Patrick Hanley, Aerodyamics Expert

AI can spit out thousands of “optimized” airfoils in seconds, but what happens when the objective is wrong, the mission changes, or the aircraft has to be certified and insured? We talk with Patrick Hanley, founder of Hanley Aerospace and author of Aerodynamics in Plain English, about why real aerodynamics is still built on fundamentals, not AI, not vibe coding. 

Patrick shares his story from St. Kitts and Nevis, tiny islands in the Caribbean, to MIT, and how a deep love of math turned into a lifelong habit of writing software to solve aerodynamic problems  —  in his spare time. We dig into how his CFD products evolved as the market shifted from paid tools to free alternatives, and why he kept climbing toward higher-fidelity capability. He explains the practical engineering choices behind accuracy, including why grid strategy matters so much, what validation really means, and how a focused tool like Stallion 3D can carve out a global niche even while giant multiphysics platforms dominate mindshare.

We also go beyond CFD theory into the real aerospace work: Terrafugia and the flying car program, the mission logic behind eVTOL air taxis, and the overlooked constraint that shapes urban flight more than range does: noise. 

Finally, we take a hard look at AI in engineering design, from automated shape generation to liability, and why prototypes and testing still catch the failures no simulator can predict.

Transcript

Welcome And Meet Patrick

Roopinder 0:18

Hello everyone, and welcome to the Masters of Technology Happy Hour, where once a week I have a drink with someone I meet in the course of business, but someone I'd like to get to know better as a person. My guest today is Patrick Hanley. He has a company called Hanley Aerospace. I'm reading his textbook, which is called Aerodynamics in Plain English. Oh hello, nice to meet you. Nice to meet you. So I was asking, okay to call you Patrick, or do you like go away? That's fine. Yeah. Okay. Okay. So I was listening to your podcast and I'm looking at your your ebook as well. Oh, okay. That was written a long time ago. You know, it still works, though. I I gotta say. It should still works. Uh principles are still the same.

Patrick Hanley 1:14

So it's very basic. And it was it was just the um, you know, a younger person writing what they knew.

Childhood Dreams And Big Engineering Rooms

Roopinder 1:24

Ah, yes, yeah. So somebody, let me just tell you or um my a little bit about my background. I'm a mechanical engineer, but I always wanted to be an uh aeronautical engineer. I just thought mechanical would cover all the bases. My my love has also always been in the air. I was designing planes as a little kid, and uh I thought, you know what cured me of that concept, of that dream? It's uh when I saw a picture of Boeing, Boeing, I think it was in design facilities, and it was a huge room of people, uh, might have been a hundred engineers, and white shirts and black ties. It was black and white photos, so maybe they were different color ties, but all on big drafting boards.

Patrick Hanley 2:15

Yeah, I I had that experience. You did? I yeah, I did a summer internship at no other place than Fear Child Republic back in the day, and um in 1982. So I spent the whole summer working on the A10 of all aircraft. I even sat in the cockpit of one of them. I I don't know which one is my A-10, but it was very interesting. But yeah, um, after it was a summer of just the same you described, um, you know, white drafting boards, people standing at boards, and uh they had an IBM um 360 computer and I fed it, I fed it on punch cards. And the punch cards was working on the air conditioning system for the uh aircraft.

Roopinder 3:04

Right. And and uh it occurred to me that I may never get out of that room and I may never be noticed. I might have been noticed because I would have been the only guy that wasn't white, I think, but maybe for no other reason, right? I thought I could I could spend my entire lifetime making uh landing gear parts or something. Yeah, yeah.

Patrick Hanley 3:24

I think I had to say more. Yeah, yeah, yeah. I think I wanted to do something myself that was more original. Yeah, but that that experience was extremely good, nevertheless.

Roopinder 3:36

Yeah, but okay. I'm I'm really I need to I need to find out more about you, so then I'll finish talking about myself. But I feel like we've known each other for a long time, but we've never we've never met until now, right? Face to face. Yeah, definitely. It's been decades. But um, I don't think we ever had a chance to meet. Yeah, so let me apologize for that, first of all, because I spent the last 30 years or so in publishing. And the publishing world, you know, is dominated by big companies, big vendors, software because they advertise, right? So consequently, we're not really uh, I wouldn't say permitted is the right word, but we didn't pay much attention to companies that had or you know that didn't advertise or were too small to matter, but that wasn't always the case. It was because they didn't advertise. Let me just be frank about it, right? So we were we, you know, I used ANSYS, but and we covered ANSYS because, well, they did, you know, they were a customer, but I was new of your products, I was new about what you had done, but didn't have time left over to I understand. Yeah. What but what made you uh you were a developer first of all? Oh actually I was a professor.

Patrick Hanley 4:57

Professor I started my career at UConn.

unknown 5:00

Okay.

Patrick Hanley 5:00

Um teaching um aerospace engineering, getting graduate students out of the door and undergrads as well. So that's where I started.

Roopinder 5:14

Okay. Yeah. They're always in the news for their basketball program, right? Oh yeah. Oh yeah.

Patrick Hanley 5:19

I watched March Madness with them in this, yeah.

