Bits and Pieces
How They Did It: Toronto's Sikorsky Prize-Winners
By Cameron Robertson, MASc, P. Eng, AeroVelo
Editor's note: The American Helicopter Society declared Canada's AeroVelo the winner of the Igor I. Sikorsky Human Powered Helicopter Competition in July after verifying its flight data from June 13, 2013. AeroVelo satisfied the three requirements put forth by the AHS Human-Powered Helicopter Committee: Flight of at least 60 seconds in duration, exceeding 3 metres in altitude, and remaining within a 10m x 10m box. The following is a project description provided exclusively to Bits and Pieces by AeroVelo's Cameron Robertson.
Back in May 2011, we saw the University of Maryland's human-powered helicopter Gamera fly. We had become familiar with the Sikorsky Prize while we were engaged in our previous project, the human powered ornithopter, Snowbird.
Ornithopter Dawn Flight
In 2009, the prize pot was increased to $250,000, which made it a break-even proposition at that point. I was working on UAVs, Todd Reichert was finishing his PhD, and we decided this would be something we really wanted to go after. The team worked with no guarantee of any kind, so we tried to do corporate, university and basic research fundraising in order to support the project with materials, construction costs, and logistics. The entire team didn't have any guarantee of pay, but worked really hard, regardless of the outcome. The prize money went to pay the bills, pay off the students and faculty, and some seed money was left over to support the next project which is currently ongoing.
It was really remarkable when we won the prize, given that we had been trying to set up a long-term career path. As we set up the project at the beginning of 2012, we started calling ourselves AeroVelo but we didn't incorporate until December of that year. This was what we had decided that we wanted to do with our lives and the ornithopter, the helicopter, and the new project, land-based "speed bikes" were where we saw ourselves spending our time: pushing the boundaries of efficient aerodynamic design, espousing the philosophy of doing more with less.
It also fits our goal of promoting engineering design education and providing inspiration to young people. We've had so much interest among youngsters in these engineering design marvels, seemingly impossible and trying to explore new paths, that that has become a basic driver setting up AeroVelo and driving our long-term mission.
Aerovelo is at arm's length to the University of Toronto. We have some funding support in terms of scholarship of the students. It's probably about 10 percent of the overall project budget, although it was a little bigger on the ornithopter project despite it being run as an extra-curricular project while we were students. We fit in very well with the university ecosystem with very strong ties to the engineering faculty, and of course the student body. We had set up a human-powered vehicles team about six years ago that does the same kind of work. They do a substantial amount of work on composites, manufacturing, aerodynamics and lightweight design, all of the topics that we have tried to institutionalize at AeroVelo. U of T is currently sponsoring our use of student project space, and we do plan to try to integrate our research with the programs of various professors there.
On the helicopter team, there were ten of us working full time on the project through the summer of 2012. We had a configuration design team as well as a detail design and fabrication team. Dr. Todd Reichert and I started in January of 2012 with the preliminary design, analysis, and development of the design tools that we would need for the most advanced stages of the project. This included the rotor design, sourcing materials right into the summer and Todd was the chief of the aerodynamics design and most of the project management tasks. I was in charge of structural design and I did many of the logistical tasks including sourcing and a share of the sponsorship tasks.
We had eight undergraduate students, mostly from the University of Toronto, mostly from materials science, engineering science and then we had a diverse group of other students that were excellent, very passionate members of the team despite not having so much specialist knowledge. For example, we invited to the team Todd's sister, a religious studies major, her boyfriend, who was in pre-med at the time, and another friend who was actually an architecture student who had worked on the ornithopter with us. The entire team came together to take on all aspects of manufacturing, design, logistics and testing and everyone was super-driven to pull this together in the tight timeline we'd set ourselves.
Todd and I did most of the configuration work January through April 2012, putting the design together including the optimization of the rotor blades, which was a very big part of the challenge. The actual build of the helicopter went from May to August, which was an incredibly fast timeline. We did final on-field integration the third week in August, leaving the last week in August for some initial flight testing, a milestone by which we'd initially intended to win the prize.
We really didn't know, going into this, what kinds of challenges we would encounter, but at this point we knew that we had a functioning aircraft. At this point, it was too flexible and it was difficult to trim. We worked through a lot of issues in the fall of that year and went on to complete multiple flight testing sessions January through April of 2013, and then finally June. We took the time to implement many design changes that we had come up with on the field. Each time we went out, the aircraft demonstrated substantial improvement. Despite having had a functioning helicopter the previous August, the helicopter that won the prize on June 13 really behaved completely differently.
The helicopter, which we decided to call the Atlas, was twice the area of any other design out there, which was something that we had decided upon very early on would be our differentiating factor. We didn't want to be constrained to fly inside a gym, which is what a lot of other human powered helicopters had done. With this increased size came a lot of additional challenges, including the added flexibility of the structure, and the control system. We broke something like six rotor blades in August alone from ground strikes, hitting structural lines and so on.
Our first control system was something known as an "aero-elastic collective." We used small canards on the rotor tips to wash-out or wash-in the rotors to increase or decrease their lift. That led to all sorts of imbalances and challenges so we eventually abandoned that in early 2013.
