Research On Robust Fault Detection Technique For Networked Control Systems

Networked Control Systems(NCSs)are closed loop control systems,where the information of control system components(sensors,controllers,actuators,etc.)is exchanged via communication networks.Compared with conventional control systems,advantages of NCSs include: reduced system wiring,ease of system maintenance,low cost,increased system agility.However,some problems such as the bandwidth?limited digital communication networks or the occurrence of network?introduced uncertainty will increase the probability of system malfunction.The safety and reliability are extremely important for the systems.Fault detection technology provides a new way to improve the reliability of systems.This thesis focus on continuous?time and discrete?time networked control systems.A full order fault filter and observer?based fault detection scheme are introduced to generate the residual signal.New models for continuous?time or discrete?time with faults are established correspondingly,and design the controller and fault detection filter at the same time,because it will improve the detectability of the fault and the reliability of the fault detection.The designed residue generator could ensure the sensitivity of the residual to fault and simultaneously guarantee the robustness of residual signal to the disturbance signal,and can also detect the occurrence of faults in time.At last,some examples are given to illustrate the approaches are feasible.The fault detection of NCSs is studied in this thesis.The main research works are concluded as follows:1.According for the problem of networked control systems with time?varying network?induced delays,packet dropouts and communication constraints,this chapter studies the NCSs fault detection under consideration of the communication networks exit in sensor?to?controller channel or in controller?to?actuator channel.The observer?based fault detection filters are adopted to generate residual signal,then the fault detection models are established correspondingly.By defining an appropriate Lyapunov function and adopting the convex analysis method,the new fault detection filter design criteria are obtained.The simulation results illustrate the effectiveness of the proposed design method.2.The problems of observer?based fault detection filter and H_?controller design for a continuous?time NCSs under both consideration of sensor?to?controller and controller?to?actuator network?induced delays and packet dropouts have been studied.The linear estimation based artificial delay compensation method has been introduced,and the non?uniform distribution characteristic of network?induced delays and packet dropouts has been considered to establish new model for a continuous?time residual system.At last,the sufficient conditions for asymptotically stability in the sense of mean?square of the systems are given.3.According for the problem of networked control systems under consideration of random packet dropout and communication constrains.By considering communication constrains exist in controller?to?actuator channel and random packet dropout,which is modeled as Bernoulli distribution,exists in sensor?to?controller channel,a new model for discrete?time NCSs with faults is established.Based on the established model,a full order fault detection filter is introduced to generate the residual signals,and the residual system is obtained.What is more,the sufficient conditions are proposed to stochastically stabilize the residual system.Finally,the results of the thesis are summarized and future research topics for NCSs fault detection are pointed out.