Roopinder 5:22

Oh yeah, yeah. Oh, so okay, I told you I was listening. I tell you I was listening to the podcast you were in by uh what's that company? CFD is a CFD company, CFD online or something. Anyhow, I was listening to your podcast. Oh yeah, yeah, CFD engine, correct. I think that's the first time I heard you speak.

From St Kitts To Physics

Roopinder 5:42

But what are you? So I can't, I can't detect your your I can't place your accent. Where are you from?

Patrick Hanley 5:49

Um I'm from the Caribbean. I'm from an island called St. Kitts and Nevis. Smallest, one of the smallest of Commonwealth countries, um British Commonwealth countries. So being a Commonwealth country is um back in that time, we had to speak the Queen's English, and we had to do all levels that were graded in England, Cambridge. So that was my earlier education, studying for O levels, um, seven, eight subjects, um, nothing really engineering French and all sorts of different things like that, and then they send them off to England, and then the results came back and they published it in the newspaper. So embarrassing both the students and their parents. So he had to do well back then.

Roopinder 6:39

You were a good student.

Patrick Hanley 6:41

It reminds me of Well, there were some of the subjects I was not stellar. I only be um things became easy when I came to America. And what I found particularly easy was um math and physics.

Roopinder 6:56

Okay.

Patrick Hanley 6:56

And where were those subjects all my life?

Roopinder 7:00

Yeah, yeah. So at what point did you discover physics? That was in in college?

Patrick Hanley 7:06

Actually, I I when I came over here with my parents, they brought us all over. I was 16, and even though I had done the old levels, the results had not returned as yet. So I had one year of high school. And so I went to that one year high school in New York City, and New York City gave uh what we call a citywide physics exam. And that's when I discovered physics. It was to me, it was very easy. I got 98 on the citywide, and it it was it was pretty cool. I I didn't realize how difficult an exam exam was, but it was all the stuff that stuff I liked.

Roopinder 7:42

So physics, physics one, which OR 101, whatever it was, that was uh in my school. I went to Drexel University, and that was a that was kind of a filter of engineers. The median grade score was the 35 on that test, and and I think half the class didn't show up the next day. They went they went to business school.

Patrick Hanley 8:03

Yeah, I think in that city by the next highest student was 78, and I think everybody else failed. I and I didn't know why, but you know, for me it was good.

Roopinder 8:15

For yeah, it was good, yeah. Most people probably wouldn't have made it anyway, right? I I I'm joking, but I you know I'm I'm of two minds about that because I think there's a lot of that a lot of people that would have been good engineers, but they didn't make it in through mathematics and and uh science, right? It just made some people out. Okay, but that's a that's another subject. So you find yourself that you were really good at, you find something you were really good at. Now, yeah, at what point did you think you were gonna go into being an aerodynamicist?

Patrick Hanley 8:47

Because I is that you might was two years or three years old, I probably think. Yeah, I chased around all the butterflies and the birds um in there on the island, and I tried to figure out how they worked. Um, we lived on the approach an airplane, and I knew all the airplanes. I set their short the radio to their broadcast, then I built a lot of things that did not work. Oh, one of the best ways to learn something is not to have it to work because if you if it works, you won't learn what's going on.

Roopinder 9:20

I I did the same thing. I broke, I broke actually a lot of things, and my mother would uh would hate it. She would hate it. There's so many things in the kitchen that are dense or she had to fix because I was busy taking them apart. But I I drifted towards the uh structural side of engineering. Oh, yeah. Because I could see and feel it. Uh I hold aerodynamicis in high regard because you guys can work with things you can't see, right?

Patrick Hanley 9:48

That's that's the fun part. Yeah, I I mean to ask many things structures. You could give me about 10 eight 10 questions and I will fail. Structures is not one of my best subjects at all.

Roopinder 10:02

Really? So sort of complimentary. I see, I see.

Patrick Hanley 10:08

Yeah, the the math and structures was easy, but you know, I just you know couldn't put my uh my mind around it at all.

Roopinder 10:17

Uh-huh. So you you went on to you got your oh, you went, you got summa cum laude at NYU. Yeah. Yeah. Okay.

unknown 10:28

Okay.

Patrick Hanley 10:28

Yeah, I did a I did a lot, I did a lot of math courses there. I did some graduate courses even as an undergrad. And as you can tell, I'm addicted to math. So and then I I in my third year um in at Polytech,

MIT, Math Addiction, Early Coding

Patrick Hanley 10:48

I found out about MIT, and um I have not heard about it up till then. So um I applied and I mean it was essentially a stroke of luck, uh, in the sense that I was just off, I mean, three years out of the island when I applied, three or four years. And at Polytech, I wanted to get a job, and also worked at a supermarket as well. And um I I well, there was a professor at Polytech. I wanted to do some some you know work study work with him, and I came, I I bought him week after week after week. Finally, I worked with him, and little did I know, just by working with him and another professor, they were working, also working in an acoustics area with a faculty at MIT. So I was really, and my grades at Polytech were excellent. So I was able to, you know, get a straight recommendation from those faculty members. Just like that. I mean, I could have gone to another school and it wouldn't have happened.

Roopinder 11:58

Yeah, that was great. So scholarship? MIT?

