Variable Drag Airbrake System

Attempting to target a specific altitude through a combination of mechanical design, control systems, and electronic integration

Defining the Problem

At our yearly competition, achieving a specific apogee is the name of the game. Maximum points are awarded for teams that have the smallest error relative to their target, in our case 10,000 feet. As the Avionics team lead for the 2022-23 competition team, I conceptualized and presented the idea of an airbrake system; with a rocket designed to purposefully fly above the target, it would utilize live input data to adjust drag throughout flight and eventually bring the rocket to a stop at the intended altitude. After discussing with the rest of our team, I was given the go-ahead to lead the Avionics team through this all-new design process.

Mechanical Design

A simple yet robust mechanism is essential to a successful braking system. Along with the rest of the my Avionics team, Quinn Edwards was one of the principal designers for this system. Check out his portfolio for more on this project and other awesome designs!

A Simple Mechanism with Advanced Components

  • Ideation

    Deciding on the motion profile was critical to a successful design. From spinning radial leaves to umbrella-style extensions, it was necessary to weigh all the options. The most important consideration was simplicity in two specific areas: manufacturing and how the added area could be considered in drag simulations.

  • Selected Mechanism

    Selecting a simple linear motion profile cascades through the rest of the project, and this was the primary reasoning behind the choice. For the control system design, extending uniform drag blades in this linear fashion results in the most straightforward drag-to-extension relationship. The central servo motor and accompanying linkages convert rotary motion into linear motion.

  • Simple Design

    I presented the idea of using a “top hat” stucture, where these two large custom aluminum parts make up the bulk of the structure. This design allows the airbrake to sit flush with the rest of the rocket while sufficiently extending into adjacent carbon fiber tubes. The individual blades sit on linear rails containing ball bearings for smooth motion.

  • Manufacturing

    As one of the largest and most involved parts on the rocket, our most seasoned machinist and 2022-23 President John Smalley took the lead on manufacturing the airbrake housings. While John developed the G-Code, I got to spend time running the parts on the lathe, Wire EDM, and CNC mill.

  • Advanced Composites

    While the initial design incorporated aluminum drag blades, our Structures team presented the idea of using “forged carbon.” Led by Vinnessa Van, they 3D printed precise molds and filled them with a mixture of epoxy and carbon fiber strands. After compressing the material for even resin distribution, the result was an incredibly strong and rigid material that was 33% lighter than aluminum.

  • High Performance Actuation

    Using the high-performance Spektrum A6350 servo motor allowed the airbrakes to actuate incredibly quickly (though slowed down in the image above). The high torque rating gave the mechanism the ability to handle the strong forces of flight and counteract the moment arm generated at full extension.

Control System and Flight Computer

The primary problem with this system is uni-directional control. Once the motor has finished its burnout, only drag can be added; no more thrust is available. This means that control system development revolved around avoiding a classic controls issue: overshoot. If the system adds too much drag and the rocket is headed below target altitude, there is no way to recover.

Theory and Ideation

Following conversations with my controls professor, my team considered a number of options to implement into this system. While many different control theories are in existence, our professor’s high level recommendation involved Model Predictive Control. However, this complex realm of control theory appeared to be a tall task in a short design cycle.

After considering other methods, we settled on a simplified version of Explicit MPC that incorporated more managable calculations for our onboard computer. To overcome the uni-directional control problem within the capabilities of our team, we instead considered a system that utilized intensive calculations and simulations before the flight. Picture this: as soon as the motor finishes its burn, the computer takes the current velocity and altitude of the rocket. Then, it compares this input data to a large matrix of pre-simulated “flightplans” that are stored onboard. Each flightplan incorporates an airbrake setpoint, the neutral state of the airbrakes for the remainder of flight.

Through simulation, each flightplan ends up as close to 10,000 feet as possible. Through the remainder of flight, the system performs proportional control around the selected flightplan based on live input data that is interpolated against a stored set of points on the plan. In order to “speed up,” the system simply brings the airbrakes to a smaller deployment level than the chosen setpoint. The opposite is also true, which gives the system to handle overshoot in either direction and theoretically target 10,000 feet. For a clear picture as to how this works, see the attached graphic (courtesy of Quinn Edwards).

Simulation and Flight Computers

In order to generate all of those flightplans, advanced simulations had to be done. Our Propulsion team lead, Sage Cooley, provided substantial work in Ansys to develop models for airbrake behavior. From there, Quinn Edwards spent some time adapting the OpenRocket source code into Python and adding his own modifications for airbrake simulation and flightplan generation. Before the flight, a 64x64 matrix of plans was produced and stored onboard the custom flight computer designed by the team.

This flight computer board was designed in-house, and programmed primarily by Luke Andresen and myself. Check out my page outlining the competition rocket for Spaceport America Cup 2023 for more details about this and the other rocket parts I designed!