Fault Tolerant Attitude Control For Spacecraft With Actuator Failure

A study of on-orbit spacecraft failures was undertaken which resulted in identifying156failures occurring from1980to2005on civil and military spacecraft. These failures were further analyzed that they may be different in types and possibly occur in different spacecraft subsystems. More specifically, nearly32%of the failures studied occurred in spacecraft attitude control system. Moreover, those failures have different impact on the mission. Although there is no perfect system design method that could prevent any failure occurring in spacecraft components, and especially in attitude control system, the lessons learned from the past years show that researchers can seek to’fault-avoidance’techniques such as enhancements in design and fabrication, and elaborate and intensive testing to improve the reliability of spacecraft attitude control system. However, faults do occur in control actuators during their long operational life in harsh space environment leading to malfunction of the entire control system. The faults in actuators may deteriorate system control performance, and even lead to the total loss of the spacecraft and then decrease the mission lifetime. Therefore, it is important to design an autonomous fault tolerant algorithm that operates satisfactorily not only in the absence but also in the presence of actuator faults.More interestingly, it is important to identify the scarcity of work essentially in au-tonomous fault tolerant strategies for spacecraft attitude control system that can be im-plemented during various phases of a mission. Therefore, this dissertation focuses on the design of autonomous fault tolerant attitude control algorithms for spacecraft with redundant actuators. Additionally, many significant challenges must be overcome before an autonomous fault tolerant attitude control system for spacecraft can be realized. The problem statements for this dissertation can be listed as:nonlinear spacecraft attitude models, system uncertainties and external disturbances, control input saturation, and ac-tuator faults. To solve those problems, the main works of this dissertation are as follows:(1) Fit passive fault tolerant control methods conceptually into the autonomous at-titude control system framework, and improve pointing accuracy and establish precise attitude maneuvers even with the failure of control actuators onboard a spacecraft. From practical viewpoint, the main work is to design an attitude control system that can achieve high precision pointing, fast slewing maneuvers in the presence of external disturbances, model uncertainties, and actuator faults.(2) We approach the validation of the proposed control algorithms from two per-spectives:conventional theoretical analysis and numerical simulation study. Analytical studies are performed using adequate nonlinear control theory to demonstrate attitude control performance and stability. Simulation studies are used as a design tool to demon-strate system performance, compliance with design requirements, and validate established theoretical framework.To finish the above stated works, an observer-based backstepping control approach will firstly be proposed, while only one type of actuator faults is investigated, i.e., partial loss of actuator effectiveness. An iterative-learning observer is developed to precisely es-timate the effectiveness factor. With this estimated information, an attitude controller is designed by using backstepping technique. The controller is able to accomplish attitude stabilization maneuver with high-pointing accuracy even in the presence of external dis-turbances and actuator control input. In comparison with the existing methods to handle partial loss of effectiveness fault, the proposed approach can decrease the conservative-ness of the controller.On the basis of the proposed backstepping control design method, immeasurable angular velocity of spacecraft is further taken into account. An attitude fault tolerant control strategy is proposed without angular velocity measurements to tolerate partial loss of actuator effectiveness fault. An terminal sliding mode observer is developed to estimate the immeasurable velocity. By using this estimated value, an velocity-free controller is synthesized to perform attitude tracking maneuver. It also can solve the actuator control input problem, and thus can achieve the objective of attitude fault tolerant control design without using gyro to measure angular velocity.To tolerate the possible faults of actuator and also to improve the fault tolerant ca-pability, the problem of finite-time attitude control is investigated with fast slewing taken into consideration. An adaptive sliding mode control methodology is presented in the framework of global sliding mode control. With application of this scheme, severe actua-tor faults can be tolerated, and the attitude tracking error is governed to be asymptotically stable. The attitude tracking maneuver is accomplished in finite-time, the objective of fast slewing maneuvers is thus realized.At the end of this dissertation, we develop a reconfigurable control strategy based on nonlinear observer techniques for spacecraft with redundant actuators possessing au-tonomous fault tolerant attitude control performance. The design of nonlinear observer is to estimate actuator faults, and this estimated information is employed to implicitly reconfigure the control algorithm. Instead of using an explicit fault detection, isolation, and identification algorithm, this methodology allows the controller to constantly update its parameters. The objective is to provide autonomous fault recovery using a reliable and const-effective control algorithm that accounting for modeling uncertainties, external disturbances, and actuator faults simultaneously.It should be stressed that, the approach for designing a fault tolerant spacecraft atti-tude control system has undergone a fundamental shift in philosophy in the last ten years. Initially, the concept was to design a control law robust enough to guarantee system sta-bility in face of the full range of potential failures. Then, upon the event of a failure, the system would identify the fault, isolate the failure mode, estimate new system parameters, and select new control law gains and/or control input distribution to regain some level of performance. The tuning of gains in control algorithms is not straightforward because the resulting closed-loop error dynamics are nonlinear; moreover, it may not be possi-ble to find constant gains that yield desirable performance for all operating conditions. Therefore, the only means of obtaining a stabilizing control law with good performance is to use an algorithm that can constantly update its parameters. The control methodolo-gy proposed in this thesis minimizes complexities in the control reconfiguration stage by employing a novel fault estimation scheme. Therefore, with the help of this reconstruc-tion approach, the method removes the toil of fault detection, isolation, and identification techniques, and thus the time required to cope with faults is reduced. Furthermore, the proposed strategy use of the reconstructed fault information to reconfigure the controller on-line and in real-time, it thus has less conservativeness.