DYNAMIC SOARING MOTOR GLIDER

INDEPENDENT ENGINEERING RESEARCH PROJECT

JUNE 2022 - MAY 2023

Full research project title:

Wind-Powered Flight: Exploring the Potential of Dynamic Soaring for Unmanned Aerial Vehicles

Summary + Context

I modified model glider to be capable of dynamic soaring by using a custom flight computer, a new dynamic soaring flight algorithm, new sensors, and new internal components. Analysis of flight data revealed statistically significant wind energy harvesting.

Dynamic soaring is a type of flight pattern that harnesses energy from the wind gradient near Earth's surface. Albatrosses, a sea bird, use dynamic soaring to remain at sea for months and even years.

This was an independent project I conducted at home. The purpose was both to advance both the aerospace field and for myself to learn more about the engineering design process.

This project won 2nd Grand Award at the International Science and Engineering Fair (ISEF).

TABLE OF CONTENTS

Engineering Visual Showcase / Gallery

Full Research Paper / Design Report

Skills Used

Design Techniques: Computer Aided Design (CAD), Printed Circuit Board (PCB) Design

Manufacturing: 3D Printing, Epoxy, Soldering, Hand-Cutting

Software Applications: Fusion 360, EasyEDA, Inkscape, Arduino IDE, Visual Studio Code, Git/Github

Programming Languages & Tools: C++, Python, Matplotlib

Project Powerpoint

This powerpoint was used to present my project at the International Science and Engineering Fair (ISEF). It provides a high level summary of the entire project, details its significance, and is useful for contextualizing the various engineering components used in this project.

ISEF 2023 Slide PNGs from ProjectBoard.pdf

Demonstration Video

Engineering Visual Showcase / Gallery

Pictures and descriptions from throughout the project.

The glider in flight at low altitude, preparing to conduct a dynamic soaring test.

Preparing the glider for launch.

The glider moments after takeoff, using its motor to gain altitude.

The original foam glider used as the airframe.

Notebook sketch of the initial design.

The CAD model of the new hardware and flight computer mount overlaid on a photo of the nose of the glider. The red box is the location of the flight computer.

Real photo of the new dynamic soaring "brain" (flight computer + mount) installed in the glider. CAD photos overlaid.

Breadboard prototyping of the flight controller. Testing IMU, compass, time of flight sensor, pitot tube, microSD data collection, and the Teensy 4.1 microcontroller.

Breadboard sensor setup on a long pole to test the pitot tube sensor by sticking it out a car window while driving.

All components mounted inside the body of the glider.

First iteration of the altitude sensor, mounted on a gimbaling servo to account for the airplane's high roll angle during dynamic soaring. Precise altitude measurements are also very important given the low altitude regime of dynamic soaring.

Second iteration, moved onto the wing to allow for a wider FOV than in the belly of the nose.

Third iteration, simplified by mounting the sensor directly to the servo, saving a slight bit of weight and cost.

Flowchart of logic on the flight computer.

Tuning and validating the pitot tube sensor using a leafblower and an anemometer.

The anemometer (and a streamer) were also used to measure wind speed and direction at the test site.

Implemented a low pass filter to reduce signal noise in the IMU. The blue line is the original signal, the red line is the filtered signal.

Labeling of the switches on my transmitter, to allow switching between manual and autopilot, and between datalogging mode and non-datalogging mode.

CAD models showing the versioning of the top hatch.

All the different versions of the top hatch, the bottom hatch, and the gimbal mounts.

Using epoxy to bind together the belly of the aircraft. This part needed to be quite strong, as it also functions as the landing skid.

The rig for hoisting the glider vertically to test the altitude sensor.

Manually holding the glider from the second story of my house.

A computer shows the glider's estimated altitude via serial connection. Markings on the wall indicate height.

Me and my mother looking over the glider after a crash.

Another crash. This crash was a result of the glider flying too low during the dynamic soaring cycle, striking the ground and cartwheeling, fracturing the tail boom. For subsequent flights, a higher margin of error was used.

The same crash, with all components recovered. The main repairs involved reattaching the horizontal stabilizers and repairing the fractured tail boom near the trailing edge of the wing.

After crashing one too many times, I built a custom gimbal rig to hold the glider to test the control loop on a car instead of in the air.

Attaching the glider and gimbal through the sunroof of my family's car to test and tune the flight control loop.

As the car drives down the road, air flows over the aerodynamic surfaces, allowing me to simulate flight conditions without risking further crashes.

The PID tuning interface via serial that I designed so I could actively tune the flight controller while the car was driving.

The glider conducting a dynamic soaring cycle. It rises from low altitude near the ground, where wind speed is low, to a higher altitude, where wind speed is faster, then returns back down, harvesting energy from the difference in wind. This cycle is entirely autonomous.

Final data comparing dynamic soaring flights with "normal" control flights, demonstrating that dynamic soaring increases the net forward acceleration of the glider by a statistically significant margin, suggesting wind energy harvesting.