Research On Fault-tolerant Control Of Nonlinear System

As an effective way of improving the security of the system,the strategy of fault diagnosis and fault-tolerant control has achieved much attentions in control community.In practical applications,the controlled systems are generally nonlinear,potentially suggesting the significance of the investigation for fault-tolerant control in the nonlinear case.In this work,the problem of nonlinear fault-tolerant control is considered and two control strategies are proposed.To begin with,for the nonlinear system with un-modelled uncertainties and external disturbances,this work presents the fault-tolerant control strategy through fault estimator.The core idea is to utilize this estimator to estimate the system state and the fault signal,based on which the state feedback control law can be designed to achieve the fault-tolerant control for the system to reject the actuator fault.Furthermore,some guidelines for the choices of control parameters are also provided to guarantee the convergence properties of dynamics both of the system state and the estimation error.Next,in order to further cope with the system with faults such as actuator fault,sensor fault and internal fault of the system,a strategy based on the techniques of neural network and active disturbance rejection control is proposed.The main principle is to treat the faults and system uncertainties as the total disturbances,which will be estimated and eliminated through the extended state observer and state feedback controller respectively,such that the performance of the system can be guaranteed even with the existence of the faults.Specifically,for the sensor faults,we apply the radial basis function neural network estimator to predict the true value of the system output,to replace the sensor output with fault,such that this sensor fault can be isolated.For the further illustrate the effectiveness of the proposed fault-tolerant control strategies,these two are verified in the numerical example of the three-tank system with the faults including actuator fault,sensor fault and the leakage fault.According to the simulation results,the tracking performance of this three-tank system is satisfactory which suggests the rejection effect for the faults.