Research On Fault-Tolerant Control Of Nonlinear Networked Control Systems

Compared with conventional point-to-point control systems,networked control systems(NCSs)have many advantages such as lower install costs,less wiring,ease of system maintenance,etc.Due to these benefits,NCSs have found applications in many areas and have been a hot research topic in the control society.Due to the complexity incurred by network,the growing demand for safety and reliability in NCSs has motivated significant research in fault diagnosis(FD)and fault-tolerant control(FTC)of NCSs,leading to a rich body of literature.Nevertheless,there are some limitations in the existing results.Specifically,most existing literatures rely on passive FTC theory and time-triggered communication method,and have failed to take into account the presence of nonlinear uncertainty,nonlinear and intermittent fault,packet disordering and medium-access constraint.This dissertation studies the FTC problems for nonlinear NCSs with nonlinear uncertainty,nonlinear and intermittent fault in the presence of delay,packet dropout,packet disordering and medium-access constraint based on time-triggering or event-triggering method by utilizing adaptive approximation and adaptive techniques.The main contents of this dissertation are as follows:An active FTC approach is developed for nonlinear NCSs subject to delay,nonlinear uncertainty and nonlinear fault by utilizing adaptive approximation and adaptive techniques.When the system is fault-free,a robust controller ensures system stability in the presence of delay and nonlinear uncertainty.In the presence of a fault,the adaptive approximation and adaptive-based fault diagnosis component is used to estimate and isolate the fault.Then the controller is reconfigured to ensure system stability by successively using the estimate and isolation result.The proposed active FTC method ensures the state tracking error asymptotically converges to zero.An active FTC approach is developed for nonlinear NCSs subject to delay,packet dropout,packet disordering,nonlinear uncertainty,nonlinear plant and sensor faults,and intermittent actuator faults by utilizing adaptive approximation and adaptive techniques.By considering Markovian delay and packet dropout,a new packet reordering method is proposed to cope with packet disordering.When the system is fault-free,a robust controller ensures system stability in the presence of delay,packet dropout,packet disordering and nonlinear uncertainty.In the presence of faults,the adaptive approximation and adaptive-based fault diagnosis component is used to estimate and isolate the faults.Then the controller is reconfigured to ensure system stability by successively using the estimates and isolation result.The proposed active FTC method ensures the boundedness in probability of the state tracking error.An event-triggered active FTC approach is developed for nonlinear NCSs subject to delay,packet disordering,medium-access constraint and nonlinear fault by utilizing adaptive approximation and adaptive techniques.The co-design of event-triggered communication,FTC and scheduling is achieved.A novel event-triggering scheme with an adjustable triggering condition and adaptive triggering thresholds is proposed.The stochastic event-driven actuator scheduling method is improved.A new packet reordering approach is used to cope with packet disordering.When the system is fault-free,an event-triggered robust controller ensures system stability in the presence of delay,packet disordering and medium-access constraint while reducing communication load.In the presence of a fault,the adaptive approximation-based estimator is used to estimate the fault.Then the controller is reconfigured to ensure system stability by utilizing the estimate.The proposed event-triggered active FTC method ensures the boundedness in probability of the state tracking error while reducing the network usage.An event-triggered active FTC approach is developed for nonlinear NCSs subject to medium-access constraint,intermittent sensor and actuator faults by utilizing adaptive technique.The co-design of event-triggered communication,FTC and scheduling is achieved.The stochastic event-driven sensor and actuator scheduling is investigated.The health statuses of the sensors and actuators are governed by two independent Markovian processes.By taking into account the stochastic sensor faults and stochastic event-driven sensor scheduling,a novel event-triggering scheme is proposed.An adaptive controller is designed by using adaptive technique.The proposed event-triggered active FTC method ensures the boundedness in probability of the state tracking error while reducing the network usage.A decentralized active FTC approach is developed for nonlinear distributed NCSs subject to delay,packet dropout,nonlinear uncertainties and nonlinear faults by utilizing adaptive approximation and adaptive techniques.The adaptive bounding technique is used to cope with delay,and the packet loss process in each link is described by an independent Markovian process.When the system is fault-free,the decentralized robust controllers ensure stability of the subsystems in the presence of delay,packet dropout and nonlinear uncertainties.In the presence of faults,the controllers are reconfigured to ensure control performance.Meanwhile,the adaptive approximation-based fault diagnosis components are used for fault isolation.Then the controllers are reconfigured again to enhance control performance by utilizing isolation results.The proposed active FTC method ensures the boundedness in probability of all state tracking errors.Finally,the conclusion and future work are presented.