Patrick Hanley 12:03

Yeah, definitely. Um, I was a research assistant from day one.

Roopinder 12:07

Okay.

Patrick Hanley 12:08

And they paid. So they paid me throughout my whole month.

Roopinder 12:11

Yeah.

Patrick Hanley 12:13

No tuition plus what was like a you know, I think they paid like for in 1983, they were paying like $900 a month for stipend. I mean, it was for graduate student salary, so it was pretty interesting. That was a lot of money back then.

Roopinder 12:32

Yeah. Yeah. But then you've been several years in the country, so you were kind of uh kind of used to the way things worked here. Uh I'm I'm saying that because my father also came. He was he came from India and a small village, and he uh he happened to get into Harvard, very close by. Uh and he got a PhD in it from Harvard. So excellent, excellent.

Patrick Hanley 12:58

I think as immigrants, we're not scared. We don't know what, we don't know how tough it is to get into these schools, so we apply anyway. And then you know, we I think immigrants have a really good work ethic. So you know, you could all learn about three or four jobs in addition to the coursework, and it still seems like wow, America is easy, you know.

Roopinder 13:24

Yeah, yeah. I remember my father's toughest, I think he got good, good, very good grades uh in all the school, just like you did. And uh and uh but I think the hardest thing for him was just getting used to the culture because he uh he still remembers being served uh pancakes at uh in uh Boston at a diner. And he took me there after what after we all moved to the country back into the country. He left to went back to India. We lived in India, uh, and then we came back to America um many years later, and uh he took me to that place where they served him pancakes, and he said, This is where I sat and I had my first meal in Boston. Wow. He didn't know what to uh yeah, he was still there, and he said, I didn't know what to do with these pancakes. It's very it wasn't very clear to him to put syrup on them or anything, right? So I think he folded them up like a chapati. That's what we had. We had uh breads, flat breads we call chapatis.

Patrick Hanley 14:26

And uh called ro is it like roti? Roti. Yeah, we have roti in the islands, yeah.

Roopinder 14:34

Oh, because your islands probably had a few Indians there, correct?

Patrick Hanley 14:38

Oh yeah, oh the island was just a melting pot. I mean, you know, it depends all the islands had people who are from India, China, Africa, all these British type colonies, people were in the Caribbean. So definitely, yeah, we we we have a lot of Indian food that that um that you know we have a lot of curry and stuff like that, and and we have different you know, African foods, and I'm beginning to learn some words that we took for granted from the on the islands, have African roots and Asi roots and things like that. So it's fun learning that stuff. But we just thought they were island road words, yeah.

Roopinder 15:23

All different cultures, yeah, right. It's uh it reminded me uh how how uh diverse that population is on the islands, yeah, yeah.

Patrick Hanley 15:34

And then of course there's cricket, you know. The West India team was really good back in the late 70s, and Clive Lloyd and all those guys, Chandler Shaker and all those guys, we know the all the Indians and the Lily and the Australian cricketers and all that stuff, just by listening to the radio on a small island.

Roopinder 15:55

Yeah, yeah. Did you develop a love for the game of cricket? I never I never did. I whenever I go back to it.

Patrick Hanley 16:02

Well see, I've been a lifelong nerd. Sports and nerdism doesn't go very well. I love cricket, but I just couldn't play. Oh, okay. There were better players than me.

Roopinder 16:14

My brother was so good. Believe it or not, in the second person, second PhD from MIT I talked to today. I had another gentleman from here, he founded another company. Um but he's Indian and he went to MIT. For a while there, we didn't have anybody from MIT. And then today we got two two uh doctors from MIT. So good, good for you, good for you, and uh good for me too. So okay, now what now to uh what led you to develop software? Because

Building CFD Products Through Disruption

Roopinder 16:53

you're familiar you're familiar with all the commercial codes, right? Like ANSIS.

Patrick Hanley 16:57

Yeah, but then I started back in 19, gosh, 1998, when I started in earnest, but I started before that in 1993. Okay, and I've always worked with software. For example, when I was working on with the faculty members uh at uh Polytech, um, I had an HP 41 CV and the code that I was stuffing into the um IBM 360, I developed an RPM version of that code just to run on my HP. Also I wrote a jack blackjack for my game for my HP, and I even wrote a H heat equations code for that and run it on that. Just for fun, just for fun, just for fun and to work with. So I I always uh once I found out about programming, I eventually said, Oh my goodness, I want to start my own um software business. But um that didn't happen until um uh 1993-94 when I decided to leave Yukon and um do my own business, which was uh a terrible decision. As far as our standard of living, of course, it was. But you know, starting from scratch and uh you know trying to develop a software package that did not exist, um and having to rely on that software package that didn't live exist, so I was I was food. So even though the competition was there, I I was able to on the price, undercut them, and so my and then I developed the software for Windows 3.1. So Windows 3.1 was user-friendly, and at that time, most aerodynamic software would require you know some bigger hardware to run. So I I w uh my first software was Visual Foil, just like Visual Basic and the Windows whole visual stuff, that's why I called it that. So I would identify with Windows 3.1. And you know, I I had a lot of customers then, but then at some point uh Mark Treller made XFOI free. So for from it was a multi-thousand dollar software, so suddenly a free software. There went my business, so I had to develop something more powerful. So I called that multi-surface aerodynamics, which was a panel method code, vortex lattice method, and then OpenVSP came out and a whole bunch of free uh vortex lattice codes. So I had to develop something a little bit more powerful still, and that's how I came up with Stalin 3D. And um, at the time the name came from the fact that we lived in Nakala and we've been living in Nakala for 30 something years now, and my daughter was doing horseback riding.

