Evolving Systems Approach to the Attitude Control of a Space-Debris Removal Spacecraft

Abstract

Space debris has become a major issue for the space industry over the last couple of years. The increasing number of uncontrolled objects in space increases the chance of collision with a spacecraft. The unexpected loss of contact with ENVISAT, the European Space Agency’s largest environmental spacecraft, added one more piece of debris to the list. It also sparked ESA’s interest in the active removal of space debris. Numerous removal methods are being investigated, one of which uses a chaser spacecraft that will attach itself to the debris by means of a robot arm and tentacles and will then perform a series of braking manoeuvres to re-enter the Earth’s atmosphere. The proximity operations require an accurate attitude control system. In this thesis, these operations have been divided into three phases: an unconnected phase in which the chaser spacecraft synchronises its motion with that of the (spinning) target debris; a semi-connected phase, where the robot arm and tentacles form a flexible connection between the chaser and target; a connected phase, where the two spacecraft are assumed to be rigidly attached to each other (and form a stack). The objective of this thesis is to investigate the stability and controllability of the system in these three phases. Because the system changes in configuration during its mission, it can be classified as an Evolving System; a system with actively controlled components that mate to form a single connected system. The connection between components becomes stronger during the evolution and can be represented by connection forces and moments. In this thesis, the connection is modelled as a rotational spring and damper acting between the chaser spacecraft and ENVISAT. By developing a representative simulator environment the stability and controllability of the system is assessed. Two major parts of the simulator comprise the attitude control algorithms and control actuators. The latter consists of a Reaction Control System, reaction wheels, and a control allocation algorithm. The attitude control algorithms that are investigated are a Linear Quadratic Regulator and a model reference adaptive controller. A linear stability analysis of the system showed that the system remains stable during its evolution when the chaser is actively controlled. This is partly because the motion of the chaser has little effect on the motion of ENVISAT due to the large size difference between the two spacecraft. However, for a system with equally sized spacecraft, the linear stability analysis showed that instability can occur. Furthermore, in stack configuration, a gravity-gradient stable attitude was found when the minor axis of the system is parallel to the Earth radial. Moreover, for a suitable deorbit attitude and an unobstructed view to the Earth, the only stable attitude is one which has a roll angle of -90 deg. Next, results of the nonlinear simulations revealed that the combination of reaction wheels and attitude thrusters gives worse performance than when only thrusters are used. This is because the reaction wheels are too small to cope with the constantly changing motion of ENVISAT. Second, for the unconnected phase, both the LQR and adaptive controller are able to synchronise the motion of the chaser with ENVISAT. However, the adaptive controller is not able to do this for all possible initial target rotations, whereas the LQR is. Third, the effect of the connection between ENVISAT and the chaser on the stability of the system is very small and both controllers are able to control the system. Fourth, the adaptive controller shows unacceptable performance during the control of the stack, because of the large difference between the reference model for which the adaptive controller was tuned and the actual system. Using a second reference model for the connected phase could improve the performance of the adaptive controller.