摘要随着计算机技术，网络技术的迅猛发展和控制系统规模的日益扩大，将通讯 网络应用到复杂控制系统中，已经成为自动化领域的发展趋势。这种通过共享网 络构成闭环反馈系统被称为网络控制系统。结合网络技术和控制技术的网络控制 系统，既是计算机和网络技术向控制领域的延伸和发展，又体现了控制系统向网 络化、分布化、智能化的发展趋势。同时网络带来的时延和丢包等问题，使系统 的故障检测设计分析变得复杂困难，也使得其在工程应用中对安全性、可靠性的 要求更高，因此对网络控制系统的故障检测具有重要的理论和实际意义。84389

本文研究了离散时间网络控制系统在随机不确定长时滞情况下和在随机数 据丢包情况下的故障检测问题。其中，在随机不确定长时滞的网络系统中，即面 向网络传输时滞大于一个采样周期并且时滞是不确定的网络控制系统，可以将系 统构造为 Markov 跳变系统，基于跳变系统再构造状态观测器，运用 Lyapunov 函数分析残差系统稳定及运用残差信号来判定系统故障的产生，运用 Matlab 仿 真验证故障检测方法的有效性；针对一类随机数据丢包的网络控制系统的故障检 测问题，将包含随机丢包的网络控制系统建模为参数具有伯努利特性的模型，设 计状态观测器进行故障检测，借助 Matlab 中的 LMI 工具箱，仿真验证方法的实 用性和可行性。

毕业论文关键词：网络控制系统；故障检测；状态观测器；随机长时滞；随机数据丢 包

Abstract With the rapid development of computer technology, network technology, and the increased scale of control systems, it is a trend to apply communication networks into complex control systems。 Control systems based on the sharing of network is known as network control systems。 Network control systems are the combination of network technology and control technology, they embody the characteristics of network and control, and they demonstrate also the trend of network, distribution, and intelligence。 Network delays and packet dropouts are inevitable, which complicates the fault detection。 The occurrence of faults impose higher requirements on the safety and reliability of control systems。 Thus, it is important to study the fault detection of networked control systems。

This paper studies the fault detection of discrete-time networked control systems with stochastic uncertain long time delays and random data packet dropouts。 For networked control systems with stochastic uncertain long time delays,  which are larger than the length of a sampling period, a Markov jump system is constructed, and then a state observer is established。 The Lyapunov function approach is used  to analyze the stability of residual systems and determine the occurrence of faults。 The simulation results illustrate the effectiveness of the proposed fault detection scheme。 For network control systems subject to random data packet dropouts, the systems are modeled as Bernoulli models, and state observers are designed to carry out the work of fault detection。 By using the LMI toolbox in Matlab, the simulation results verify the feasibility of the proposed methods。

Key words: networked control systems; fault detection; state observer; random long delays; random data packet dropouts

目录

第一章 绪论 1

1。1 选题的背景和意义 1

1。2 网络控制系统的发展过程 2

1。3 网络控制系统的基本问题 4

1。3。1 网络诱导时延 4

1。3。2 数据包丢失 4

1。3。3 网络调度和其他