Roopinder 20:01

Oh, I was wondering why, if you're into airplanes, would you name it after a land adult?

Patrick Hanley 20:07

Yeah, I recorded Stalin 3D, and my daughter, who was about eight or nine, she drew the logo because she's really good at drawing, so they're very good at my kids. I'm very good artists, they have envelopable artist talents. I could do stick drawing, they can do masterpiece.

Roopinder 20:30

Really? So they are your children. Tell me how many children? Six, six, six kids. Six children. Yeah, yeah. All right, I'm spanning what ages.

Patrick Hanley 20:41

Spanning from 21 to 31, going on the tools in Spain. Yeah. Okay. Did you make any engineers out of them? Oh, my son, my first son is an engineer. He he okay. Yeah, he he he did computer science.

Roopinder 20:56

But a few of them are are more of a creative type, you would say. Yeah, exactly. Yeah, they're only extremely creative. Okay, I see. Yeah, but yeah. Yeah, I've have you found this to be true that engineering is not hereditary? Are you you got lucky with your son? Yeah, yeah.

Patrick Hanley 21:15

I I when my uh when I I I was doing some outreach work at MIT when I was in Almight, Connecticut, and I took my son who was like two or three years old, and we were speaking to the dean. And the dean said, Your son is going to be an artist because it's not hereditary. That was the first time I've heard, but you know, he became an engineer and the others became artists. Not because I was more influential than him, but I took him to most of my um meetings and stuff that I did, promotions and and um AIAA um meetings. And like it was 1617 and he was at my booth. He was the attendant at my boot for free. So he got it in Providence. He's pretty good at engineering, so and public speaking as well.

Roopinder 22:11

Oh wow. Well, that's great. Uh what kind of what type of engineer did he become?

Patrick Hanley 22:16

Computer science, which you know, as a new graduate is one of with AI. Apparently it's not good, but he has embraced AI. And um he has been waiting for AI for the last 20 years.

Roopinder 22:33

I see. So he's ready to jump on it and take advantage of it. Right. He's probably using Claude code now.

Patrick Hanley 22:40

He uses everything and he's telling me how I can get into it. And um I'm like, you know, I just use it to buy code and I use it as a means to an end. I don't use it as, you know, I'm gonna learn anything about AI. No, I learn AI to do my work and to accelerate my coding, to accelerate my geometry creations and to accelerate my advertising.

Roopinder 23:06

Uh advertising. Oh, you do advertising. Oh yeah.

Patrick Hanley 23:11

I usually do, I just write posts and I more or less boost those posts on mostly on Meta and sometimes on LinkedIn. And I find that is quite effective.

Roopinder 23:23

Okay. So you kept up your software. I noticed that you updated, it's it's you you introduced what did you introduce Stallion a while ago, a few years ago?

Patrick Hanley 23:33

Yeah, I um in 2017, Stallion 3D was an Euler code.

Roopinder 23:39

Okay.

Patrick Hanley 23:40

Um, and then I had to figure out how to put Navi Stokes equations in an Euler code. And I had issues with um a Cartesian grid and um the wall. And anytime I tried to run a Cartesian grid on uh without cut cells, it would get terrible, terrible results. And at the and at the time my kids were in um elementary school, and I would have to, I would go to pick them up in the afternoon, and all I was looking at on the ground was tiles. So I was looking at tiles, and it came to me how in the world

Accuracy, Structured Grids, Finding Customers

Patrick Hanley 24:22

to make a Cartesian grid without cut cells that work with high accuracy. I'm not gonna tell you, but it worked. Oh, you've you figured it out. Definitely, I had to.

Roopinder 24:37

Okay. Now you so how did you but how did you sell your code when you're comp now you're competing with big guys, the big guys, right? There's a lot of CFD companies now. Uh there's probably two big ones that are into uh multi-physics, so they can do structural and CFD, right?

Patrick Hanley 24:59

It's surprising how the customers find me. They find me from all over the world. Uh, and you know, Ocala is a very remote town in Florida by um the national forest that is good for bombing runs by the Air Force. Yet still my customers found me here from out of nowhere. And the reason why there is well, I pray a lot, and the reason why that happens otherwise is that uh the Stallion 3D carves out a very narrow niche. So if you have a uh if you're just starting out, yeah, uh a lot of the big companies, when they were just starting out, they bought my code, you would be surprised. And after they graduate from my code, that's when they buy answers and starts to see a plus. But before they get to that point, they had to buy my code. And and uh you'll be but yeah, they found me, and uh basically I I had a I started out with I had a website from day one, so that's how I was able to be found, and I do a lot of search engine optimization, and I think I was one of the first companies to do to do um e-commerce for CFC.

