0:00 Perfect. All right. Thank you guys for having me and it’s really awesome to see everybody back in New York coming here. I feel like each of these years the events getting bigger, more popular, more awesome. And so my name is Brad Rothenberg. I’m the CEO, founder of nTop. And I’m going to talk today about advanced aircraft design and running rapid trade analysis in it. So, first, you know, if I just take a step back and what and and look at what’s happening in the industry, like the new programs that the customers are putting out in the market.

0:31 This happens to be an announcement from the Navy recently. They’re asking for models to be delivered faster than ever. This specific program is saying you need to go from zero to coder, which is conceptual design review in 30 days. That means the submission after 30 days is a three do or a six dooff model that is of the air vehicle that’s going to go into a mission analysis tool and they’re going to say does this meet the mission or does this not meet the mission and so what’s happening is these faster design cycles essentially just amplifying what’s called the lock-in risk and you know this this is in aircraft design aircraft development but it’s also across other industries, which is early on in a program.

1:21 When a program is being formed, all of these decisions get made that have massive cost, massive budget, massive time implications. Because if you submit that model after 30 days and say the mission is going to do one thing and then later on in the program you realize, -oh, it doesn’t do that thing. You’re going to eat that in cost and time and budget. And so, you know, what do we do as an industry?

1:46 You know, we have a decision that we need to make if we’re submitting to this program and this is our customers do this. We don’t submit for programs at nTop, but our customers have a choice to make. Do you go low fidelity and fast or highfidelity and slow? And there was a really awesome chief engineer I was talking to who said, you know, lowfidelity tools give me lowfidelity insights.

2:09 And that’s the the the core issue. Like if we want a lot of confidence, if we want to reduce the risk in a program, we need to be able to both have highfidelity tools but also go fast. And so the bottleneck to that is really that building parametric models, which are the kind of core way that we understand systems, building those parametric models in the high fidelity CAD tools that we use every day, it’s really hard.

2:37 The the feature-based CAD has inconsistent behavior. The models break. And so these parametric models, it might take six months to build a model. You don’t have that if you have to go from zero to coder in 30 days. And so with NTOP, you know, all of the core modeling capabilities, the implicit modeling that we’ve built and developed over the last 10 years has proved you the the the most robust, fast, and also high fidelity model that allows you to iterate really quickly and actually allows our customers to kind of hone in with a high confidence that the results that they’re going to deliver are okay.

3:17 And so there’s really three reasons. One, it’s the implicit model that’s at the core of NTOP. Two, it’s that the notebooks in NTOP, they’re not just, you know, 3D modeling operations and features, but the notebooks capture the model as code. So, it’s replayable, it’s reusable, it’s sharable, but also all the engineering knowledge that goes into building those different components. And three, the implicit model makes combined with the knowledge allows for very fast robust parametric models to be built.

3:53 And so this parametric model is really the core of moving quickly having a high confidence in results. And what we’ve been doing is we’ve been proving it. We’ve been sitting shoulder-to-shoulder with our customers building out advanced parametric models that help them win programs. And what that looks like is starting from some requirements, parameterizing up a model, and then exploring the design space and closing a design that then you feel confident about and can move forward with.

4:20 And so the real key unlock for us internally, so the the key unlock that we’ve made in the last year is actually just being able to loft precise implicit surfaces that form the core of a lot of of many of these different models. And so you still are creating parametric geometry, but it’s a very very high fidelity. For example, you know, you can put together a model like this liquid rocket that’s kind of a early preliminary design style model, but you’ll notice as you move the different features, and this is kind of in real time in the software, all the blends, all the fillets, all the connections update smoothly, reconstruct.

5:04 And you know, if you as a human could click on a parameter in nTop and move it and the model updates, that means these models are perfect for integrating with all the different tools that we’re seeing for building surrogates because you can then spit out at each variation of the model a version of that to start to train. You can create a synthetic data set really easily.