Our revised control methodology was to use the incredibly flexible structure as a method of controlling the position of the aircraft. In the video you will notice Todd, who was also our lead test pilot, leaning a lot of the time, to vector the thrust of the rotors one way or the other. This turned out to be a very fast and incredibly intuitive control mechanism, so our flexibility problem turned out to be a huge asset.
Much of our time was spent re-jigging the overall structure, trimming to avoid rotor ground strikes and tuning all of the lines that connect the arms to each other. The tips of the arms and the structure are connected to each other with bracing lines that keep everything locked together. In this video from August 2012, you can see the helicopter kind of dancing all over the place and then if you look at the Sikorsky Prize flight the thing is rock solid. It's a substantially different aircraft although it may look to the outsider like the same one.
Atlas Human Powered Helicopter on the Day of the Flight
The other problems were kind of pervasive. We had to have the lightest weight solution absolutely everywhere and that was pretty absorbing throughout the design process. Crashes resulted in repairs, which resulted in weight - but with the abandonment of the original control system, which weighed eight pounds, and going to this new control system which weighed something like a quarter of a pound, that was a huge savings.
We had to remain flexible and creative every step of the way. The materials we used were carbon fibre for the primary structure. The cross bracing lines and trusses were made of Vectran, which is a manufactured fibre spun from a liquid crystal polymer with very high strength and low creep/extension under sustained loads. We used Kevlar for the tube joints and the yellow spools you can see in the video, and for the rotor hubs. We used a lot of insulation foam, balsa wood, and then Mylar film for the flying surfaces.
One of the Rotors and Hubs
The whole helicopter is balanced on a knife edge in terms of controllability, weight, and stiffness. The safety factor is something like 1.3. We had to build and break many test specimens in order to figure out how close they were to failure. One of the early teams in human powered aircraft had a saying: "If it hasn't broken yet, it's too heavy."
Another constraint was, "How much weight do we want vs. the time investment it would take?" With our timeline, we were forced to take a lot of pragmatic decisions. We had fantastic support from Cervelo, which produces some of the lightest bicycles in the world. The company donated the frame of the bike used in the Atlas. This was an example of using something off the shelf that was close to what we needed rather than reinventing the wheel.
After having decided that we didn't want to be constrained by the size of the average gym, we finally had to accept that we had to do the testing indoors. We found an indoor soccer field in Vaughan, Ontario, which we could get from 7 a.m. to 5 p.m. on the days that we were testing. Two hours of that was spent setting up the machine. Four hours of adjustments and trimming the rotor blades left us only about two hours for actual forward progress in learning the flight capabilities and problems. Then we'd have to pack it up.
There was a huge compromise in terms of being methodical and safe in our testing vs. always having to focus on a very tight timeline, especially in terms of the competition. We were always acutely aware that there are no points for second place.
The soccer field was in substantial demand in the evenings so we had to invest about seven lost hours a day for the benefit of about two to three hours of progress. It certainly hobbled our forward momentum.
Human-powered flight is normally done outdoors. Using multi-disciplinary optimization, which is used in advanced aerospace these days, in our early design we had concluded that an ideal rotor radius would have been something like a 12.5-14 metre rotor radius. We ended up going down to 10 metres because it was a leap over previous teams and still close to what the perfect helicopter would be. We felt comfortable with this because we decided we did not have to operate inside a gym and that we'd flown outside before, with the ornithopter.
After that we quickly ruled out flying outdoors because of structural problems. Making it gust tolerant would have meant adding a lot of weight and we realized that just wouldn't work. It's never going to be flown outside. You could never fly it to work for example.
Atlas is an excellent demonstration of the pursuit of aerodynamic perfection which could actually be applied to other flight vehicles but you'll never see a practical human-powered helicopter.
In pursuing the prize, we viewed directional control as fundamental. The successful flight had to climb to 3 metres, stay in the air for 60 seconds be stable within a 10m x 10m square and it couldn't rotate more than 180 degrees. That is pretty challenging when you consider that the aircraft itself is around 40 metres on a side, so this flight box was quite small in comparison. We learned literally 45 minutes before the successful flight that we could retrospectively superimpose the ten metre square box in the flight track. In other words, we didn't have to start in the middle of the box according to the FAI and AHS. We could post-define the box.
We had to breach three metres and then come back down into ground effect, but stay flying for 60 seconds. Along the way, we figured out that all these human-powered helicopters were designed to stay in extreme ground effect, like a foot off the ground. That's the sweet spot for most efficient flight. At 3 metres up, the energy expended just to stay there was more than double.
We had two exciting crashes in learning exactly how to do this, and by that I mean devastating for the team. We learned that we came down from the top of the flight too fast. Helicopters have a phenomenon known vortex ring state. If the helicopter is descending faster than the in-flow velocity, the lift is lost. Because our disk loading is so small, the in-flow velocity is extremely small. Any noticeable descent rate is going to induce vortex ring state and that's what happened on our first two flights to three metres.