Roopinder 26:19

Yeah, okay. So you're kind of like their what did they call it, a crime, a gateway drug?

Patrick Hanley 26:25

You're like the gateway drug, and then they're exactly exactly, exactly, exactly, but of course the the code has to be extremely accurate. Yeah, I mean when you do something on style like the the M6 wing at um 0.8, um YO would be equal to 0.8. That's very difficult to get. And when you look at the first NASA validations of the Onera M6 wing, there's no shocks, there's no double shocks at 0.8. But with Stallion 3D, when you look at a small, you know, 200, sorry, two million cell that runs on the PC, yeah, I do get that distinguishing shock. That's the accuracy. And that's why I had to find hot cell would have killed me. I would not have been able to do that. I had to find, you know, that that accuracy to to do that. And uh as you know, when you look at cubes and write rectangular prisms, yeah, they're 10 I've rather 10 to 30 times more accurate, say than uh unstructured grids.

Roopinder 27:38

So that's your main selling point, is the accuracy.

Patrick Hanley 27:42

Yeah, and um 3D isn't it's um it's structured on struct um but it still has uh it still has it it's uh what should I say it's effectively is structured, but it has a connected connectivity matrix, nevertheless. So it can make the it can make the grid without any fear of being or uh unstructured, following all the structured gridding, like for example, car 3D, not car 3D, but uh that uh NASA structured grid code. Yeah, okay. Why is somehow I'm forgetting all the structured grid codes.

Roopinder 28:26

Oh, you probably have you haven't had to remember them for years, probably. Uh okay, so okay, so you're selling you're selling your own software, and and you're it sounds like you know all the fundamental theory behind it, and you can go for it.

Patrick Hanley 28:42

Yeah, um, I looked at the basically all I was doing was solving math equations. I did everything mathematically, yeah. So and then programmed it afterwards.

Roopinder 28:52

How do you feel about there's a lot of people

AI Design Tools And Liability

Roopinder 28:55

now that are pretending to know what they're talking about, and and they're coming at it from a whole different point of view. Like they don't necessarily have to know the fundamentals, they don't have to know what the what's happening in the black box, they can just tell from the input and the output what happens in the black box. And I'm not saying this company is like this, but this is what reminds me of a recent conference I was at. It was by nTop. Yeah. You know nTop? Yeah, I think so. Yeah, I know Brad Rothenberg.

Patrick Hanley 29:29

He's a uh is it the FD or more structure or more geometry creation?

Roopinder 29:35

Both. It does both. Uh it does both, but it the his the advantage of nT top is it's called nTop comes from nTopology, and it can vary the topology, optimize the topology, uh, so that it can generate not a single solution, but it can look at hundreds, maybe literally thousands of solutions and look at all different things. The reason I'm reminded of it is because in that conference he showed, oh, he does workshops too. In a workshop, he showed normal people, people who had no idea of aerodynamics, probably had no idea of a what an Euler equation is or Bernoulli or anything, right? Didn't know any of the fundamentals. And they were just able to use their software. And sure enough, it would generate airfoil shapes that were that had good things that so what do you think of that? It's like are you is are people like like you real experts that know the fundamentals going to be displaced by people who don't not necessarily know the the well that that would be good because um for example uh several years ago I was hired by a firm to design a drone, yeah, and um uh basically I have I back in that day, back in the day, then maybe about 20-15 years ago, I knew airfoils.

Patrick Hanley 30:57

Um I've had my airfoil code, yeah, and basically instead of training, the code trained me. So when I worked in um when I went and did real physical work instead of just playing with mathematical equations, I knew exactly what airfoil would be good because doing all the airfoil analysis, yeah, and all the six series, four five series, and UIUCE airfoil database, I was able to say this one is going to work well for this particular airplane.

Roopinder 31:35

You developed an intuition from from watching what your software did. You developed an intuition. So even though you couldn't see the fluid moving, you knew from enough analyses that this is how it was going to operate.

Patrick Hanley 31:47

You got to feel but also also you have to know about what the mission profile is for that aircraft as well. And for the you have to pick an airphone, not because it had good lift drag moments and all that stuff. So I don't know what so my biggest concern as it asked before is what parameters are you optimizing for? Are you optimizing for lift drag moments? But the plane is not always lift drag moments, uh, and moments come in because it has control surfaces, and not all airplanes work the same. All that some airplanes have a wing without any control surface, and all it has maybe was working with, all it has was a rudder and elevators, and you just have to rely on just those tail control surfaces to fly the whole aircraft. So, what do you optimize there to get a good working aircraft? So, yeah, it's good that you could use AI to come up with good lift dragon moments, but you have to tell the airplane, they have to tell AI the whole story, right? And the whole story has to come from you. Um it's like it's like going to an ice cream store as opposed to making it at home, and maybe you could only make vanilla. Once you get to the ice cream store and you see 20 flavors, you have to figure out which one is the best.

Roopinder 33:19

But you know what? If I if I was making ice cream at home and I went to an ice cream store that made all those flavors, I'd wonder if they knew what they were doing. What kind of ingredients are they using, right? I understand it could probably be cheaper than what I make, and because they're mass producing it, but I wouldn't trust them, right? Because none of these software companies, I'm not I'm not saying this about NTOP, but none of these companies make any claim like for accuracy.

