This thesis first addresses an important topic concerning space structure control systems, namely, attitude stabilization and control, which is followed by a study on subsystem interactions of general Multi Input Multi Output (MIMO) systems for better performance and actuator fault tolerance. A novel and simple output feedback stabilization approach is proposed for a space structure system characterized with kinematics and dynamics. The approach globally, asymptotically stabilizes the plant and the closed-loop stability is proved using Lyapunov analysis. The simplicity and robustness of the designed controller are demonstrated by investigating the closed-loop response after reducing the degree of freedom in control structure. The stability of the closed-loop system is further analyzed and the performance is compared with two other robust control approaches. The study carries on to another space plant, a Large Space Telescope (LST). Its dynamic model which is fitted with reaction wheels initially developed by NASA is analyzed and the fully coupled dynamics are derived by taking into account the nonlinear coupling phenomena and other terms neglected in their original (NASA) form. The dynamics are combined with Quaternion based kinematics to form an intricate yet realistic LST attitude model. The attitude of the nonlinear LST model is stabilized using a state feedback controller and the LST model is shown to track a time varying attitude reference. Structure configuration is an imperative task in the design of MIMO control systems. In order to make use of interactions between multiple channels so that the system can deal with vulnerability due to actuator degradation, a novel interaction measure is proposed. It is defined as Relative Dependency Index (RDI) and is based on H∞ norms. Such a measurement is effective in understanding the influence of the jth input on the ith output of a system. RDI based guidelines are outlined for configuring a system towards coupling/decoupling. RDI is further extended to the Input Impact Index (i.i.i.) which helps in determining how much an actuator degradation would affect the output of a system. The validity of RDI and i.i.i. is illustrated by simulation results and tested on the linearized spacecraft attitude model presented in the former part of the thesis.

Identifying the detrimental effect of environmental factors and subsystem interactions are one of themost challenging aspects of early hazard assessment in the design of complex safety critical systems.Therefore, a complete understanding of potential failure effects before the catastrophe happens is a verydifficult task. The thesis proposes a model-based hazard analysis procedure for early identification ofpotential safety issues caused by unexpected environmental factors and subsystem interactions within acomplex safety critical system. The proposed methodology maps hazard and vulnerability modes tospecific components in the system and analyzes the hazard propagation paths for risk control andprotection strategies. The main advantage of the proposed method is the ability to provide the designerswith means to use low-fidelity, high level models to identify hazardous interactions. Using thistechnique, designers can examine the collective impacts of environmental and subsystem risks onoverall system during early stages of design and develop a hazard mitigation strategy.