The way we avoided it on the successful flight was to descend really slowly, even though it took more out of Todd to do it that way. The very gradual tapered flight down took almost the entire minute. There is a risk of destroying the aircraft every time we go to three metres, usually, it seems, on the way down. Todd had trained extensively for this particular flight profile and he knew that he had the power to achieve the climb and sustain the very slow descent. We had optimized the helicopter for, say, a 15-second climb, 15-second return, and 30 seconds in ground effect. That had to be abandoned and we were just fortunate that Todd was in incredibly good shape, and that the rest of the design allowed us to adapt to this new flight profile.
The day prior to the successful flight, we had spent most of the day trimming and adjusting, and we were very fortunate that we were able to leave the aircraft assembled overnight. We saved hours of time on the day of the flight, but we only had the place until 12:30 and there was a huge soccer training camp coming in after us. In the end, they were gracious enough to let us use the arena until 2 p.m.
We started our test flights with 1M, 1.5M, 2M and then we went for the prize. We were always concerned to have sufficient control authority. Todd was able to get in a full 45-minute warm up, which was crucial to getting the power required. We were able to clarify the box rule. We were always aware that in order to win the Sikorsky prize we were going to need everything to come together perfectly in that one sixty second opportunity.
Whenever we made an attempt, everyone had to be in the best shape possible and so did the helicopter. The previous two tries had resulted in four weeks, and then six weeks of rebuilding. There had been 75 flights beforehand so it took a lot to get it all into that one package. It was a wild moment. We had to trust the spotters to blow the horn to say we were at altitude. We had cameras set up to prove that we'd not only achieved but recorded the altitude with targets using photogrammetry.
I was there under the helicopter making sure we stayed within the box and there was a moment of tense silence at the end of the video and I yelled "Woah"! There was incredible sense of relief and elation. We'd aimed for a height of 2.8 metres because we'd learned that there was actually quite an overshoot from the moment that Todd stopped pedaling hard, and finally the lowest rotor was actually at 3.26 metres from the ground. The average altitude of the helicopter was 3.5 metres with this crazy burst of power that Todd was able to give us. You can see the successful flight here.
We wanted to document the project from the summer of 2012, not only from a social media perspective, but also for posterity. There are now at least a dozen human powered aircraft teams around the world who are doing this for the first time. This was our second successful aircraft. We've studied some of the best and have produced some of the first-hand information that will be useful to these other teams. EAA members will find a lot of this very interesting, the way we made the pre-preg tubes, for example. Some of it is based on model aircraft technology. The blog posts of the summer of 2012 on the AeroVelo web site are largely about the fabrication.
In terms of the competition, it was really very intense. The University of Maryland, one of the top rotorcraft schools in the U.S. and the world, was very close. As the prize was first upped in 2009 they had just previously started a human-powered helicopter project and they spent four years on it. They had their first iteration of the aircraft in 2011. Gamera II first flew in May of 2012 and then they made a number of modifications including making their aircraft slightly larger such that in the summer of 2012 they were right on the edge of a successful prize-flight. Gamera II had flights of 70 seconds. They had flown to 8.5 feet and they were very close, which was driving our timeline, our motivation and how we were going about our testing.
It was really an unbelievable situation to be in. This 33-year historic challenge was now being chased after by not just one team but two. It was really an interesting relationship. In August 2012 we were both flight testing. At Christmas 2012, I went to visit them.
Our two teams weren't really worried about, say, industrial espionage. Our designs were so different. We had contrasting approaches to almost everything. We developed a fascinating camaraderie since we were such a small and elite community chasing after the same objective. That doesn't often bring contemporaries together, especially in the science community.
The documentation of the competition was somewhat lopsided. CBS in the states did a huge piece on the Maryland team and we got no mention at all. They concluded that Maryland had it in the bag only days before our successful flight! It was somewhat of a David and Goliath position to be in and we felt free and unconstrained just to keep on going.
The fundamental difference seems to us to have been that they looked at it as a helicopter problem that needed to be powered by a human, where we saw it as a human powered aircraft that just happened to have wings that rotated. We'd already built the ornithopter, so we'd covered a lot of that ground already and four years of that experience really gave us a leg up. It wasn't as much of a lopsided matchup as had been portrayed.
Where do we go next? We're excited to be working on a very fast human-powered land based vehicle otherwise known as a speedbike! We have been working on a breed of recumbent super-aerodynamic bicycles.
The current world speed record is 133.8 km per hour and we're aiming to break this by going at least 140 kmh. We're very excited by this because a lot of the technologies demonstrated with the helicopter are directly translatable to the speedbike and to automotive applications.
We hope that by showing what you can do with human power, which is less than a horsepower, the opportunities for improvement are limitless. These bikes have less than a tenth of the drag coefficient of the best car. We can do highway speeds on one hundreth of the power of the most economical car and this is something that we hope to be able to showcase to the world in the next few months.
We hope to be setting a new land speed record in September 2014. Producing such a commercial vehicle is not necessarily within our milieu. We are more focused on design, prototyping, and testing, but we're certainly interested in developing technologies that might someday go into commercial production at a much lower cost than we're used to, as a transportation solution, or these technologies might be more directly translated into automotive, and that's certainly something we have on our horizon.