5:25 And so, you know, starting with an initial idea, it’s very very quick to put that model together into a parametric geometry model that you can gain insights on, do calculations on, connect up to the physics to calculate your seal over CD or your lift over drag of your airframe. And those airframes can become more advanced. And so once you have that model, you know, you could tie it.

5:50 You know, in NTOP, we have our own NTOP fluids solver, but we also work very closely with the team at Luminary for kicking out models into Luminary. We also connect with Star CCM with Fluent. And so that that parametric geometry model becomes the kind of core, fast, quick, iterative model, but it’s still a very high fidelity, right? Like I could write a spreadsheet and model an airplane in a spreadsheet.

6:16 But is it going to close out later on? Is it going to package? I don’t know. But you can build that that into a model like this. And then run your design of experiments and start to make a trade and say, okay, you know, overnight let’s run hundreds of designs and then make a trade and pick, okay, this these three meet the mission, so let’s continue detailing those.

6:38 And so I want to kind of shift into a program that our interns did for the summer because I think this is a really good example of a fast design process. And so we put together this internship program. I was getting like hammered last year from this guy Sam from USC and he was like, “Can you guys hire me as an intern? Can you hire me as an intern?” And I was not thinking about hiring interns this summer at all.

7:00 And I was like, “Okay, well, if we hire an intern, you know, let’s make a program out of it. And what’s the internship that I would actually want to do? I’d want to get something flying by the end of the summer.” And so what I said to Sam was, “You can come on board, but you have 12 weeks, and in 12 weeks, you’re going to get a group one UAS to fly up in the air.” And so he was like, “I’m on board.” And I was like, “All right, awesome.

7:28 You’re starting May May I I forget, maybe the end of May.” And I hired one other intern as well from Penn State Katya and the two of them came on board with the goal of flying an airframe in 12 weeks. Four weeks into the internship we also I also called up our intern from the summer before who’s in high school, Zeni, and said, “Zenhi, why don’t you join as well?

7:51 Do you want to do the electronics and the wiring to get this?” So he he should talk to you guys next. He’s amazing. And so what that looked like is they started with the components that we had from these existing aircraft models and those were kind of up to preliminary design level. And here they’re like okay no we need to go into detailed design and so they took they built a parametric model really really fast in a couple of couple of days to build in all of the fastener all of the things that attach all the motors.

8:20 We picked some existing motors etc. Tied it up to the analysis to to validate some of the assumptions that were made to check the air foils. They connected X foil up to the tool as well and built a parametric model that you know this model or the airframe model itself is parametric all the way down to the design detail. So you can make parametric changes of the sweeps, the spans, you can look at the r how that impacts the range basically in real time in this model that they built all the way down to a detailed design level.

8:54 They made a couple ugly configurations also. Not not hard to do that as well. Still updates, you know, is that like the guppy maybe, right? There are ugly planes out there also. It’s one of my favorites. What they settled on was this configuration that was about 40 in wingspan about 30 in long. It was a penetrator designed for a group one UAS sized to have a flight time of a little over an hour and it was about 23 different components that can be 3D printed overnight and assembled in an assembly in nTop.

9:34 And it again, it still is equally as parametric at this detailed level. And so the program really is like design, sleep and print, wake up in the morning, put it together, which is quickish, and then fly. And so, you know, how could you get a new capability out into the market faster than ever? The solid model is the core to that. And so they put a number of these models together.

10:00 This was the first airframe that they built. Got out to the field. Floyd Bennett Field is an old Navy base kind of close to JFK and there’s an FAA there’s a there’s a there’s a the FAA allows you to fly whatever you want there. So you can go to the RC field as long as you’re part of the club and you can you can fly stuff. And so we became friends with the pilots out there.

10:25 This happens to be George Irving who’s a configurator at Loheed for like 33 years. He was advising the team on the on the design. And so they got out there, first flight, everybody’s like, “Yeah, the model’s awesome. Ready to hand launch it.” And they go and they crashed. And so, you know, it’s good learning. The spec sheet of the engines and the motors said that the power was going to be X.

