Ignoring the Old Math, a San Francisco Startup Reinvents UAV Wings
What if you could eliminate turbulence around an aircraft wing and get twice as much lift for the same amount of engine power—or five or 10 times as much? What if the outer surfaces of a wing weren’t rigid, but were more like a living thing, moving relative to the wing’s interior parts?
To most aeronautical engineers, these questions would sound heretical, or ridiculous. But it turns out that if you go back to the fundamental mathematics of flight and start monkeying with assumptions engineers haven’t questioned in decades, some interesting possibilities arise.
At a garage in the Dogpatch neighborhood of San Francisco, a team of recent UC Berkeley grads is teasing out some of those possibilities. If their calculations hold up and their prototypes can be scaled up—and those are big ifs—it could cause airframe builders to rethink the whole concept of the wing.
The young founders of Vires Aero realize this is a big vision for a tiny startup, but they don’t have time for modesty. “Anytime I try to do any type of engineering, I want to be completely devoid of anything that has come before it,” says Vires co-founder and CEO Harshil Goel, a 2014 Berkeley graduate in math. “That comes from my mathematical ego. Mechanical engineers use a lot of calculus and partial differential equations that are over 100 years old. There is more sophisticated stuff out there that could create much more interesting developments.”
While the math is complicated, the basic idea Goel and his co-founders Jordan Greene and Zachary Hargreaves are testing is simple. Rigid wings with static surfaces usually create a wake of turbulence in the air around them, and that increases drag. By giving an aircraft wing a moving surface—in effect, wrapping a nylon conveyor belt around it—Goel and his team think they can delay onset of turbulence and, as a result, reduce drag and win back a measure of lift. And not a small measure: Vires Aero’s modified wings can carry up to 500 percent more weight, depending on factors like the angle of attack, Greene claims.
Star investor Tim Draper, who contributed $250,000 to Vires Aero’s $1 million seed funding round, has called the idea “the first true innovation in aviation since the jet engine.” That can probably be discounted as Silicon Valley bravado, and there’s actually a long history of experimentation with wing technologies that combat turbulence, from twisting “aeroelastic” wings to “circulation control” systems that emit compressed air to smooth out air flow. But it’s true that Goel and his co-founders—who see unmanned aerial vehicles (UAVs) as the first commercial market for the technology—are going back to first principles.
Goel, a math wunderkind, says he first got interested in aerodynamics in 2012, when he had a summer job at a wind tunnel facility at the Indian Institute of Technology Kanpur, in Uttar Pradesh. His supervisor knew that he was searching for tough math challenges, and told him to look at the phenomenon of boundary layer separation.
When air or water flows around a solid object such as a wing, viscous forces usually keep the molecules closest to the surface “attached”—moving in smooth parallel layers. This is called laminar flow. But sometimes friction at the wing surface causes the flow to become “detached,” forming tiny eddies and vortices. That’s boundary layer separation, and it gives rise to turbulent flow, increasing drag and decreasing lift.
In the traditional fluid-mechanics equations describing air around a wing, Goel explains, there’s an assumption called the “no-slip condition,” which dictates that air at the very surface of a wing has zero speed, due to friction. Some of that stalled air usually gets turned around and goes in the wrong direction, against the flow of the air above the wing, which is what starts the eddies.
Goel’s leap—he says it came to him one hot, sleepless night in Kanpur—was to modify the no-slip condition in the equations. “I said, well, okay, if this is zero, what happens if I made it some constant I could control?” he recounts. “Then, because the air is being essentially speeded up at the surface, it has more resistance to being turned around.”
In other words, if the wing surface were moving in the same direction as the air passing over, it might help stave off separation and maintain laminar flow. “The ‘holy crap’ moment, one morning at 3 am or 4 am, was that we could build a really awesome plane around this,” Goel says. (Mathematically, giving a wing a moving surface also changes a quantity called circulation, further increasing lift.) Here’s an animation produced by Vires Aero that makes the whole concept clearer.
Of course, it’s a long way from an idea and a pretty graphic to a working wing with a moving control surface. That’s the gap Vires Aero is now working to close.
Goel knew Greene, a business major, from their freshman dorm at Berkeley. Greene, in turn, knew Hargreaves, a computer science major with lab experience programming swarms of small UAVs, from an engineering class they’d taken together. The three decided to team up. “As soon as I saw some of the data Harshil was showing me, my ears started to perk up and I wanted to be involved,” Hargreaves says. “UAVs are a passion of mine, and I wanted to see if I could make a career out of a hobby.”
