Integration Of Model Based Safety Assessment For Fault Tolerant Flight Control System Design PDF Download

Safety is one of the major concerns in the aviation community for both manned aircraft and unmanned aerial vehicles (UAVs). The safety issue of manned aircraft, such as commercial aircraft, has drawn great attentions especially after a series of disasters in recent decades. Safety and reliability issues of UAVs have also attracted significant attention due to their highly autonomous feature towards their future civilian applications. Focusing on the improvement of safety and reliability of aircraft, a fault-tolerant control (FTC) system is demanded to utilize the configured redundancy in an effective and efficient manner to increase the survivability of aircraft in the presence of faults/failures. This thesis aims to develop an effective FTC system to improve the security, reliability, and survivability of the faulty aircraft: manned aircraft and UAVs. In particular, the emphases are focused on improving the on-line fault-tolerant capability and the transient performance between faults occurrence and control re-configuration. In the existing fault-tolerant literature, several control approaches are developed to possess fault-tolerant capability in recent decades, such as sliding mode control (SMC), model reference adaptive control (MRAC), and model predictive control (MPC), just as examples. Different strategies have their specific benefits and drawbacks in addressing different aspects of fault-tolerant problems. However, there are still open problems in the fault-tolerant performance improvement, the transient behavior management, consideration of the interaction between FTC and fault detection and diagnosis (FDD), etc. For instance, MPC is recognized as a suitable inherent structure in synthesizing a FTC system due to its capability of addressing faults via solving constraints, reforming cost function, and updating model on-line. However, this on-line FTC capability introduces further challenges in terms of fault problem formulation, on-line computation, transient behavior before reconfiguration is triggered, etc. Designing an efficient FDD is also a challenge topic with respect to time response speed, accuracy, and reliability due to its interaction with a fault-tolerant controller. In the control design framework based on linear quadratic (LQ) cost function formulation, faults can be accommodated in both passive and active way. A passive FTC system is synthesized with a prescribed degree of stability LQ design technique. The state of the post-fault system is obtained through state-augmented extended Kalman filter (SAEKF), which is a combined technique with state and parameter estimation. In terms of reconfiguration capability, MPC is considered as a favorable active FTC strategy. In addition to MPC framework, the improvement of on-line computational efficiency motivates MPC to be used to perform fault-tolerant flight control. Furthermore, a Laguerre-function based MPC (LF-MPC) is presented to enhance the on-line fault-tolerant capability. The modification is based on a series of Laguerre functions to model the control trajectory with fewer parameters. In consequence, the computation load is reduced, which improves the real-time fault-tolerant capability in the framework of MPC. The FTC capability is further improved for accommodating the performance degradation during the transient period before the control reconfiguration. This approach is inspired by exponentially increasing weighting matrix used in linear quadratic regulator (LQR). Two platforms are used to perform the evaluation of the designed FTC system. A quadrotor UAV, named the Qball-X4, is utilized to test FTC designed with exponentially increasing weighing matrix LQ technique and FDD designed with SAEKF. The evaluation is conducted under the task of trajectory tracking in the presence of loss of control effectiveness (LOE) faults of actuators. The modified MPC is utilized to synthesize an active FTC system to accommodate the elevator stuck fault of a Boeing 747-100/200 benchmark model. The exponentially increasing weighing matrix LQ technique is further implemented in LF-MPC framework to improve the fault-tolerant capability before the control reconfiguration. A time delayed FDD is integrated into the evaluation process to present the effectiveness of the proposed FTC strategies. The designed FTC system is evaluated under the emergency landing task in the event of failure of elevators.