the PR that I first saw, first first time I heard about this helmet design about the Olympics, it didn't mention the company. So I've this is one A design company. Are you allowed to say which company that was? No, oh okay.

Jason 16:40

I mean I could say D2H, which is the services provider in the UK that we work with, but can't share their company name or their customers' name.

Roopinder 16:49

Uh but that is sort of what you do for companies. If they need simulation, you'll go to them. And they they came and said we need a better helmet for impact, correct? Not for not so much for aerodynamics. This was just for impact.

Olympic Helmet Project: Scope And Stakes

Jason 17:02

Well, no, uh 100%. I mean, aerodynamics is part of it, right? If you read through the case study, step one is they needed to create a skin or a shell of the uh of the helmet that met all the you know uh lift and drag, you know, aerodynamic performance goals that they had. Frankly, that was the easy part, you know. But but then, you know, with some of these sports there, uh, and there's a variety of sports, we just all watched the Olympics where where helmets are required. So some of the impact that these athletes can feel, you know, when they when they take a spill, can be, well, frankly, they can be fatal. You know what I mean? So the challenge with this, much more so than the aerodynamics, was designing the interior, the guts of it, the polystyrene, you know, that would absorb that uh in such a way that um, you know, the G forces on the brain were not extremely high and and catastrophic, right? And that's where Dyna came in, you know, a very really the the the best in class explicit solver, which Turner has his PhD uh in this, not only a ton of experience in LS Dyna specifically, but in impact analysis with LS Dyna. So Dyna is used, repender, when you know, and this is why, you know, crash, whether it's a a helmet or uh a vehicle crash, it's very big deflections over a very, very short time frame, like milliseconds, microseconds, uh, which is a completely different set of physics. And that's where Dyna really shines. So that was the challenge for this company in designing the helmet was the internal such that it would absorb it quickly without having G forces that were catastrophic. Turner, I don't know if you would add anything else to that as the expert with it, but yeah, no, I think you summarized it really well.

Turner 18:48

I mean, the challenge there is you have a very thin space between the outer shell of the helmet and the skull, and you need to absorb a lot of energy in that very tight gap, you know, absorb or redirect somewhere else so that it's not entering the head, entering the brain. Got it, got it.

Roopinder 19:06

In that small space is essentially hair. And I heard it said uh you can't depend on hair for any type of insulation, even yours.

Jason 19:16

Yeah, now if if he wrapped his beard around the hair, yeah, if I kind of pull this around, maybe we can get it. But but that's that's really where the value, even outside of Dyna, you know, but the value of this one, they had to do it in short order because you know the Olympics were right around the corner at the time. Now now they're behind us. But you think about the helmet, not just the aerodynamic aspects that they did pretty quickly, but think about the different shapes and sizes of athletes' heads. So one helmet has to accommodate all of that, how it's going to react and how much space between the skull and the polystyrene foam is gonna be different every time. Then what about the the you know unlimited uh permutations of speeds, of angles that it's impacted, of temperatures, because you know, the foam that they were using was very temperature sensitive. So all those material properties varied with temperature. So there were just literally an infinite number of permutations that in a perfect world needed to be analyzed. So they were able to set up DOEs, you know, do a lot of automation and run uh you know a huge permutation to very quickly, in the matter of days, repender, identify the worst case scenarios so that then they did all their development effort, you know, around the worst case. Because of course, if you if you pass in worst case, you know you're gonna pass in in all the less severe uh scenarios.

Roopinder 20:44

Got it. So this company said, oh, we got we have we want to make a better helmet, and you just happen to have a helmet specialist, somebody who had their doctorate in.

Jason 20:54

Kind of ironic, isn't it?

Impact Physics And Why LS-DYNA

Roopinder 20:57

I don't think you had a lot of competition here in that for that project. So but tell me this. I have to ask a very it's probably a dumb question. Okay, so at that level at the Olympics or any major sport, by the way, just blink once if it was the uh if it was the high jump. It was or that's the only thing I can think of that's related to really cold weather, winter Olympics, and really have high speeds.