For its first few months, Vires worked from an incubator space at Lawrence Livermore National Laboratory, pursuing an unrelated idea Goel had dreamed up for a new type of vehicle transmission. They picked the name Vires because it means “forces” or “powers” in Latin; Greene says it was also an acronym for “virtually infinite rotary exponentiation system.”
But the wing idea proved much easier to explain to potential customers. “Everyone was telling us how amazing the drone market was going to be, and that we should definitely pursue that,” Greene says. “We said all right, let’s pivot.” (Which also meant a slight change under the hood: Vires now stands for “velocity injecting rotary enhancement system.”)
The Vires Aero team is hardly the first to consider the idea of building a system into wing surfaces to alter air flow. For decades, engineers have experimented with circulation control technology that prevents boundary layer separation by blowing compressed air through slots in a wing’s leading and trailing edges. But one longstanding problem with circulation control is that the compressed air is usually diverted from the engines. That decreases their power, which cancels out the added lift from the blowers.
The Vires Aero team felt they could achieve greater lift improvements at far less energy cost. To try the idea in the field, the company raised seed funding in 2013 from Draper Associates, Promus Ventures, and a group of angel investors. It also joined Lemnos Labs, the San Francisco-based hardware accelerator that has incubated companies in music, food, transportation, robotics, and even nanosatellite technology. At Lemnos, the Vires team has focused on building a prototype: a two-meter wing assembly with rip-stop nylon belts that attaches to a small remote-controlled drone.
Several times a week, the team goes to an open space and flies that craft at various speeds, weights, and angles of attack, gathering data that can be compiled into what Greene calls a “performance envelope.” Once the tests are done, the company will be able to predict how much the technology might increase the lift coefficient for other types of aircraft, Greene says.
“Customers always ask us the same thing: what can you do for us?” Greene says. “We say, tell us about your wing and we will tell you the performance specs we can get you.” Recently the startup began taking pre-orders for a kit called the Aquila MK-1 that will allow UAV makers to test the technology on their own.
One attraction of the belted-wing technology is that it can, in theory, be programmed to respond to sensor data in near-real time. Hargreaves, whose research at Berkeley included a project to turn beetles into cyborg UAVs, is in charge of the needed software. “You have to take into account the angle the vehicle is at, the wind speed, the speed relative to the ground and the air, the pitch and yaw, the inertial moments; and according to all that you can adjust the system dynamically to perform optimally for that regime of air flow,” Hargreaves says. “That’s where I come into the equation, in terms of giving this wing a brain.”
The extra motors and processors in the Aquila wing add only add about 300 grams to its 3.4-kilogram mass, and the motors draw only about 5 to 10 percent of the power normally going to the drone’s propeller. It all pays off in terms of increased lift, Goel says.
The main complaints of many UAV buyers, he says, are that the craft can’t fly far enough and can’t carry enough weight. He does a back-of-the-envelope calculation to shows how Vires Aero could help with that headache. A Boeing ScanEagle, a UAV widely used by military organizations, weighs 18 kilograms and has a 22-kilogram maximum takeoff weight, so it can carry a 4-kilogram payload—say, cameras, radar, or other reconnaissance gear. If Vires Aero’s wings can double the craft’s lift coefficient, the new maximum take-off weight would be 44 kilograms, which means the ScanEagle could carry 26 kilograms, minus the weight of the new wing mechanism.
“You can play with the numbers, but it’s easy to see that the difference is astronomical,” Goel says.
At this stage, Vires Aero is all about numbers, and testing whether the flight measurements live up to Goel’s projections. The company will also have to figure out how to handle everyday problems like malfunctions in the moving parts, water and dirt on the wings, and the like.
But even if the Aquila wing proves itself, the tiny company—which has only one full-time employee in addition to the three co-founders—probably won’t evolve into a full-fledged UAV manufacturer. More likely, Greene says, it will just make wings, or license the technology to other companies. The company is going to focus for the time being on small craft, up to ScanEagle size—the Boeing craft has a 3-meter wingspan. Beyond that, it would need to partner with, or raise capital from, larger aerospace companies.
But those problems are in the future. Right now, the biggest challenge for Vires Aero is just proving that Goel’s late-night math breakthrough in Kanpur will work in practice.
“People are very reluctant to move away from the conventional processes that have been used for the last 100 years,” Greene says. “Trying to convince people who are not aerospace engineers and mathematicians that this will work is important, and is a feat in itself.”