• 01 Mar 2012 10:45 AM
• 0

From Concept to Prototype in 180 Days

By Pascal Chretien

Vertiflite, March/April 2012

The Project

The advent of high power density batteries has prompted the emergence of highly efficient electric powertrains now flying on light fixed-wing aircraft. Rotary-wing aircraft, with one of the highest power demands in the aviation world, has been left outside of the electric realm. Designing a 100% electrically-powered manned helicopter is a highly challenging exercise and hasn’t been successfully attempted until today.

In late 2009, Eric Chantriaux, the managing director of Solution F, a French-based company, decided to design, build and fly the world’s first fully electric powered helicopter. The deadline was aggressive: 6 months from concept to prototype. Real design work started in August 2010 and machining started the following month. All development was conducted in Venelles, France.

Design

With a consumption of 8 to 10% of total hover power, a tail rotor was not welcome. Side by side, intermeshing, propeller arrays, and coaxial configurations were all studied. It was found that a coaxial twin rotor would lead to the best compromise between airframe weight and rotor efficiency. Computational Fluid Dynamics (CFD) analysis using ANSYS Fluent showed that, for power optimization, unequal rotor diameters were appropriate.

Conventional cyclic and flight controls were replaced by a weight shifting system, resulting in significant weight savings. However this arrangement required flying with reversed roll and pitch controls, and a mechanical flight simulator was subsequently developed for training. Free flight stability analysis was conducted using MSC Adams.

An energy absorbing composite landing gear was designed to withstand a 2 m (6.6 ft) drop and protect the 58 kg (128 lb) of lithium-ion polymer batteries located under the pilot’s seat. Due to cost and time constraints, the airframe was made of welded 7020 aluminum tubing. Although a few kilograms heavier than a composite airframe, this could be produced in days while still possessing fair and predictable crashworthiness.

Rotor System Analysis

Two distinct methods were used to determine power requirements in hover and in flight: an analytical method run on MATLAB based on relations derived from flight tests, and by 3D CFD analysis using FLUENT (Figure 2). The convergence of both methods validated power requirement estimation.

With a main rotor disk plane located 2.5 m (8.2 ft) above the skids, and a radius of 2 m, the rotor system operates out of ground effect most of the time.

A somewhat unorthodox rotor system is used because given a fixed pitch system, rotor inertia is the enemy of power response; consequently, the blades had to be made as light as possible. An extruded multi-cell structure (Figure 3) keeps the blade’s weight down to 1.9 kg (4.2 lb), yet offers outstanding torsional stiffness. Although not being optimum as far as fatigue is concerned, 6063 T6 aluminum alloy offers slow crack propagation, and is an acceptable solution considering the short life of this demonstrator.

Leading edge balancing masses that usually bring the center of gravity near the center of lift were not used. The retention plates are located at 25% chord length, and the whole blade is pivoted forward by the lag link, at such an angle that the midspan center of gravity projected at the hub level is located at 25% chord.

The retention plates were bonded onto the airfoil to spread the stresses out, as blade skin is only 1.1 mm (0.043 inches) thick and fasteners lead to unacceptably high localized stresses. The Finite Element Analysis (FEA) graphic in Figure 4 depicts the situation.

Although this helicopter is unable to autorotate – the machine exhibits an equivalent hover time (t/k factor) that violates common best practices guidelines – this has to be put in the context of a demonstrator designed to hover, and/or fly close to the ground, and this drawback was accepted in view of the substantial weight saving achieved by removing the swashplates, and control linkages.

Drivetrain

The entire machine’s design revolves around the powertrain. At a takeoff weight of 247 kg (545 lb), 32 kW (43 hp) distributed over two Agni DC motors is necessary to hover out of ground effect at 1000 feet above mean sea level, at ISA conditions.

Brushed motors were employed due to their fair performance, both in torque and efficiency (91.5%), along with ease of integration (no need for optical encoders, Hall Effect sensors, etc.). Controllers for DC motors are simple and reliable, yet very efficient.

Custom-made metal oxide semiconductor field-effect transistor (MOSFET) controllers offering 98.5% efficiency were used, with no need for a heavy cooling system.

The drawback of such a solution is the relatively low operating voltage of such semiconductors, requiring higher currents, hence heavier bus bars; however, those feather-light controllers (1.7 kg for 20 kW continuous – 3.75 lb for 27 hp) offered significant weight saving (9 kg/20 lb), compared to an insulated gate bipolar transistor (IGBT)/brushless solution and the resulting drivetrain’s weight budget is remarkably good. End-to-end drivetrain efficiency from battery terminals to rotor shafts is 87.5%.

Energy

The Rechargeable Energy Storage System (RESS) is designed around Kokam’s Li-ion polymer pouch cells, providing 160 W-h/kg (0.1 hp-h/lb), and using a sophisticated battery management system. A combination of conduction and convection cooling based on aluminum honeycomb heat sinking structure enabled the RESS to deliver 43 kW (58 hp) continuous and 52 kW (70 hp) peak for 10 seconds. The tricky thermal instability of lithium/cobalt chemistry does not leave room for error, and advanced thermal and electrical analysis had to be conducted. The most drastic temperature rise of large pouch cells when discharged at high current for short duration occurs in the terminal area where current density is the highest. If not properly cooled, localized aging can occur, hence reducing cell capacity. Battery bank design and thermal analysis represented a substantial part of the whole project time. An ultralight aerogel-based firewall was used between the battery bank and the seat.