Jason 21:25

Well, think about it. You got luge, skeleton, bobsled, downhill, super G. I mean, there's there's a lot of helmets in the Olympics.

Roopinder 21:32

Okay, all right, okay, got it. I won't press on that one, but uh oh okay, so I gotta ask this. Is my dumb question. So you said there's a space between your hand, between your skull and the helmet. Uh can't that, knowing what we know about 3D printing, why can't I make a 3D printed helmet there? Where there's no space, right? Would that be like the perfect helmet?

Jason 21:56

Yeah, I mean, some people are doing that now. A lot of the newer football. Helmets have 3D printed lattices in that space, and there's uh a lot of research in that area right now. So it's coming down the pipeline for sure. The 3D printing gives a lot of flexibility in tuning the parameters of force deflection in that little gap there. And um, yeah, you know, it's one of the technologies that's out there, and simulation does a lot to enable it, right? You can figure out what the optimal force deflection is and then go to your 3D printing software and create a lattice that will exactly match that. So correct me if I'm wrong. I think one of the challenges with that is you know, there's only certain materials you can use today, anyway, for 3D printing, right? So I suspect the polytyrene foam that you know typically is going into these helmets, which is mainstream today, probably can't be used in 3D printing. So that's probably another reason that folks aren't doing you know what what you're saying there, Ropindender. We probably will in the future, as as you know, technology evolves with 3D printing, but I suspect that's an issue. Uh Turner, please correct me if I'm wrong there, but no, you're 100% right.

Turner 23:04

And you know, that's another case where Dyna really shines, is there are so many different material models and you know, ways to capture the complexity of a lot of different specific materials. And so the helmets come in as a really good use case because there's all these really complex foam models that give you strong accuracy in modeling these different you know, helmet liner materials.

DOEs, Temperature Effects, And Worst Cases

Roopinder 23:30

So is it am I correct in thinking that if you reduce the air gap to zero and you have like a somewhat of a harder material, more than soft foam, you could actually make a make the perfect helmet? Perfect uh is there such a thing? Maybe there's no such thing, right? Yeah, I mean, uh I heard that by the way, I heard it said that you can add all the protection you want to athletes, and athletes will still find a way to hurt themselves. They'll just like in football, they'll just run, they'll just run at each other faster because now they think they're protected.

Turner 24:03

Yeah, yeah. I mean, the the perfect helmet is probably six feet of foam, but then no one would wear it, right? So there's there's always the balancing act there of how do you make something that's safe that is also performant for the task that you know you're you're working on, be it football or be it the Olympics. Uh, you know, you need it to be comfortable and effective for the task at hand as well.

Roopinder 24:26

I got okay. I want to I want to talk more about the Olympic helmet and what you did. But I've got to ask one more question because I watch these football games, right? And I'll oh it's over now, Super Bowl. I'm in some kind of a withdrawal. But everyone, somebody on the team always has extra liner on their helmet. What is that? And why isn't everybody wearing that if it is if it is useful? Is it a matter of vanity? Like it makes you look funny or what? You know that you know what I'm talking about?

Jason 24:54

That extra make them look like a mushroom.

Roopinder 24:57

Yeah, all right.

Turner 24:58

Right. Who wants to look like a mushroom? Yeah, I think that might come down to personal choice. I know it's a lot more common in practice sessions, you know, they'll wear that extra soft shell cap on the outside, right? Um, you know, just an extra bit of padding to keep the head safe. But um I don't know, I've never worn one.

Roopinder 25:17

Has that shown to be effective, that extra padding?

Turner 25:20

I imagine it is, but I believe there's been some research that's shown it reduces you know the force transfer onto the head.