The conventional approach to vehicle design is restrictive, limited, andbiased. This often leads to sub-optimal utilisation of vehicle capabilities and allocated resources and ultimately entails the repercussions of designing andlater on an using an inefficient vehicle. To overcome these limitations, it is important to gain a deeper understanding of the interaction effects at component,subsystem, and system level. In this thesis, the research is focused on identifying appropriate methods and developing robust models to facilitate the interaction analysis. To scrutinise and identify appropriate methods, criteria were developed.Initially, the Design Structure Matrix (DSM) and its variations were examined.While DSM proved to be fundamental for capturing interaction effects,it lacked the ability to answer questions about the structure and behaviour ofinteractions and to predict unintended effects. Therefore, network theory wasexplored as a complementary method to DSM which was capable of providing insights into interaction structures and identifying influential variables. Subsequently, two criteria were established to identify subsystems significant to interaction analysis: high connectivity to other subsystems and multidisciplinary composition. The traction motor was observed to satisfyboth criteria as it had higher connectivity with other subsystems and was composed of multiple disciplines. Therefore, a detailed model of an induction motor was developed to enable the interaction analysis. The induction motor model was integrated into a cross-scalar design tool.The tool employed a two-step process: translating operational parametersto motor inputs using Newtonian equations and deriving physical attributes,performance characteristics, and performance attributes of the motor. Comparing the obtained performance characteristics curve against existing studiesvalidated the model’s reliability and capabilities. The design tool demonstrated adaptability to different drive cycles and the ability to modify motor performance without affecting operational parameters. Thus validating the capability of the design tool to capture cross-scalar and intra-subsystem interaction effects. To examine inter-subsystem interaction, a thermal model of an inverter was developed, capturing temperature variations in the power electronics based on motor inputs. The design tool successfully captured interaction effects between motor and inverter designs, highlighting the interplay with operational parameters. Thus, this thesis identifies methods for interaction analysis and develops robust subsystem models. The integrated design tool effectively captures intra-subsystem, inter-subsystem, and cross-scalar interaction effects. The research presented contributes to the overarching project goal of developing methods and models that capture interaction effects and in turn serve as a guiding tool for designers to understand the consequences of their design choices. / Det konventionella tillvägagångssättet för fordonsdesign är restriktiv, begränsat och partiskt. Detta leder ofta till en suboptimal användning av fordonets kapacitet och tilldelade resurser och innebär i slutändan att konsekvenserna blir att använda ett ineffektivt fordon. För att övervinna dessa begränsningar är det viktigt att få en djupare förståelse för interaktionseffekterna på komponent-, delsystem- och systemsnivå. I denna avhandling fokuserar forskningen på att identifiera lämpliga metoder och utveckla robusta modeller för att underlätta interaktionsanalysen. För att granska och identifiera lämpliga metoder utvecklades kriterier. Till att börja med undersöktes Design Structure Matrix (DSM) och dess variationer. Medan DSM visade sig vara grundläggande för att fånga interaktionseffekter, saknade den förmågan att besvara frågor om interaktionsstrukturer och beteende samt förutsäga oavsiktliga effekter. Därför utforskades nätverksteori som en kompletterande metod till DSM, vilket kunde ge insikter i interaktionsstrukturer och identifiera inflytelserika variabler. Därefter etablerades två kriterier för att identifiera delsystem som är betydelsefulla för interaktionsanalysen: hög anslutning till andra delsystem och mångdisciplinär sammansättning. Dragkraftmotorn observerades uppfylla båda kriterierna eftersom den hade högre anslutning till andra delsystem och var sammansatt av flera discipliner. Därför utvecklades en detaljerad modell av en induktionsmotor för att möjliggöra interaktionsanalysen. Induktionsmotormodellen integrerades i ett tvärskaligt designverktyg. Verktyget använde en tvåstegsprocess: att översätta operativa parametrar till motorinsatser med hjälp av Newtons ekvationer och härleda fysiska egenskaper, prestandakaraktäristik och prestandaattribut hos motorn. Jämförelse av den erhållna prestandakaraktäristikkurvan med befintliga studier validerade modellens tillförlitlighet och förmågor. Designverktyget visade anpassningsbarhet till olika körcykler och förmågan att modifiera motorprestanda utan att påverka operativa parametrar. Detta validerade designverktygets förmåga att fånga tvärskaliga och intra-subsystem interaktionseffekter. För att undersöka inter-subsysteminteraktion utvecklades en termisk modell av en inverter, som fångade temperaturvariationer i kraftelektroniken baserat på motorns styrning. Designverktyget fångade framgångsrikt interaktionseffekter mellan motor- och inverterdesign och belyste samspelet med operativa parametrar. Därmed identifierar denna avhandling metoder för interaktionsanalys och utvecklar robusta delsystemmodeller. Det integrerade designverktyget fångar effektivt intra-subsystem-, inter-subsystem- och tvärskaliga interaktionseffekter. Den presenterade forskningen bidrar till det övergripande projektets mål att utveckla metoder och modeller som fångar interaktionseffekter och i sin tur fungerar som ett vägledande verktyg för designers att förstå konsekvenserna av sina designval. / <p>QC 231003</p>

The research reported in this dissertation focuses on the development of optimization tools for the design of subsystems in a modern aircraft power distribution system. The baseline power distribution system is built around a 270V DC bus. One of the distinguishing features of this power distribution system is the presence of regenerative power from the electrically driven flight control actuators and structurally integrated smart actuators back to the DC bus. The key electrical components of the power distribution system are bidirectional switching power converters, which convert, control and condition electrical power between the sources and the loads. The dissertation is divided into three parts. Part I deals with the formulation of an optimization problem for a sample system consisting of a regulated DC-DC buck converter preceded by an input filter. The individual subsystems are optimized first followed by the integrated optimization of the sample system. It is shown that the integrated optimization provides better results than that obtained by integrating the individually optimized systems. Part II presents a detailed study of piezoelectric actuators. This study includes modeling, optimization of the drive amplifier and the development of a current control law for piezoelectric actuators coupled to a simple mechanical structure. Linear and nonlinear methods to study subsystem interaction and stability are studied in Part III. A multivariable impedance ratio criterion applicable to three phase systems is proposed. Bifurcation methods are used to obtain global stability characteristics of interconnected systems. The application of a nonlinear design methodology, widely used in power systems, to incrementally improve the robustness of a system to Hopf bifurcation instability is discussed. / Ph. D.