A Study On Nonlinear Self-Repairing Control For An Aircraft

NASA brought forth a research project of self-repairing flight control system (Self-Repairing Flight Control System) by United States in 1984, so the self-repairing control has became a focus of research in aerospace fields. Self-repairing flight control system is developed on the basis of fly-by-wire flight control system, and it is an advances flight control system combined with active fault-tolerant technology. SRFCS turns into an important aspect in future. The developments of fault diagnosis and fault-tolerance control technology provide an effective way of improving safety and reliability for complicated flight control system. The basic idea is following: under the premise of without increasing the hardware devices, SRFCS can make full use of the capabilities of inherent function redundancy and improve the reliability, survivability and maintainability of flight control systems, at the same time reducing the maintenance costs for flight control systems. Fault detection and isolation is one of core technology for self-repairing flight control system, which can determine faults cause and location in time, and also can predict a possiblity of potential failure, and then provide a relevant information for alarm and reconfigurable flight control law and self-diagnostic maintenance. The reconfigurable method for flight control law is another core technology in self-repairing control system. In the emergence of the larger fault, a control law is reconstructed, and can ensure a safety of the flight after failure, thereby increas the flight control system reliability and maintainability, reduce maintenance costs, while reducing the burden on pilot.The main works of this thesis are shown as follows:Firstly, in this thesis, according to some basic concepts in SRFCS, the relationship between all of control surfaces for aircraft with five control surfaces single engine is analyzed, by combining with the basic assumptions for a self-repairing flight control system, an architecture of self-repairing controller is proposed, which is implemented in a retrofit fashion, i.e. a self-repairing controller consists of a basic controller and a reconfigurable control vector. The baseline controller is retained, and the reconfigurable controller is implemented as an add-on signal. Based on the controller structure, a new description for actuator/control surface failure is also proposed, then a problem of design the self-repairing control law is transformed into the solution of a reconfigurable control vector. The design process of self-repairing control law is simplified.Secondly, a speed, flight path angle and heading of an aircraft are viewed as a generalized velocity vector, an adaptive tracking control law based on backstepping technology is designed, which can guarantee the generalized velocity and sideslip angle tracking desired value of velocity and sideslip angle with fault-free aircraft. The backstepping approach, while suffering from the problem of "explosion of terms"guarantees boundedness of tracking errors globally. In this paper, an version of adaptive tracking control law based on backstepping combined with dynamic surface is derived, and the control law can guarantee closed-loop system with input-state stability properties, and strict proof is given. Based model of a fighter, two super maneuver numerical simulations is carried out, and the results show that the control law can guarantee aircraft without fault to has a good performance and to meet one of the basic assumptions for a self-repairing flight control. It also provides a basis for solving reconfigurable control vector.Thirdly, based on the output matching conditions, the analytical solution of a reconfigurable control vector is derived, its structure is a linear combination of the basic control law and coefficient in connection with failure, and the coefficient is a function of fault parameters, then a problem of solution of repair control vector is transformed into a calculation of fault parameters. For for the calculation of fault parameters, two different methods for solving parameters are proposed, namely, the parameter estimation based on adaptive thinking and based on fault diagnosis.Fourthly, according to the thought of adaptive control, on the basis of theσ, d parameters of the model for the actuator/control surface stuck and efficiency loss, the adaptive law of parameters for reconfigurable control vector is derived, and the closed-loop characteristics of the restored system is also given. Few parameters will to be estimated and the structure of reconfigurable control vector is simple in this method. The simulations of the stuck 10°for left aileron and the efficiency loss 90% for right elevator failure, show that the algorithm is validity. Since adaptive control law is working in the whole of process, the controller also can improve the tracking performance for overall system.Finally, adaptive control law based on adaptive thought can not guarantee that fault parameters estimated value converge to the true value. Taking into account the dynamic characteristics of typical second-order for aircraft actuator, the dynamic of actuator will be introduced into the nonlinear aircraft dynamic equations, by using the actuator/control surface of the single (?) parameter model, a on-line fault parameter identification method based on high-gain observer is designed, which accomplish the purpose of estimating actuator states wile simultaneously identifying fault parameters. Then a reconfigurable control vector is implemented by using the estimated fault parameters and actuator states, and compensates the influence of actuator faults, and can guarantee the tracking performance for fault system. Nonlinear model of an aircraft simulation results, in two scenarios of the stuck 5°for right aileron and the efficiency loss 90% for left elevator failure, show that fault parameter estimated value can quickly converge to its true value and the reconfigurable vector can also fast restore fault system performance to the levels of fault-free system. The repairing process is relatively stable, and on-line identification of fault parameter is accuracy. The proposed algorithm is effective.