Jason 25:27

I I suspect, Roopinder, what that is is the players that are using it probably are the ones that are coming off of uh concussion protocol and they can't get back on the field if if they don't have it on, would be my guess.

Roopinder 25:39

But uh that makes sense.

Jason 25:40

That makes sense because I you know, you you know these mainly football players. I doubt any of them want to go out there looking like that. You know, they they're not they're not doing it on purpose.

Roopinder 25:49

Right okay, all right. I gotta I can't stop. I got I have experts, I have experts at my back end call. I can't stop asking questions. I just asked one more. Which sport is the would you say is in most need of good helmets and yet doesn't wear them? Because I just watched yeah, hockey. And those guys, that puck is really hard. And most of the guys aren't the guys aren't wearing helmets, right? The players? Okay.

Jason 26:15

No, they are.

Roopinder 26:16

I think they're wearing helmets most of the time, right? Yeah, yeah. Okay. Okay. I totally drew a blank, but but not enough. I don't think the helmet's good enough.

Jason 26:23

I don't think it's it's a great question for Turner. I mean, this is what he got his PhD in. What's shocking to me with hockey is yes, they they are wearing helmets these days. They didn't use two decades ago, but I'm still surprised that like the lower portion of their jaws are not covered with the helmet, and those pucks flying at very high rates of speed seems like it could do some serious damage. Um but as to which sport is in the most need of uh you know better uh helmets. Uh Turner, that's that's a great question for you, man.

Turner 26:54

What do you think? Yeah, yeah, it's a good question. I hockey's definitely one where the technology can always be improved. I mean, you watch those people get thrown into the boards and it it's scary. Rugby's another one. You know, they've started to make more soft shell caps to protect rugby players, but you know, there's there's definitely adoption to be done there.

3D Printing Lattices And Material Limits

Roopinder 27:19

Rugby players just don't have helmets, they're they're too manly.

Jason 27:24

They're animals. I don't know how they do it. Yeah, takes some gumption to run run at a 250-pound grown man running full speed, hit him head on in no pads whatsoever. Those guys are nuts.

unknown 27:36

Yeah, yeah.

Roopinder 27:37

So rugby is uh sort of what like American football started out at. I read the most interesting short history of uh American football. It was Ivy Leagues, and it took them a long time. It actually took a presidential decree. I think it was Teddy Roosevelt, whose son was playing at one of the Ivies, and his he was afraid his son was going to get killed because it was very routine for players in the Ivy League to well, not wear helmets, be gashed, bloodied, concussed, but they were proud of it. And then they forced the league to adopt leather helmets or whatever, but for what they're worth, right? Almost nothing. But yeah, I think that rugby was a very strong proponent of the original American football. It's a very interesting story. But how far we've come, and yet okay. Whatever you can, whatever you're allowed to tell. So about that, Helvin.

What Makes A “Perfect” Helmet

Jason 28:39

Well, I mean, I I think you know, it's the work that D2H did. I mean, D2H is a customer of ours. You know, we help kind of solution map the right technology to you know the application need that they had. Um, but you know, they did the work and they're very strong analysts. I mean, they're they're great, but I mean the work, the stages they did on this, again, they they first designed it aerodynamically, the the skin, that was the easy part. That's classic, just kind of external C you know, arrow or CFD. The most challenging part, which is where Dyna really came in, because the external arrow was done more with fluent CFX, you know, more traditional CFD tools. That the big challenge was all the nonlinearities of the polystyrene phone. And this is just a testament to how strong D2H is. They didn't just throw everything together in one model and see what happens. They already had real-world test data from a lab on how this material responds under pressure and short duration impacts from lab tests. So they first isolated that in an LS Dyna model with all those unique material characteristics and models that Dyna has. Turner mentioned earlier. And you know, even if you read the press release, the white paper there, they were really pleasantly surprised at just how accurate the correlation was between what the Dyna model showed on that sample of polystyrene foam and what the physical test showed. So that gave them the confidence repender to then go off, put the the foam in the actual skin of the helmet itself and do more of a system level analysis. So they they did it pragmatically. They they they got confidence at a smaller level, uh, so that when they went to the system level and they started getting results, they knew those results were dependable. There's nothing worse than just assuming the model's right, you know, and then you go run this huge DOE only to find and you come up with what you think is the best design, but it's actually a terrible design. You never validated it against something. So that was the step for them was validating the uh the foam to reality, which gave them the confidence to blow through all of the iterations, identify the worst case, and optimize around that. All right.

