Research On Joint Faults And Disturbances Estimation For Networked Systems

Networked Control Systems(NCSs)is a typical spatially distributed system.Its sensors,actuators and controllers,as well as other system components,are equipped with network interfaces and communicate through a shared digital communication network.It has unique advantages such as simple installation and maintenance,low cost,reduced system wiring and flexible communication architecture.However,unexpected faults or failures in engineering practice may induce performance degradation and even catastrophic accidents.Consequently,it is paramount to detect and identify faults so that some necessary protective actions can be taken in the incipient stage of the occurrence.Generally speaking,a typical fault diagnosis process consists of fault detection,fault isolation and fault estimation.System faults and interference are the main reasons that cause system performance degradation.If fault and disturbance information can be effectively estimated,active fault-tolerant control algorithms can be designed specifically to improve the fault-tolerant capability and safety of the system.The introduction of the control system complicates the analysis and synthesis process.Under this background,the joint estimation of fault and disturbance in networked control system is mainly studied in this paper.The main research contents are as follows:The joint fault and disturbance estimation problem is investigated for a class of networked control systems with random packet losses and input/output disturbances.By introducing an intermediate variable,a novel intermediate estimator is proposed to estimate the states,faults and external disturbances simultaneously,which can be used for the systems in which the strictly positive real condition and the observer matching condition are not satisfied.It is proved that the states of the error system are uniformly ultimately bounded.A simulation has been carried out on a continuous-stirred tank reactor is exploited to demonstrate the effectiveness of the proposed fault estimation approach.The joint fault and disturbances estimation problem for a class of networked systems with random packet losses and sensor saturation is studied.The output measurements are affected by sensor saturation which is described by sector-nonlinearities.The missing measurement is modeled as a linear random variable satisfying the Bernoulli distribution.It is proved that the augmented states of the error systems are uniformly ultimately bounded by employing Lyapunov stability theory and linear matrix inequality technique.Finally,an illustrative example about two-stage chemical reactor systems is exploited to demonstrate the effectiveness of the proposed joint estimation approach.The fault estimation problem is investigated for a class of discrete networked control systems with deception attack and input/output disturbances.We model the malicious deception signals as unknown-but-bounded signals in such a way that no prior knowledge of the attack signals is required.By introducing an intermediate variable,a new specially designed security-guaranteed intermediated estimator is proposed to estimate the faults and disturbances in the presence of attacks.It is proved that the states of the error systems are uniformly ultimately bounded.A simulation has been carried out on three-capacity water tank level control system is exploited to demonstrate the effectiveness of the proposed fault estimation approach.