If put in short circuit, Li-ion polymer batteries can deliver extremely high peaks of currents that will vaporize small conductors, or induce battery thermal runaway if the short is of a heavier gauge (e.g. airframe). To avoid this danger, common sense dictates distributing fuses along the chain of cells. However, due to the high currents involved (620 amps max), the fuse weight was unacceptably high – 1.5 kg (3.3 lb) per fuse – so it was decided to trade safety for weight. It is important to note that the cells themselves are not a danger; Kokam nail-tests all their products and they are proven to be safe, but the particular assembly of cells makes the pack less safe without fuses. In the event of a crash, the possible short circuit could be problematic.

Flight Controls

Roll and pitch are achieved through a gimbaled front end. Yaw control is achieved through a combination of electric controls in the form of resolvers linked to the yaw pedals and acting on the controllers, as well as mechanical linkage acting on a tail fin that intercepts the rotors’ downwash. The tail fin produces instantaneous yaw response by deflecting rotor downwash. As for yaw, “collective” pitch is an electrical control, in the form of a flat wheel located on the control stick.

Electrical flight control management was achieved by a triple redundant op-amp-based processor. Going analog instead of using a common microcontroller may appear like a Stone Age choice, but it was a deliberate decision made to speed up development, and prevent possible crashes due to program glitches. Digital flight control systems

take a lot of time to be fully tested as programming faults can lie in the software for a long time before being detected. Once fine-tuned, this analog processor proved to be extremely resilient to electromagnetic interference and worked flawlessly. The development, construction and integration took less than two weeks.

Assembly

Airframe and electronics assembly started at the end of 2010, and took less than three months.

By the end of March, the first ground testing on wheels could validate the effectiveness of mechanical and electrical flight controls. A lift test with the machine resting on four raised scales, revealed that the out-of-ground effect thrust was 260 kg (573 lb) with NACA0012 airfoils, later replaced by 8H12 airfoils. In the later test the lift went up to 310 kg (683 lb). With an all up weight of 247 kg (545 lb), the power margin was sufficient for takeoff.

By the end of June, the tail fin was installed, the landing gear adjustment was completed, and the electronics modifications were finished; the first flight tests could start, preceded by a delicate period of adjustments to the rotors and motors. Strong aerodynamic coupling between rotors meant that it took several iterations to achieve a precise torque balance on the rotor shafts. An initial series of tethered flights, where the machine was restrained by four small Spectra lines proved that roll and pitch controls were working as expected.

In order to preserve the battery pack’s life, only shallow discharges were allowed, equal to about four minutes of flying, although a few flights, intended to gather thermal data on the batteries, lasted up to six minutes. High summer temperatures precluded day flights, and the machine was flown at sunrise only. A hover, witnessed by a court bailiff was carried out on August 12, 2011. It was logged as the first takeoff and flight, under its own power, of a manned electric helicopter.

The craft was dismantled and transferred to a larger pad, where it could be safely flown. Free flights were resumed in the second half of August 2011. The picture below is a capture from the videos shot during free flights.

Flight Results

Despite initial concerns, the machine showed an excellent power reserve and the Agni motors, although working at a high rating, behaved flawlessly. To date, the craft has accumulated 99.5 minutes of flying time over 29 flights. A typical flight lasted four minutes, with a demonstrated maximum of six minutes. Additional battery cooling would be required for longer flights.

Ground resonance was a concern during the initial design phase. As both rotors may rotate at slightly different speeds producing low frequency beating transmitted through the gimbaled assembly, it could drive the airframe into ground resonance. This constraint drove the design of the landing gear, and a highly dissipative structure presenting a flat response over a broad range of frequency was employed. Furthermore, oscillations of the gimbaled front end are damped by dissipative bungees. Using rubber compounds has proven to be a light and effective solution. In practice, each rotor was separately adjusted to 0.1 inches/sec, and then only could both rotors be run together. No ground resonance was ever experienced.

What Next?

This challenging exercise provided a wealth of data on electrical propulsion. Although less than practical, the purpose of building this machine was to show that it could be done and to provide a springboard for future research. The most cost-effective way to “recycle” this machine would be to convert it into a hybrid UAV/optionally-manned vehicle, by integrating an autopilot that would enable autonomous flights, or an interpreted flight control system in a manned configuration, as the present pilot’s workloads are quite high.

The end goal of this demonstrator is to pave the way to a hybrid helicopter, where gears, clutches and shafts will all disappear. Instead, copper, batteries and an electromagnetic transmission will be used, in a highly redundant and flexible way. Key advantages of this configuration range from preventing autorotation through the use of sufficient onboard batteries enabling powered landing in case of generator failure, to significant power reserve at takeoff, with a drivetrain’s weight no heavier than that of a conventional architecture.

Several patents have been filed jointly by Solution F and Pascal Chretien. They cover, among other things, the architecture of a new electromagnetic transmission, the details of such a transmission, new battery pack architecture and others features. They may contribute to the development of a new class of rotary-wing machines offering unprecedented levels of safety and resilience to ballistic impact, at low operating costs.

About the Author

Pascal Chretien has engineering degrees in electronics and aerospace, as well a commercial helicopter pilot license (CHPL), with aerial work experience in Canada and Australia. He led the design and development of the Solution F aircraft.