ESA GNC Conference Papers Repository

In the last decade, the space sector has seen rapid growth with new players entering the business, bringing with them a wave of innovation in spacecraft design and development solutions. This innovation includes the attempt to simplify processes to aid reduction of development cost and time through the use of existing solutions in other sectors and applying them to the space domain. Case in point is the use of Commercial-Off-The-Shelf (COTS) equipment for space applications in general, and specifically for the AOCS subsystem. Their cutting edge technology (high processing power/lower mass/volume/power usage) and low cost are a definite advantage over standard radiation hardened space qualified components. Their across-the-board adoption in space, however, is currently confronted by their robustness and reliability issues, coming from both design as well as cost control practices, like lack of characterisation in radiation environment, shorter lifetime at design, poor lot control etc. Their susceptibility to single event upsets, unpredictable end-of-life performances (for example due to the radiation environment), and maybe even permanent failures, directly impact mission availability and performance. To address these particular challenges, this paper proposes another set of existing solutions, namely advanced and innovative control, estimation, and FDIR techniques. In particular, the aim is early FDIR targeting a varied selection of failure cases for a variety of COTS AOCS equipment. A range of state-of-the-art algorithms are investigated, some of which have shown great potential in other flight control applications like UAVs. Innovative FDI techniques like Adaptable estimation, Model-based FDI, and Machine-learning based FDI, as well as advanced control techniques like adaptive control, and Non-holonomic control are chosen to be studied in detail to assess their potential benefits when coping with COTS failures. These techniques were then prototyped and implemented in an FDIR focussed AOCS simulator (GAFE). To demonstrate possible advantages for space missions compared to standard FDIR practices in the industry, the techniques were applied to two space mission use cases, EarthCARE and OneSat, covering different applications, orbit regime, environment, and design philosophies, thus providing a varied set of mission performance and availability requirements to be fulfilled by the AOCS subsystem. The paper goes on to present the test results for each of the techniques as applied to chosen fault cases with varying fault types, magnitudes, and dynamic conditions. The detection and isolation capabilities and recovery performance are used as criteria to compare to classical fault management concepts for the particular fault cases in the respective missions. The paper also discusses driving aspects and challenges of implementation of the methods for the missions as well as the potential benefits of their use as compared to classical FDI and control techniques for low-cost COTS equipment.