Browsing by Author "Buysse, Aaron"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- ItemDevelopment of a flight control system for the subsonicwing deployment of a reusable rocket booster(Stellenbosch : Stellenbosch University, 2018-03) Buysse, Aaron; Steyn, W. Herman; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: A flight control system and a navigation system have been developed for the subsonic wing deployment of a reusable first stage rocket booster. In previous research, the booster’s wing was designed to rotate about a single pivot point, resulting in an oblique wing aircraft while partially deployed. Additionally, the booster was designed to utilise deployable propellers powered by a piston engine for thrust during the flyback portion of the return flight as an unmanned aircraft. This work defined a concept of operations for the booster for the subsonic return flight prior to starting the piston engine. The aerodynamics of the booster were modelled in the rocket configuration, in the oblique wing configuration and in the aircraft configuration to form the basis for the controller designs. Furthermore, the criticalMach number of the booster’s wing in the oblique wing configuration and in the aircraft configuration was determined, and constrains the airspeed at which the wing may be deployed. Since the critical Mach number increases with increasing wing sweep in the oblique wing configuration, the booster first partially deploys the wing to generate lift at a higher velocity and altitude. A pull-up manoeuvre was then performed to further reduce the airspeed by exchanging kinetic energy for gravitational potential energy, before fully deploying the wing. With the wing fully deployed, a minimum sink rate glide was established in order to maximise the time available to start the piston engine. A cascaded control architecture was designed to handle the significant variation in the booster’s geometry throughout the wing deployment, as well as the dynamic cross-coupling resulting from the oblique wing configuration. The control architecture consists of four parts. An outer loop uses single-input, single-output controllers to achieve the pull-up manoeuvre and the gliding flight by commanding an attitude. Two separate inner loops use multiple-input, multiple-output controllers to first track this attitude by commanding an angular velocity, and then track this angular velocity by commanding a body moment. Finally, control allocation is utilised to distribute the required moment to the six aerodynamic control surfaces on the booster, which vary in effectiveness and availability throughout the wing deployment. The navigation system uses low-cost, commercially available sensors fused with two separate Kalman filters to estimate the attitude, position and velocity of the booster. A multiplicative extendedKalman filter estimates the booster’s attitude using a gyroscope and measurements of inertially referenced vectors in the body frame. The reference vectors utilised were Earth’s magnetic and gravity fields. The second Kalman filter estimates the position and velocity using the accelerometer and a Global Positioning System. The inertial velocity output from this filter was differentiated in order to estimate the dynamic acceleration, which was subtracted from an accelerometer measurement to provide the gravity vector measurement to the attitude filter. The control system and navigation system were verified in a six degree-of-freedom simulation. Atmospheric disturbances were modelled by static winds, wind gusts and turbulence. Imperfect sensors were accounted for by introducing characteristic measurement noise in the simulation.