Roopinder 30:58

So if I got this right, you validated the test, you validated the simulation based on the test results, and then you're able to do a lot more what-ifs, a lot more variations, and you were confident your variations would work. Was that the did you validate again at just that initial time with the test?

Football Caps, Concussions, And Adoption

Jason 31:20

They I honestly I don't know if they did some you know, system-level stuff on the helmet at the very end, but they did not do any more uh validation on the foam material itself because that the first one correlated so well, and they correlated it across a variety of different uh you know, anvil impact speeds, forces, temperature ranges. They really vetted that out because that's the complex part. And again, a testament to Dyna. You know, I kind of equate this to like uh tires, you know, trying to simulate do an analysis with a rubber tire is actually quite complex. That's a very complex material. Dyna is one of the only tools, if not the only tool, that can handle such a complex material. And this foam kind of fits into the same category. So they vetted that out, confirmed the validation so that they can move forward with the actual design aspects of the project with complete confidence, knowing that not just the trends, but the qualitative numbers and quantifiable numbers that they were seeing were very representative of reality.

Roopinder 32:24

So they're playing what they are they playing with, it's still the same material, polystyrene, but it just doesn't uh you they're playing with the shapes. In fact, they're varying the shapes and everything to make it correct to test. Okay, got it, got it. Very interesting. I I'm remembering when I went to a there's a computer museum out here in the Bay Area near San Jose, and uh I saw one of the first cases of fragmented fragmentation. It was a vehicle crash back when they were first doing simulations to um to validate or to to look at crash simulations. So they wouldn't have to test so many dummies that they would actually do crashes. And I was amazed by how the pieces flying off the car, I mean literally coming off the car had the same shape uh in real life as they did in the simulation. That to me just seemed like magic. It's somewhat chaotic how the stress could move through the material, but to actually predict where the fracture would occur, that is that that that's the power of Dyna. Yeah, I I don't what how does it do that? I how I can imagine if we get the number of pieces right, perhaps. But can is it that accurate? You could actually see the shape of the pieces, or is that was that just a lucky simulation? This was a function four years ago, right?

Turner 33:58

Yeah, no, it can be that accurate. And it I think that's one of the really, you know, for me, that's one of the fun parts about being in simulation, right? Is every now and again you'll throw together one of those results and you'll pull it out and you'll have something to compare it to, and you're like, man, this is an exact match to you know what we're seeing in the real world. It's a really cool moment when you you get to that point. Yeah, yeah.

