Adaptive Fault-tolerant Control Of Nonlinear System With Constraint Characteristics

With the increasing complexity of modern industrial systems,system components such as actuators and sensors are inevitably subject to failure,causing performance degradation and even system collapse.Therefore,the safety and reliability of the system have become an important research topic in the field of modern control.The fault-tolerant control method has provided a new effective way to solve this problem.The fault-tolerant control scheme is proposed to ensure the stability of the system and maintain a certain control performance even if some components of the system failed.Most of the actual engineering systems are nonlinear systems with high nonlinearity,uncertainty,coupling,and other characteristics,so the fault-tolerant control for nonlinear systems is not only more difficult and challenging to design,but also has important theoretical significance and practical application value.On the other hand,in the actual engineering environment,control systems always have different constraint characteristics,such as input saturation constraints,prescribed performance constraints,etc.These constraint characteristics often affect the control behavior,resulting in system performance degradation or even instability.Meanwhile,the constraint characteristics also increase the complexity of the fault-tolerant control problem and create a greater challenge for the controller design.The current research on fault-tolerant control of nonlinear systems with constrained characteristics is not sufficient and needs to be further developed.Thus,in this thesis,the adaptive fault-tolerant control methods are developed for several types of nonlinear systems with different constraint characteristics,and the main contributions are as follows.1.For a class of high-order nonlinear uncertain systems with actuator bias faults and input saturation constraints,the prescribed performance fault-tolerant control with partial unknown dynamics and unknown external disturbance is investigated.First,the neural network-based finite-time observer is proposed to approximate the system uncertainties using neural network technology.Furthermore,the prescribed performance constraint is tackled by the error transformation function,which can generate an equivalent constant-constrained tracking error.By constructing a barrier Lyapunov function,an adaptive saturated fault-tolerant controller is designed for high-order nonlinear systems using backstepping technology,which can ensure the system stability and the satisfaction of the performance constraints.2.For a class of nonlinear unknown systems with actuator faults and prescribed performance,the asymptotic fault-tolerant tracking problem under the actor-critic learning mechanism is investigated.The actor-critic learning mechanism of reinforcement learning is introduced into the design of prescribed performance fault-tolerant control.Specifically,the critic network is utilized to approximate the performance index for evaluating the steady-state of the controller system,and the action network is used to deal with the system uncertainties,which is updated by the output of the critic network and tracking errors.Moreover,a low-complexity error transformation function is designed to tackle the prescribed performance constraints,which can convert the time-varying constrained tracking error into unconstrained tracking error.Based on the error transformation,an adaptive fault-tolerant controller is proposed using the robust integral of the sign of the error method,which can achieve the asymptotic tracking control,and guarantee the prescribed performance.3.For a class of nonlinear uncertain systems with actuator faults,input saturation,and prescribed performance,the fault detection and the controller reconfiguration-based active faulttolerant problem under actuator fault are investigated.First,the extended state observer scheme is used to reconstruct the system state and external disturbance,and the residual observer error is employed to determine whether the actuator is faulty.Furthermore,based on the results of the fault detection,the nominal controller based on the state observer and the reconfigured controller based on the fault estimation are designed.After a fault is detected,the controller is reconfigured to guarantee the prescribed performance constraints,and to ensure stability and excellent tracking performance.4.For a class of nonlinear uncertain systems with sensor faults and input saturation,the active fault-tolerant and asymptotic tracking control problem with unknown dynamic parameters and external disturbance is investigated.Based on the faulty sensor state information,an online updated fault estimator is designed to obtain a real-time estimate of the fault information.Based on the obtained value,a saturated fault-tolerant controller based on the robust integral of the sign of the error method is proposed,which can effectively compensate for the effects of faults and input saturation,and guarantee the asymptotic tracking control performance with inaccurate state information.