Patrick Hanley 33:49

And sometimes yeah, my biggest worry about software like AI design and aircraft is liability. And I've that liability I've had been on my mind from day one.

Roopinder 34:03

Yeah.

Patrick Hanley 34:04

I do not design airfoils because of liability. I do not claim my software has to have the user put in their own airfoil. I'll analyze it using mathematics, and uh the disclaimer is this is what the math says, which is usually a good match for the physics. But I'm not going to say I have this airfoil that have very good at drag, very good at lift, the six series airfoils, for example, but have terrible stall characteristics. Airplanes designed with six series airfoils tend to hit the ground. Um, just because they're very good, efficient airfoils. Yeah, as soon as you exceed a certain angle of attack, you lose your lift. And so, yeah, there's a lot of liability. So a lot of all of these AI automatically designed airplane companies, uh, they probably have to carry a lot of um insurance for liable. Because um you're you're giving a customer physical, a physical property, uh uh something that is physical that either uh people are going to be on, or it could destroy property and things like that.

Roopinder 35:23

Yeah, I'm looking at your career over the years. You talked about how you developed software, kept developing software all this time, but you worked in some very interesting companies too. Uh I want to talk about each of the some of these

Terrafugia And Learning FAA Reality

Roopinder 35:46

companies. I'm looking at one, uh, I looked at several of them, and these seem to be like cutting-edge, almost I wouldn't call it, I don't know if dream companies is the right term, but if I was an engineer starting out from school, I'd think that is a cool job. All except this one, Terrafugia. Terrafugia? Fugia? Yeah, Terrafugia, yeah. Yeah, I I was thinking that that's gotta be your toughest job you ever had, because that thing should doesn't look like it should fly.

Patrick Hanley 36:15

No, it wasn't, it was one of the funnest. It was the most fun I've ever had.

Roopinder 36:19

Really?

Patrick Hanley 36:19

They all kind of yeah, Terrafugia. It started out with the flying EV tolls. Um, Carl Dietrich is one of the smartest persons that you will ever find. Okay. And um, he's an MIT graduate as well. He has his PhD from MIT, along with uh his wife is also from MIT. And they're the kindest people, and you know, but um working with them, he hired me um just speaking with him for the first time. And I was he I I applied for being an engineer, and he said, No, you're going to be a technical fella. That's how he hired me. He created a job in the interview for me. And um I came up with the TerraFugia, and it was a lot of fun. And there were a lot of challenges coming up with TF2, and it had funding from Gili. And then once that um project um was closed down, we went and we did the flying car, the original flying car, the transition, and working with that um flying car, I was able to be at the airport, I was able to be in the truck, chasing that airplane taking off, was able to see takeoff, and really write code instead of doing CFD, but to reduce data from an instrument that took eight points of data every second for two hours and put it in an Excel file. So I wrote code for that. Um, let learn what it was to get something certified by the FAA on the light sports aircraft. Right. So yeah, fun.

Roopinder 38:03

That's it, it was extremely interesting work. I guess I'm looking at it if I was first given that task of hey, you have to have a car, and by the way, it has to fly too. So that to me seems like it's at best, it's going to be a compromise, right? Because it's not gonna be a very good car, it's probably not gonna be the fastest airplane, uh, it's not gonna be good at either task, right? But it has to do both. Uh right. Yeah.

Patrick Hanley 38:29

But

eVTOL Missions And The Noise Problem

Patrick Hanley 38:30

speaking about EV tolls, uh, if you want to transition to that, the flying car was named the transition, nevertheless. And um, when you look at an EV toll that can only go 25 miles on a single chart, on a charge, and it could maybe they could go longer and maybe they could go faster. Right. But look at the mission. The mission is only to get you from the city to the airport, which is about 10-15 miles away. Yes, and you can do it in five, 10 minutes. And if you were gonna hop a taxi in New York City, yeah, you have to leave your work. I mean, maybe hours earlier because you're gonna run into New York City traffic, you'll run into Boston traffic going through the tunnel as well. So even though that seems like, oh, we can only travel 20 miles or 25, 40 miles, maybe even 100 miles, getting to the destination that you want to go to, and it's not a helicopter that you do not, you're not going to incur the cost of a helicopter, just leaving where your office building or close proximity to your office building and fly for five, 10 minutes to the airport, you'll save yourself an hour in any major city.

Roopinder 39:44

Yeah, yeah, because you're going as a crow flies, you're going a straight line. You're getting I call it the you know, they call it the in package delivery, they call it the last mile problem. This is like the last 50 mile problem because airports are always 50 miles away from downtown, and everybody, uh it's it's I can understand the need because everybody lands at an airport, everybody has to get downtown, and they kind of leave it up to you to get from here to there. Whereas this there's nothing more logical seeming than why not fly the last 50 miles at a straight line, right?

Patrick Hanley 40:19

Yeah, and then for EV calls, because they are leaving from the city, there are all the requirements. There are the requirements that come in, major one being noise. You do not want to hear these things buzzing you every five seconds.

Roopinder 40:35

How can you avoid that? Because even with a uh electric motor, you're gonna hear a considerable amount of worrying, right?