10:54 And when we measured the power, it was actually like 30% of X. And so instead of a 3 to1 thrust thrust to weight ratio, it was about 0.9. And so hand launching was not going to work for us. So, you know, over a overnight in about an hour, Zeni, our our high school from Avenues High School intern put together this landing gear assembly. And, you know, at the field, Bruce, who’s an old-timer, he’s he was our pilot.

11:23 He’s 80 years old. Amazing guy. He’s like, “You know, you guys, you better put steering on that nose gear.” And George Irving, also an old-timer from Lockheed, ex Lockheed, who’s working with us, is like, “Don’t put the steering on it. It’ll work.” Like, “This thing’s twin engine. You can deal with variable thrust. It’ll be fine.” So, we listened to George Irving and we put the they the the team put together this airframe and got out to the field again.

11:55 Let’s see. Is this going to go to the next slide? And we got out to the field and everyone’s ready to go. I should note there was two other hand launches that they also crashed in this process. This is a learning experience. And so we got out to the field, everyone’s psyched. And Bruce is like, should have put should have put steering on it, but it’s okay.

12:16 We’re going to try it out. And press go. And we did ground loops the whole time again. Learning there is listen to the people in the field that are operating, not us engineers building models. That’s where the learning happens. And so we were like, okay, there the internship is ending in a week and a half. We got to get this thing in the air. And so one of our team members, Yuki, was like, okay, I have a Mini.

12:42 Let’s put the Mini on the runway. Let’s get the car up to speed and at 40 miles an hour, we’re going to launch it out of the top of the out of the top here. And so I was actually That’s me. And Zeni was recording and also I don’t know if you can see I should have blown it up but it’s also crashed and so but this was the longest flight yet.

13:02 This one flew for like maybe 30 seconds 30 time you know 300 times longer than our last two flights. The team was psyched. Like Zeni was so excited when the thing was leveled off and flying perfectly that he dropped the camera and he was running over. At which point the the vortex coming off of the car, the wake from the front of the car caused the plane to just completely turn over and go straight into the ground.

13:31 So it crashed. We didn’t have the car in the simulation set up in nTop. So maybe we should do that next time. And so finally team went back to the office with like a week left in their internship and listened to Bruce and put some steering on the nose gear and got out to the field with they built two airframes this time and they’re like we’re going to build two airframes and we’re going to get these things in the air.

They just had their design review that day and I think the next day was their last day at NTOP. So this was it. They had two airframes. And so they got there, landing gear with steering, takes off, and Bruce, amazing save on takeoff. The trims were not right exactly, but this thing got in the air and had an amazing flight. Perfect first flight. Bruce nailed two landings.

14:27 This thing was just awesome. It was fast. Flew really perfectly on the first flight. I don’t know if I have video of the landing but it lands really fast also and this was a you know this is a learning experience and so for the team the learning is really like you know the tool set is one enabler like the tool set should allow you to move as fast as possible given whatever scenario given whatever situation is out there the tool should be right there with you and help you iterate and so it’s really from nTop like I want to build tool sets that help you fail fast, help you learn faster and the the missing link that we’ve added is really the parametric geometry model.

15:11 That’s where NTOP fits within the market. And so with that, there’s kind of a nice happy ending as well. I guess it kind of is a happy ending too from their internship, but they originally designed the aircraft to be hand launched. And so one of my buddies who’s in Dayton, Ohio a few weeks ago was like, “Hey, I saw that nTop model. Can you send me that?” And so I I got him access to NTOP edu version, sent him the NTOP model and said, “Go for it.” The engines we picked were undersized.

15:44 If you can fit a better bigger engine on, do it. And a few months later, he went out into a like a few weeks later, he went out into cornfield and had a very successful hand launch with the same vehicle with the correct size motors. And so, this was like 3 days ago. I got that. He’s in Dayton. He’ll be making changes to that. We’re going to do another internship program this year as well.

16:05 So, I’ll put together a more formal IP team. Maybe we’ll bring the arena guys in, have some of the wiring more automated. But with that, thank you guys for for having me today and we’ll talk soon. To see the full recording of this and previous presentations as well as information about future CDF events, visit CDFAM.com.