Which Sports Need Better Protection

Jason 34:21

Roopinder, little little fun tidbit for you. So this was over 20 years ago. Uh this is when I was actually driving a lot of simulations myself these days. You know, I not so much. Turner, Turner gets that that joy, and I get to scratch my itch by watching the fun things that that he and the rest of the engineers are doing. But back in 2002, when Dale Earnhardt died in the uh Daytona 500, the company I was with at the time, we were we were contracted by the big three to simulate that crash. Uh, and and we used LS Dyna. So a lot of the car of tomorrow, if you follow NASCAR, came from that work. University of Nebraska was heavily involved, you know, in that. Um, and that was again 20 some years ago. So you can only imagine how much you know that technology has evolved, you know, over the past couple decades uh and more. So, and back then, I mean I I can't show it, but I mean we had side by side images where you know it was his crash because we simulated that crash at you know that angle of attack, that speed. Uh, and it was amazing when you had it side by side, because in a simulation, you can make it go as fast or slow as you want. And it was just like you described, you know, not only was the deformations and the magnitude of those deformations on point and in the right areas, but things that were flying off of the car were picked up in the simulation. And again, that's what makes LS Dyna and probably arguably the most powerful all-encompassing simulation tool on the planet, which frankly is why ANSYS acquired them a few years ago. Uh, I mean, they have phenomenal technologies on the FEA, the structural side, and ANSYS mechanical, ENCODE, they got a plethora of tools over there. But I mean, Dyna just had the brand recognition of being the explicit tool out there. So it was a it was a wonderful acquisition for them, and it's really opened us up to you know, new markets, additional technologies that we can offer to uh as a complement to some of the folks using mechanical, you know, where appropriate. And that's why we brought Turner on board, you know, because Dyna just is it has such a loyal following, and we have a lot of mechanical users where mechanical, and says mechanical, their flagship structure analysis tool is all they need. But but then there's certain applications where it's like, you know what, we probably need to educate them, you know, maybe Dinah's a bit better for that one specific application. So Turner's been really instrumental in helping us uh better represent Dyna to our existing and and prospective customer bases.

Roopinder 36:57

Okay, very impressive.

Jason 36:59

Because Dyna also does, I mean, everybody knows it for like crash analysis, all this explicit stuff, but uh it does a lot more than uh than just FEA structural analysis. Maybe maybe Turner can speak to that a little bit, but it's a it's a very powerful, diverse simulation tool set.

Roopinder 37:16

Large deformation, too, if I'm not mistaken, right? So okay so low, so short duration, low, low, uh low deformation, small deformation, all the way to large deformation, like for rubber and everything, right? So definitely the high end of the physics uh spectrum as far as uh as far as tools go, we're talking about democratization at one end of the spectrum with doing simple stuff, and then at the other end, really intense stuff like this, right?

Correlation Moments And Crash History

Speaker 3 37:46

Well, I I can see Turner smiling because if there's a tool, at least as it exists today, that should not be democratized, it's LS Dyna. I mean, LS Dyna is a power user's tool.

Jason 38:00

You know, it's it can do whatever you want it to do, but you have to know how to make it do that. If you know what I mean. So it's uh, I don't know how what how would you describe it, Turner?

Turner 38:13

I think you you hit the nail on the head. It's a steep learning curve when you get started. You know, it's it's got three manuals that are each a couple thousand pages long, and you really have to go through it all to understand what's going on. But you know, once you get into it, sky's the limit on what you can simulate, everything from these rapid crash problems to you know metal forming and manufacturing. You can simulate your entire product lifestyle uh lifecycle from creation to eventual failure. Uh, and then it's just down to sky's the limit on what you can simulate, what is the case where simulating it will save you time, save you money, save you risk. And you know, that's our job here at RAND is to help the customers kind of figure out what are those cases where this is a strong tool for adding value to the current you know product development cycle.

Roopinder 39:07

Yeah, I totally agree. Uh that this, you know, this some tools should just not be democratized and certainly LSD being a very sophisticated uh software.

Jason 39:17

Well, and that's the business challenge, right, Roopinder, that you know, companies like ANSYS and all the other players actually have at this point, right? I mean, from a market size, yeah, you want to democratize it for obvious reasons. You know, for every hardcore analyst, there's probably a thousand you know, designers uh or entry-level engineers. So it makes sense why, as a business, they want to you know make it more automated, more user-friendly, you know, trustworthy without having the PhDs and uh you know, computational fluid dynamics, for example. But at the same time, some of these tools like Dyna, just for the time being, I mean you just cannot simplify it to that point. And that's why, again, I chose ANSyS to partner with when I decided to go become a reseller back in 2018, because they're they're going after and penetrating and have phenomenal tools for the larger audience, you know, the designers like Discovery, but also still have those really sophisticated, specialized, high-end tools that you got to be a certain kind of caliber engineer to even use. So they they really have solutions for every engineering type, every industry, every piece of physics. That's why they're the 800-pound gorilla.