Patrick Hanley 40:41

Yeah, the the the most noise comes from the propeller. Yeah, so the propellers are are designed so that they will be more, they will be quieter than a helicopter, way, way more quieter than a helicopter.

Roopinder 40:55

How? Tell me how that works. Is that how's that one of the uh how is acoustics coming?

Patrick Hanley 41:02

Aero acoustics is one of the biggest requirements for these EV torks. And um just about how the blades are set up, um, just by the multi rotos, you know, the most efficient blade that you could have is one blade, just one blade. I didn't know that. But when you have two blades like a helicopter, then you have a lot of noise too, and the more and you start to lose efficiency by adding more and more blades.

Roopinder 41:32

Well, one blade, uh one blade is intrinsically unbalanced, correct? So but are you saying that's the most efficient aerodynamically? Is one blade? Because yes, because it less it's less it goes less into the uh foil there.

Patrick Hanley 41:47

It it's more it's it's more efficient, I've been told. Um, I haven't done a CFD to do it, so I don't have uh built in. Oh my goodness, this is extremely efficient compared compared to two blades, three blades, four bit. But apparently uh five blades might be very efficient and done by testing and twisting and all of that type of stuff. But also, you by having multiple rotors, you cut down on the blade road arm loadings a lot. And by having multiple rotors, you also have redundancies, so it's uh safer, right? So yeah, so these are the technologies that were that has to be perfected and get certified for the VTOLs, flying cars.

Roopinder 42:40

Are you saying that's the I almost thought it was the range that was that was the problem or the uh the uh the range is might be perceived as a passenger problem, yeah.

Patrick Hanley 42:52

But thinking about it and you know going back to getting to the airport in five minutes, you must say I'll get to the airport in five, ten minutes, so it's a big deal. Yeah, when you look at the alternative, right? Getting to the airport in two hours, otherwise, yeah, you mean you have to let you have to start two hours early, and then four hours early when you add what the TSA requires that you get there.

Roopinder 43:19

Yes, of course, right?

Patrick Hanley 43:21

So that five minutes take away the two hours traffic, and that two hours more work you have to do, you could be doing at the office, and because e veto might be ubiquitous, the the price will go down. So, you know, Uber was the one who started it, they were the one to Uber elevate. So you pay your $50, maybe even $100, which is not that, you know, with inflation, that's not exp seem expensive these days.

Roopinder 43:53

Yeah. Um does a helicopter ride, is we'd just say.

Patrick Hanley 43:57

Oh, a helicopter ride is more expensive than. hundreds of course the helicopter cannot uh land near your um office building or in close proximity to right yeah and um basically helicopters sometimes for EMT for emergencies you actually want the helicopter to be noisy so you can hear it coming to give you a good feeling but uh what buzzing around right um you're just taking people to the airport on business trips you you do not really want that uh for a while there I was hearing nothing about this uh air taxis I was like what happened to them how come we still don't have them we google sorry no archer and Joby these guys are very close to um certification within a year or so and they've done a lot a lot of tests on flights and things like that i work for supernal on day technology is very difficult and you know so these um people like i i i keep saying google because it's um google was one of the first companies to get into e V TOLt whisk if it it grew into visc and everybody everybody knows about Google's Waymo but they also had air taxi we yeah yeah they did they had a company called Kittyhawk and um kitty hawk bought my software there back in the day they were that much of a startup they bought my they bought one of my software but then they grew up into a big company and they don't remember me but um yeah okay so every everybody yeah everybody wanted to be the next air taxi company right oh yeah there were there were a whole bunch of them yeah um um going into coming out of night 2016 and going into the pandemic there was a okay it sounds like what do you you believe that all the technical difficulties have been worked out and now it's just a matter of certification oh oh yeah I think if you get the correct routes uh where you could travel mostly over water not over too many heads and you probably have reduced noise i mean uh Joby archer uh and the leading companies they have um uh work with the noise you're not hearing the whap wop of helicopters and things like that with these aircraft for example five blades going around you're gonna hear you're not gonna hear it five times you're gonna hear mostly broad broadband noise um instead of all these these spikes of shock waves from a helicopter are you worked for uh shield are they are all military they made all the make drones that can be used by the military and uh a lot of the UAVs can also you know be used for surveillance and things like that.

