离散切换时滞系统的鲁棒故障检测方法研究

发布时间：2018-01-13 11:01

本文关键词：离散切换时滞系统的鲁棒故障检测方法研究　出处：《长春工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：在实际工程应用中,由于控制系统规模不断扩大,复杂程度日益提高,以及控制系统长时间处于连续的工作状态,人们迫切地需要提高系统的可靠性和安全性。另一方面,时滞和参数不确定现象普遍存在于控制系统当中,上述因素会导致系统性能下降甚至会造成系统的不稳定。当系统的执行器和传感器发生故障时,传统的控制器不能保证闭环控制系统期望的性能甚至会导致整个系统不稳定。故障检测(FD)和可靠控制技术是解决这一问题的有效方法。为了提高离散切换时滞系统的故障检测性能,本论文基于Lyapunov-Krasovskii Function(LKF)稳定性理论和鲁棒线性矩阵不等式(LMI)技术,进一步深入地研究了离散切换时滞系统的鲁棒性能、状态反馈控制器和故障检测滤波器(FDF)协同设计。由于在时滞的处理过程中,采用输入-输出方法和二项近似方法,所得结果具有更小的保守性和计算负担。全文的主要工作总结如下:(1)针对带有时滞和数据包丢失的离散切换网络控制系统,协同设计了故障检测滤波器和控制器。不同于以往文献的结果,本文设计策略为闭环故障检测策略。通过将基于观测器的故障检测滤波器作为残差生成器,网络时滞切换系统可以表示成模型匹配问题。采用LKF法和平均驻留时间方法给出了时滞依赖的充分条件,确保残差和故障之间的估计误差尽可能小,同时满足闭环网络切换系统的指数均方稳定。最后,数值仿真结果验证了所提出方法的有效性。(2)在工作(1)的基础之上,考虑了一类具有模型不确定性的离散切换时滞系统的故障检测滤波器和控制器的协同设计问题。为了提高系统的性能,使系统更加便于检测,不同于上一章的研究方法,本章引入故障加权矩阵(),限制加权频率在故障信号的频谱范围内。并应用输入-输出和二项近似方法将离散切换时滞系统转化为互联的两个子系统。设计的FDF,对所允许的数据包丢失条件,保证故障检测动态系统为输入-输出均方稳定,并且满足期望的性能。通过选取一个新的LKF泛函,获得了FDF存在的充分条件,相应的FDF增益的可解性条件利用锥补线性化迭代算法转化为凸最优问题。最后,对全文所做工作进行了总结,指出了目前切换时滞系统(网络控制系统)故障检测理论研究中存在的一些问题和进一步的发展方向,并对未来的研究工作进行了展望。
[Abstract]:In practical engineering application, the control system is in continuous working condition for a long time due to the continuous expansion of the scale of the control system and the increasing complexity of the control system. There is an urgent need to improve the reliability and security of the system. On the other hand, time-delay and parameter uncertainties are common in the control system. The above factors will lead to the deterioration of system performance and even the instability of the system, when the actuator and sensor of the system fail. The traditional controller can not guarantee the desired performance of the closed-loop control system and even lead to the instability of the whole system. And reliable control technology is an effective method to solve this problem, in order to improve the performance of fault detection for discrete switched time-delay systems. This paper is based on Lyapunov-Krasovskii function LKF) stability theory and robust linear matrix inequality (LMI) technique. Furthermore, the robust performance, state feedback controller and fault detection filter (FDF) co-design for discrete switched time-delay systems are studied. The input-output method and binomial approximation method are used. The results are less conservative and computational burden. The main work of this paper is summarized as follows: 1) for discrete switched network control systems with time delay and packet loss. The fault detection filter and controller are designed in cooperation. Different from the results of previous literatures, the closed-loop fault detection strategy is designed in this paper. The observer based fault detection filter is used as the residual generator. The LKF method and the average dwell time method are used to give the sufficient conditions of time-delay dependence to ensure that the estimation error between the residual error and the fault is as small as possible. At the same time, it satisfies the exponential mean square stability of closed loop network switching system. Finally, the numerical simulation results verify the effectiveness of the proposed method. The cooperative design of fault detection filters and controllers for a class of discrete switched time-delay systems with model uncertainty is considered in order to improve the performance of the system and make the system easier to detect. Different from the research method in the previous chapter, the fault weighting matrix is introduced in this chapter. The weighted frequency is limited in the spectrum range of the fault signal, and the input-output and binomial approximation method is applied to transform the discrete switched time-delay system into two interconnected subsystems, the designed FDF. For the allowed packet loss conditions, the dynamic fault detection system is guaranteed to be input-output mean square stability and meet the desired performance. A new LKF functional is selected. Sufficient conditions for the existence of FDF are obtained, and the solvability condition of the corresponding FDF gain is transformed into a convex optimal problem by using the cone complement linearization iteration algorithm. Finally, the work done in this paper is summarized. This paper points out some problems existing in the research of fault detection theory for switched time-delay systems (NCS) and its further development direction. The future research work is also prospected.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP13

【参考文献】