The design of an Attitude Control System for the SPS-2 satellite

Abstract

This thesis covers the design of the Attitude Control System for the SPS-2 configuration, developed at Airbus Defense and Space the Netherlands. The start of the SPS-2 mission consists of three phases: rate damping, Sun Acquisition and Sun Pointing. The SPS-2 satellite has five available mounting places for different payloads. In the first part of the thesis work, different configurations have been considered, of which three were chosen to be considered further. A choice has been made to develop a simulator in Matlab/Simulink, with inclusion of existing models of the GGNCSim repository. The correctness of the implementation of those models has been verified with ADS, where it has been discovered that there was a sign ambiguity in the aerodynamic model. At this stage, the assumption of modeling the complex SPS-2 geometry as a simple cylinder shape has been verified. The second part of the thesis describes the development of the sensor and actuator models, and the control algorithms which have been designed. The sensor and actuator model error parameters have been obtained from the data sheets of different manufacturers. The models have all been verified. The controllers which have been designed are a nominal B-dot detumble controller in the RD mode, a simple PI controller on the velocity in the SA mode and the quaternion-feedback controller in the SP mode. The gains for the different controllers have been selected such that these work for all configurations. The stability has been looked at briefly (gain and phase margins), but no requirements have been imposed on those. In the last part of the thesis, the simulation performance runs have been performed. Here the uncertainties in the error parameters, uncertainties in the satellite geometry, variations in initial conditions and the impact of different actuators have looked at. It has been found that the detumble time is within the requirement of 1.5 orbit, but that the maximum magnetic dipole moment shall be a minimum of 15 Am2. The SA and SP results show that the Sun is in all cases acquired within 1 orbit and that the time to be actual Sun pointing is less than 10 minutes. The pointing accuracy is within 1.92 deg for all cases. For both reaction wheels which have been considered, more than enough margin was found for the angular momentum build-up. As such, there is no need to offload the wheels during the first part of the mission. The SunSpace wheel will suffice. The simulation performance runs have shown that the requirements are met, even in presence of the maximum sensor errors. Critical may be the impact of the Earth-albedo error, which has been roughly estimated. More research shall be performed here. The current simulator structure and models which have been developed give a good basis for the further design of the ACS of the SPS-2, but also for other projects.