The First Prototype

Like many things, Project Everloft began with a simple question: Is it possible, using solar cells and batteries, to build a plane capable of indefinite flight? And if so, could we build one as students? After preliminary calculations, simulations, and planning, we concluded the answer is yes.

With that, we started investigating the design options for this solar UAV (Unmanned Aerial Vehicle). The most important design objective, we found, is to increase the efficiency of every part of the system. With a vehicle as energy constrained as a solar plane, every decrease in power consumption means that less batteries and solar panels are needed, reducing the mass and in turn reducing the power consumption again. This positive feedback loop makes relentless optimization necessary.

One way we found to increase aerodynamic efficiency and reduce mass is removing the tail that typically includes a horizontal and vertical stabilizer, turning the plane into a flying wing. This makes the plane aerodynamically unstable around the pitch and yaw axis, requiring a combination of special “reflexed” airfoils, sweep and twist to make it stable again. These changes reduce aerodynamic efficiency, partly canceling out the effect of removing the tail.

But there is another option: Keeping the plane unstable. Modern flight controllers are more than capable of actively stabilizing everything from propulsively landing rockets to fighter jet. With this design decision, the plane can be designed with efficiency as the single focus, letting the flight controller do the work of stabilizing.

The Design:

To prove that we can build an unstable UAV, we set out to design the simplest, easiest to build test platform we could think of. This prototype could have the aerodynamic efficiency of a brick, as long as it demonstrated the necessary things.

The main design decisions are as follows:

• Choose an airfoil that is only slightly unstable, in our case a NACA 23012. Make the wing planform a simple rectangle.
• Use an aluminum profile as a “fuselage”, increasing the inertia around the pitch axis. This makes it easier for the controller to stabilize the plane.
• Use the PID controller integrated into Ardupilot for stabilization. Ardupilot includes a robust controller architecture and good autotuning capabilities. A well-designed PID controller is sufficient for our application.
• Make the battery moveable on the aluminum profile, to change the center-of-gravity. This allows a gradual change from unstable to stable flight, retuning the controller gains at each step.
• Use a standard vertical tail. Stabilization in the yaw-axis would later be done using differential thrust, but as this model focused on instability in the pitch axis and only included one motor, this was not tested.

The plane was built within a few weeks. The wings were cut from XPS foam using a self-built CNC hot-wire cutter, a piece of wood served as the main structural member, and the surface received three coats of high-visibility orange, later dubbed “structural paint.”

Simulation:

Parallel to building the plane, we worked on a 3 degree-of-freedom Simulink model. This would allow us to predict the behavior of the plane around the pitch axis and compare it to flight test data. The model was fed with aerodynamic data from xflr5 and inertia/mass data from the finished prototype. A partial recreation of the ardupilot flight controller was also included to allow for pretuning of controller gains in the future.

Using this simulation, we found that mass and inertia are the most important parameters to predict how easy active stabilization will be. A high mass but low inertia means that the controller has [add more text]

Flight Tests:

During the first flight tests, still in a stable configuration and without any active stabilization, the aircraft was hard to control and sensitive to wind gusts. Still, the plane was successfully landed every time, in part due to great piloting by Johannes, who has years of experience flying gliders and experimental RC planes. To alleviate the controllability problems, a pitot tube was added to give the flight controller accurate air speed data, and the area of the vertical stabilizer was increased. This largely mitigated the problem and allowed us to tune the controller gains and decrease stability in the next flights.

[add text for for other flights and crash]