Study On Fault Detection Method Of Discrete Networked Systems Under Communication Protocols

In the early stages,the equipment detection and maintenance mostly focus on post-maintenance,that is,examine and repair equipment when it is not working properly.This detection and maintenance method not only consumes manpower and material resources,affects the normal operation of the system,but also may cause serious accidents.Right now,accompanied by the high speed advances of computer technology and the comprehensive application of network technology,all kinds of detection devices are developing towards the direction of intelligence and automation.However,the introduction of the network makes the signal transmission inevitably affected by network-induced phenomena,such as time delay,sensor measurement loss and sensor data drift,which may cause insensitive fault detection.Therefore,constructing the networked system based on equipment and communication network to study the fault detection problem can guarantee the stable operation of facilities in the network environment.Networked systems are easy to install and maintain,which have been widely concerned in recent years for their high flexibility and reliability.Currently,networked systems have been utilized in many practical fields,such as industrial control systems,intelligent transportation systems,and remote-control systems.Because the signal transmission in networked systems shares the same communication medium,the actual network is usually equipped with the corresponding communication protocols to promote the effective use of network resources and prevent the occurrence of data collisions or communication channel congestion.Up to now,the fault detection issue of discrete networked systems subject to the communication protocols has not been fully studied and discussed,and there still exist many issues that need to be researched.Therefore,the regulation function of several communication protocols on measurement information is discussed in this thesis,and the fault detection methods are proposed for discrete networked systems under different communication protocols.This thesis intends to conduct in-depth study from the following aspects:1.The fault detection issue is investigated for nonlinear systems under the RR protocol.Firstly,a fault detection filter incorporating sensor node information is designed for discrete nonlinear uncertain networked systems.In accordance with the Lyapunov approach and matrix inequality techniques,the stability criteria of the augmented system are developed.Thus,the robust fault detection issue is handled for nonlinear systems under the RR protocol.Secondly,the fault detection problem is studied for delayed nonlinear systems with uncertain sensor measurement loss.Via designing a delay-dependent Lyapunov functional,the sufficient conditions are proposed under which the resultant augmented system is globally asymptotically stable in mean-square sense,and then the fault detection strategy is presented.Therefore,the fault detection issue is handled for delayed nonlinear systems under the RR protocol.2.The robust fault detection issue is investigated for nonlinear systems with sensor data drift under the WTOD protocol,where a second-order Markov chain is used to describe the randomly occurring faults.Moreover,the sensor data drift is characterized by a set of mutually independent random variables and their probability distributions.In order to save the network resources and decrease data conflicts,the WTOD protocol is introduced to regulate the transmission of measurement output,and then the corresponding protocol-based fault detection filtering algorithm is given.In consequence,the robust fault detection issue is solved for nonlinear systems with sensor data drift under the communication-constrained case.3.The robust fault detection issue is addressed for delayed systems with measurement outliers under the SCP scheduling.Specifically,a fault detection filter with saturation constraints is designed to achieve the purpose of filtering abnormal measurement values of the system.Via constructing a delay-dependent Lyapunov functional,the stability criteria of the augmented system are given with the aid of the generalized sector condition.Then,a new fault detection filtering algorithm is proposed for a class of delayed systems with measurement outliers under communication constraints.Besides,an unsaturated constraint fault detection filter is also established to show that the influence of outliers on the detection effect can be appropriately eliminated by designing a fault detection filter with saturation constraints.4.The fault detection problem is studied for delayed nonlinear systems with finite-time constraint under the dynamic event-triggered communication scheme.The dynamic event-triggered mechanism can dynamically adjust the trigger threshold according to the changes of the system and environment,and judge whether the measurement output is updated at the current moment,so as to reduce the data transmission.By resorting to the Lyapunov approach and matrix inequality techniques,the sufficient criterion is proposed such that the augmented system is finite-time stable with a satisfactory H∞ performance.In line with the solution to certain matrix inequalities,the expression of the desired filter parameter matrices is obtained.Consequently,the finite-time fault detection filtering algorithm is presented for delayed nonlinear systems subject to limited communication.5.The problem of finite-time robust fault detection is studied for a networked DC motor under the dynamic event-triggered mechanism.In line with the loop equation for the electrical circuit and the mechanical torque balance of a DC motor system,the discrete mathematical model for the DC motor system is derived with the aid of state augmented approach and Euler’s discretization approach.Using the network as a communication medium enables the DC motors in different positions can transmit data through the network,which facilitates the signal processing.Then,the measurement output is utilized to design the traditional fault detection filter and the memory scheduling fault detection filter,respectively.Based on the Lyapunov method,the stability criteria of the networked DC motor system are derived,and two fault detection strategies are presented so as to cope with the finite-time fault detection problem for networked DC motor subject to the communication protocol.