LS-DYNA Beyond Crashes And Metal Forming

Roopinder 40:33

Jason, you mentioned earlier on you mentioned AI. So I'm gonna make this the final question because you guys are probably have better things to do than talk to me. But uh, and we're approaching our time limit. But I wanted to talk, you know, no conversation can be complete without AI. So I gotta ask about it just as an obligation. I'm also gonna play devil's advocate here because you said, okay, we could never, we both agreed we can never democratize something so sophisticated.

Speaker 3 41:02

I didn't say never.

Roopinder 41:04

Okay, all right, all right, all right, good. Because okay, here I go with this. I you could argue that tools as sophisticated like this, because FEA was at one time, even simple FEA, if I can call it that, was thought of as very sophisticated stuff. You can only have PhD engineers do it. Now we're saying, okay, you can have PhD engineers do the sophisticated stuff like simulational crashes. Uh but this is a moving scale, right? And AI has the promise of making these things go under the hood, as it were. Like I can use natural language interfaces because LLMs, you gotta admit, are really good at reading documentation, remembering documentation, like the whole manuals for all the manuals that you know you spent months absorbing for LSDA. Give it to the AI, right? Give it to LLM. And now I can ask simple natural language interface questions to Ellis Dida. I don't have to maybe read. Do I can I ever say I don't have to read the manuals now? Because I can just say ask simple questions using AI. Could you see a world where that could exist? Where that I could just a typical engineer.

Tools For Every Physics And User

Jason 42:24

So, you know, Turner, you know, he he's the one living and breathing these things, but in the short time, because AI, I mean you're right, it's the buzzword. It's not just in simulation, it's in note takers, it's in it's in PowerPoint present, it is everywhere. And it's actually scary, quite frankly, how good a lot of these are. And ANSYS, you know, started thinking about this quite a few years ago. They're always on the cutting edge. So, you know, without getting into too many details, because I'm not a developer over there, but they're using you know large language models. So whether it's you know, fluent or dyna or mechanical or you know, SPIOS, doesn't matter what it is. Think about some of these larger, more established companies that have years, some decades worth of analyses that they have run in that program, right? That it spans the gamut for any possible type of you know design they would ever want to simulate. That now can be fed into a large language model that just keeps getting smarter and smarter and smarter to the point where you could be an absolute novice, you know matter how complex the numerical methods are behind the scenes. You just basically say, here's my model, here's what I want to get out of it. And it just because it has thousands of data points, millions of data points that it's learned from, you almost can't mess it up. So, you know, again, I I think Dyna probably would be the most challenging to do that for for the reasons we've already talked about. But as quickly as AI is developing, I think it's just a matter of time where something that complex could be. Driven by, and I don't even know if driven is the right word. I mean, there comes a point where you almost, you know, just uh type in this is what I want, and it just does it for you. You don't have to know anything about simulation. So that time is definitely coming. Turner, what do you think?

Turner 44:15

Yeah, I mean, I feel like I have watched the progression of some of these models over the last couple of years. And, you know, they're still not at a point where I would trust it to tell me how to set up a model and here are all the steps you need. And I would I would just go and trust that. But looking at where it was four years ago when I would, you know, try and ask a question about LS Dyna and then looking at it now, uh, it's leaps and bounds better, and it will just keep moving that direction. So, you know, we're not at a point where I would say it's a tool that you know I use every day as my co-pilot for LS Dyna, but you might get there sometime pretty soon.

