Fault Detection For Networked Systems With Packet Dropout Using Delta Operator

With the rapid development of computer information technology and the cross-penetration of network communication technology and control technology, in the complex control systems, communication networks are often used to connect controlled objects, sensors, controllers, filters, actuator together to form the networked control systems(NCSs). NCSs have many advantages such as lower cost,higher reliability, simpler installation and maintenance and more flexible structure,etc.However, due to the introduction of the network with limited-bandwidth, the nodes compete for resources and there will be the induced delay, packet dropout,quantization error,network congestion in the control loop network. On the other hand,with the increasing complexity of NCSs, it may bring about casualty and property losses in the event of failure. Therefore, it is significant to study the fault detection(FD) for NCSs.While most of these studies focused on the continuous-time systems and discrete-time systems respectively and the discretization methods are mostly traditional shift operator. However, when sampling is fast, shift-operator-based designs may exhibit the problems of crowding poles located in the boundary of stability circle which make the system unstable. This thesis introduces the Delta operator method into the FD for NCSs and several different NCSs models are established. The FD for NCSs with induced time delays and packet dropouts are investigated and the major contributions of the thesis are as follows:First, focusing on the phenomenon of data packet dropout, the problems of fault detection for networked control systems(NCSs) via Delta operator are addressed. The packet dropouts of NCSs are modeled as a stochastic variable satisfying Bernoulli distribution and the fixed short-delays are considered. By using Delta operator approach, a new model is constructed to depict the behavior of the networked control systems. Based on the Delta operator system, an observer-based fault detection filter is given to detect faults. By using Lyapunov stability theory and linear matrix inequality(LMI),the stability and H∞ performance of the Delta operator system are deduced.Then the network control system with uncertain time delay and packet dropouts are further discussed. The packet dropouts are still modeled as Bernoulli random variable, and the uncertain time-delay are convert to the parameter uncertainty of the system. Based on the established Delta operator model, the state observer-based fault detection filter is designed. The sufficient conditions of the asymptotic stability and H∞ performance of the system are deduced by Lyapunov stability theory and LMI method. By using LMI toolbox, the designed observer are solved. A numerical example is used to illustrate the effectiveness of the proposed method.