Shield AI, Swarms, And Meshing Craft

Patrick Hanley 47:06

I see they can be used for keeping the peace and and so forth. Okay. All right and are they somewhat competitive with uh oh there's a lot of talk about enduro now and yeah but the Shield AI um have an advantage a lot of all of these companies because they also make a one of um a software called hive mind which is their AI wing in California. And that software is being used by a lot of audit room companies so that allows the aircraft to virtually fly by itself even in deny areas that don't have to rely on a whole lot of um um you know I I would say radar and satellites and all that type of stuff using the their AI pilots so yeah that that that's one of their big advantages by by the name alone I would think that it uses uh like aircraft communication as well as yeah exactly like it can fly in a swarm or a formation easily so they they find their customers are other drone companies as well yeah or UAV or UAS as you want to be called were you working on the design of the particular specific drones? No just CFD okay um just pressing the button on the HPC and having it run. Now there you were using commercial commercial software right commercial CFD software oh yeah yeah you know star 50 plus and point wise on 3D um uh and also Lumin not luminaric but what's the other one uh boy now I remember Luminaric but it's competitive yeah interesting company flex compute actually okay I remember 360 you know 360 is really a joy to use you know it's yeah if you are interested you like that because of what is interface is good or yeah it's interface is excellent i mean you could just throw something on it and it would work yeah and um you don't really have to go to point wise or anything like that yeah point wise is very therapeutic to use and um it's therapeutic because the challenge is to come out with a grid yeah and you know you want your grid to be perfect but you still want to get it done in less than an hour. You have to get all your surface grid to be really really really nice. So if you do anything bad it's not going you're gonna have bad CFC results but sitting there and the challenge is let's get the best grid that we could within an hour so we don't waste time and then it also has a lot of automate the latest version of pointwise as fidelity point wise as it's called now has a lot of automation it it's a beauty it might be one of the only software that I might buy for myself for something you like it that much. That means a lot coming from you because wasn't pointwise bought by I think Altair Altair that's what I'm thinking of Altair bought it yeah Altair was very very uh active and buying companies and uh yeah I think that I met the gentleman from pointwise and uh then and then I I kept up with it a little while after uh after the acquisition but yeah that always impressed me as good as good uh good software it was a way that the mesh could adapt itself and reorient itself to for uh to make smaller cells where the was critical it in high flow areas it would it would keep getting the sm the cell or yeah I don't know if it has adaptive uh mesh refinement I do not I'm not sure if it does have adaptive maybe it does I'm not sure I've never used the adaptive mesh refinement for me you had to make the mesh a priori and then run it. Hence the therapeutic effect yeah definitely and it's only therapeutic if you can do it within an hour do a 100 million cell mesh within an hour and um it comes from um oh my goodness once it's done and it it runs without any errors or any type of compromise it's very good yeah and once and then drop it off you want your Fun 3D to run without any crashes. So everything went smooth and that makes you a point wise or Fun 3D expert instantly yeah uh you said you say your children are the artists but I'm thinking you're an artist with a mesh that's your creativity my codes are uh automatic gridding so um they had to be because um I do not know how to write the mesh generator I could write O grids uh most of uh my two 2D airfoil code is an O grid I could write that it's an algebraic not it's a finite difference grid elliptic grid generator the math again I mean the math is really cool there just do the math and you can get an elliptic O grid but when you can have a complex grid that will do you know any shape then you have to get point wise or do automatic grid um you know something that adapts to the surface. Anything else you'd like to point out as as your technologies you worked at or or uh companies that you worked at or mentioned not necessarily um CFD is is it's uh I would like to with AI I mean AI for my a pet peeve if you want

Why Prototypes Still Matter

Patrick Hanley 53:16

to call it that.

Roopinder 53:17

Oh okay yeah yeah that's discussed because we can't have a conversation complete these days without talking about AI.

Patrick Hanley 53:26

I would say AI generate a thousand airfoils instantly or two thousand or ten thousand yeah and I'm thinking you know the aircraft only needs two or three why not sit down for a day and figure out which two or three might be the best one based on your experience something that's not going to crash because working with small drones not necessarily expensive aircraft that you like to crash and you're not going to get optimized a drone unless you crash it a hundred times. Yeah not because it's a bad design but there's so much to work out and you want to get a good handle on what is happening. Is it the airport oh I don't want to open the book and review 1000 because it crashed this morning. So yeah that's a lot that that's what I have to add.

Roopinder 54:21

No I I think that's a good note to to that's a good thing to consider because all the time even just yesterday I heard a company make a claim that they want to make prototypes obsolete and it was a simulation company and you just reminded me of how it's important for the opposite you need to make a lot of prototypes because you know what they always say in engineering it's or one of the things they say in engineering you can't make anything foolproof because fools are so ingenious right yeah it's always what it's always what you didn't think of and you can't think of everything when you do simulation but yeah especially AI because nature has a way of making nature is the best simulation simulator yeah and um it it has a way of coming up with stuff that you can't duplicate in any software you can not not in the math I mean you remember that uh Galloping Gurdy bridge it's like yeah yeah yeah yeah like that nobody nobody ever considered resonance as a failure mode right until that bridge happened and now it's so many people I swear they almost like would pin our eyes open and make us watch that video as as uh engineering students because look it's what you don't consider that's going to wreck your bridge yeah and then after you have your perfect air pull attached to your airplane yeah then the airplane mission would change a hundred times okay let's add cameras here oh let's oh let's add this radar let's add more passengers let's take the same wings that we develop and add you know more feet so yeah and then everything changes in the smallest thing yeah perfectly multi billion dollars just to go to the AI and act a good multi billion dollar road backwards yeah because you have to go and act the AI for five cents and then you have to add a few billion to that five cents to get a new airplane. Well Patrick it's a great talking here I think it's been a joy to the conversation you're welcome welcome thank you bye bye bye and that concludes this episode of the Masters of Technology Happy Hour. I hope you enjoyed our chat and join me as we talk to some of the most interesting people in the design and engineering software community