Fault Diagnosis And Fault Tolerant Control Based On Communication Schedulings

With the development of control science,computer and communication technology,the structure of control system is more and more complex and the scale is larger and larger.It has been a development trend that the closed-loop system is constructed by using data communication network.In order to avoid congestion and improve the accuracy of information transmission,it is of great theoretical and practical significance that the communication scheduling strategy is introduced to allocate the limited network resources reasonably.However,up to now,little attention is focused on the fault diagnosis and fault-tolerant control problem under the communication scheduling schemes.In this paper,fault diagnosis and fault-tolerant control problems are systematically investigated for some kinds of networked systems subject to communication scheduling schemes.The main content of this paper can be divided into three aspects.The first part mainly discusses the problem of fault detection based on network-induced phenomena and communication scheduling protocols.By means of Euler's lemma,Lyapunov stability theory and linear matrix inequality technique,some sufficient conditions are established to satisfy the predetermined performance index in finite frequency domain,and the desired fault detection filters are obtained.In addition,by using ellipsoidal estimation technology and convex optimization theory,a distributed fault detection mechanism is designed based on set-membership filtering method.In the second part,the problem of fault estimation subject to communication scheduling schemes is investigated for discrete time-varying system.Based on the Riccati difference equation technique,the desired fault estimators are designed.Furthermore,in the third part,the finite-time fault tolerant control problem is studied with communication protocol.By resorting to Lyapunov stability theory and the stochastic analysis technique,an observer-based fault-tolerant controller is designed and the performance of system is ensured.Specifically,the main researches and contributions of this paper can be summarized as follows:First,the problem of fault detection in finite frequency domain is discussed for a class of discrete-time time-delayed networked systems subject to Round-Robin protocol.In order to prevent the data from collisions,this protocol is a periodic one to schedule the information transmission via a shared communication network between the sensor and the filter.Based on the version of generalized Kalman-Yakubovich-Popov(GKYP)lemma and Lyapunov stability theory,some sufficient conditions are obtained to ensure fault sensitivity and disturbance attenuation level.According to such a framework,the event-triggered distributed fault detection problem is investigated for a class of discrete-time uncertain systems in the finite frequency domain.An event-triggered protocol is introduced with hope to reduce the communication burden and the energy consumption.A distributed fault detection filter is presented based on the information of itself and neighbor node.By using Euler's formula and Lyapunov stability theory,some sufficient conditions are derived to ensure the required performance index and the gain of distributed fault detection filters.Second,the problem of the distributed fault detection is discussed under the Weighted Try-Once-Discard protocol.The Weighted Try-Once-Discard scheduling is considered with hope to avoid congestion and improve the accuracy of information transmission.A novel fault detection method is proposed to determine whether there exist the unknown but bounded faults in the discrete time-varying networked control system.To this end,two ellipsoids are obtained which include prediction ellipsoid sets and measurement updated ellipsoid sets of all possible real states.Then,a fault is detected when there is no intersection between the prediction set and the estimation updated set.Two recursive optimization problems are developed to look for the minimal ellipsoids in order to improve the fault detection performance.Third,the finite-horizon fault estimation issue is analyzed for discrete time-varying systems subject to torus event-triggered protocol.The torus event-triggered protocol is introduced in order to relieve the communication burden.Besides,the multiple fading measurements are considered which is described by an individual stochastic variable meeting a certain probability distribution.By utilizing the stochastic analysis theory and completing squares technique,sufficient conditions are obtained to achieve the finite-horizon H performance constraint.Then,the desired estimator gains are calculated by solving two backward recursive Riccati difference equations.According to such a framework,the Round-Robin protocol-based finite-horizon fault estimation problem is investigated for the discrete time-varying systems subject to uniform quantization and randomly occurring faults.In order to reduce the data conflict,the Robin-Robin is introduced to schedule the network resource.Moreover,the error from the uniform quantization is described as a series of uniform distributed noises.For the discrete time-varying system with the Round-Robin protocol,some sufficient conditions are established to achieve the H?performance index.Furthermore,the desired estimator gain is obtained by solving two backward recursive Riccati difference equations.Finally,the finite-time fault tolerant control problem is investigated for a class of the discrete-time stochastic parameter systems subject to stochastic communication protocol and censored measurements.A stochastic communication protocol governed by a Markov chain is employed to determine which actuator has the right to access the network at each transmission instant.Besides,the censored measurements are described by the Tobit I model.An improved performance index dependent on the predetermined censored threshold is constructed to evaluate the disturbance rejection level of the fault tolerant controller in the simultaneous presence of both external disturbances and censoring effects.The purpose of this chapter is to design a fault-tolerant controller such that the closed-loop system satisfies both the stochastically finite-time boundedness and H? performance requirements.In light of the Lyapunov theory and stochastic analysis technique,some sufficient conditions are derived skillfully to guarantee the performance index,and the desired controller gains are calculated by solving a set of linear matrix inequalities.