AI, LLMs, And Simulation Co-Pilots

Jason 44:58

Well, and that's that's combined, repender, with uh, and this was a big part of the strategy behind Synopsis acquiring ANSYS, right? Ys whole kind of silicon to systems. But I mean, look at silicon and semiconductors these days. And I forget what the latest statistic is, but you know, the data transfer rates, even though the chips are getting smaller and smaller, the data transfer rates are doubling. I forget about it. You know, it used to be like every seven years, now it's every 18 months. I mean, don't quote me on that. But uh the point is we have the computer hardware evolving at an even more rapid pace where you can do these large language models and have it, you know, scan all the previous results that you've done. So it's not just simulation improving and evolving. It's you have other technologies like hardware that support that, also developing at just as fast, if not faster, a rate, all kind of holistically merging together, which is making AI just an absolute beast these days. My my concern personally, as a human being with AI, forget, you know, simulation is just the atrophy of the human brain. You know what I mean? Like we still should be using our brain. We still should think creatively and critically. And I think a time is coming where, I mean, listen, the computer is always gonna be smarter than the human being. You know what I mean? But I think we're designed to have human relationships. So I think the people that are gonna win in this AI race are the ones that, yeah, pragmatically and efficiently use all this amazing AI technology, but don't lose sight of the importance and criticality of human-to-human interaction. You know what I mean? And still the ability to think on your own and not just have a machine tell you every step of what you should do in your life. You know what I mean? Not to go on a tangent or my own soapbox, but uh, I think AI, if used appropriately, is a wonderful, phenomenal new arrow in our quiver that can make all of our lives easier, happier. We just can't lose sight of the fact that we still have brains that we need to use and we need to learn from the AI. You know, it's it's I mean, even if it's just chat GPT, the things that you can learn on chat GPT in a matter of seconds is nothing short of amazing, but like at least read it, learn from it. You know what I mean? So when you're in front of an executive and you don't have the bot telling you exactly what to say to sound knowledgeable, you've learned it, you've absorbed it. So people still need to learn from the AI.

Roopinder 47:30

Totally agree. I think uh we I think AI is getting better at answering the questions, it's getting better with the answers. It still remains for us to ask the right questions, though. We it's not capable of doing that whole thing. It's just but think about it.

Jason 47:46

There's billions of people around the planet every nanosecond of every day that are asking questions. So we're feeding that machine very, very rapidly.

Roopinder 47:55

Yeah, we are still a place for us. I hope that uh hope our children will still have a place in this world and not be confined to a world where they world of Netflix and consumption, other consumption. We do still need to have smart people in the loop at some point.

Jason 48:13

Well, you you still again not to get on soapbox, but you still need you know human-to-human interaction, right? Uh I genuinely believe we're we're we're designed for that and uh you know, socializing with one another and breaking bread with one another, and you know, all those sorts of things. So technology is is a wonderful thing. We just got to keep that uh at the top of mind of what we do for our own happiness, you know.

Silicon To Systems And Human Judgment

Roopinder 48:36

Great, agreed. We have to stay on top of the technology, and uh yeah, and I think uh I think we're I think we're there. I think we have the uh what you guys are doing. I've always admired simulation as being the top of the uh cream of the crop or the top, the top of engineering. I used to do a little bit myself a long time ago, though never reached levels that you guys have obviously have, but uh I think that's the high point of engineering. I don't think it's about ready to go completely under the hood. Uh I think it's good that there's some democrat democratization, but uh the real sophisticated stuff, as I've said, it's not there yet. It can't do that yet. We need uh we need it, still need experts. So glad you guys are there. Keep up the good work. And uh great having you on the show. Thank you, gentlemen.

Jason 49:27

We appreciate it. Thanks for having us. Take care.

Roopinder 49:30

All right, take care. Have a good rest of the day. Bye-bye. Bye-bye. Thank you for listening to FoDES the Future of Design and Engineering Software show, brought to you by ENGTechnica. I hope you have learned of a new application or technology that will help you with your job. If you have an application you think would be of interest to other engineers, please let me know by emailing me at roopinder at engtechnica.com or message me on LinkedIn.