离散事件触发通讯机制下NNCS主—被动混合鲁棒容错控制研究

发布时间：2018-04-24 21:57

本文选题：非线性网络化控制系统 + 主-被动混合鲁棒容错控制　；参考：《兰州理工大学》2017年硕士论文

【摘要】：近年来,网络控制早已成为控制领域中的一个热门研究方向。然而,网络化控制系统(Networked Control System,NCS)中存在的问题是比较复杂的,如:网络诱导时延、数据丢包、模型不确定性、有限网络资源的浪费等现象的存在以及随时可能发生的各种故障,均会导致系统性能下降甚至使系统变得不稳定。除此以外,由于在实际的工业系统中或多或少都存在着非线性特性,因此,通过容错控制设计使非线性NCS(Nonlinear NCS,NNCS)具有较高的安全可靠性,具有深远的意义。现有的NNCS容错控制研究中,基本可以分为两大类:被动容错控制(Passive Fault-Tolerant Control,PFTC)和主动容错控制(Active Fault-Tolerant Control,AFTC)。尽管这两种方法有各自独特的优点,但是前者的保守性较大,无法对未知故障进行有效容错,而后者在进行控制器重组/重构时,无法保证系统在此期间的稳定性。除此以外,现存成果大多采用“时间触发通讯机制”(Periodic Time-Triggered Communication Scheme,PTTCS),数据的传输间隔等周期,这一特性无疑会造成无效数据的传输,严重浪费本来就有限的网络资源,更加加剧网络诱导时延和丢包现象的发生。基于此,本文以一类具有时变时延和外界有限能量扰动的不确定NNCS为研究对象,采用更加节约资源的“离散事件触发通讯机制”(Discrete Event-Triggered Communication Scheme,DETCS),在执行器任意失效故障情形下,从系统建模、观测器与控制器的设计与分析、DETCS与控制器协同设计等方面研究了DETCS下NNCS主-被动混合鲁棒容错控制问题。主要研究内容包括以下几个方面:1)DETCS下闭环故障NNCS建模本文引入一种与系统状态息息相关的DETCS,同时具体考虑到网络诱导时延和外界有限能量扰动的实际存在,以T-S模糊模型为基础,分别建立了标称、不确定以及不确定受扰闭环故障NNCS模型。2)NNCS主-被动混合容错控制与DETCS协同设计研究在上述所建立模型的基础上,利用并行分布补偿方案(Parallel Distributed Compensation,PDC)以及构造适当的Lyapunov-Krasovskii泛函等方法,推证出了在DETCS下使NNCS具有完整性、鲁棒容错以及鲁棒H_∞容错的时滞/事件依赖充分条件;同时以线性矩阵不等式(Linear Matrix Inequality,LMI)形式给出了主-被动混合容错控制器与DETCS协同设计的求解方法,并基于H_∞控制思想设计了故障检测观测器(Fault Detection Observer,FDO),以实现对故障的在线检测;随后设计了PFTC和AFTC,使得在发生已知故障类型时,系统能够维持自身的稳定性,在发生未知故障类型初期阶段,通过PFTC的作用能够降低系统性能下降的速度,为FDD子系统检测准确的故障信息以及AFTC的重组/重构赢得宝贵的时间,一旦得到准确的故障信息,AFTC将立即重构新的控制器用以补偿未知故障对系统性能带来的影响,进而维持系统的稳定性。3)DETCS下不确定NNCS主-被动混合鲁棒H_∞容错控制器无冲击切换研究考虑到在不同控制器之间进行切换时,由于控制信号不匹配而容易影响控制器的性能,进而影响到系统安全的问题。因此,在NNCS主-被动混合鲁棒H_∞容错控制器的设计中需要考虑重构控制器的平滑切换问题。通过引入适当的平滑切换函数,实现了切换瞬间控制器信号的匹配,减小了切换时的抖动以及跳变现象,从而保证了在执行器任意失效故障情况下系统的安全性。4)DETCS下NNCS主-被动混合容错控制仿真程序的设计与实现针对上述研究,仿真结果表明采用文中所设计的FDO可以准确地检测出故障信息;通过对AFTC、PFTC和主-被动混合容错控制器的效果比较,发现文中所设计的混合容错控制器的性能优于两者中的任何一个:即对已知/未知故障类型的故障均能实现有效容错;平滑切换函数的恰当选择实现了控制器之间的“无冲击”切换;除此之外,事件触发条件的引入大大节约了有限的网络资源,提高了网络资源的利用效率。
[Abstract]:In recent years, network control has been a hot research direction in the field of control. However, the problems in Networked Control System (NCS) are complicated, such as network induced delay, data packet loss, model uncertainty, the waste of limited network resources and so on, and may occur at any time. All kinds of faults can cause the degradation of the system performance and even make the system unstable. In addition, because of the nonlinear characteristics more or less in the actual industrial system, the design of fault-tolerant control makes the nonlinear NCS (Nonlinear NCS, NNCS) have high security reliability and far-reaching significance. The existing NNCS fault tolerance is of great significance. In control research, it can be divided into two main categories: Passive Fault-Tolerant Control (PFTC) and active fault-tolerant control (Active Fault-Tolerant Control, AFTC). Although these two methods have their own unique advantages, the former is conservatively conservative and can not effectively fault the unknown fault, and the latter is under control. In addition, most of the existing achievements are Periodic Time-Triggered Communication Scheme (PTTCS) and the interval of data transmission. This feature will undoubtedly cause the transmission of invalid data and seriously waste the originally limited network. Based on this, this paper uses a class of uncertain NNCS with time-varying delay and external limited energy disturbance as the research object, and uses the "discrete event trigger communication mechanism" (Discrete Event-Triggered Communication Scheme, DETCS), which is more resource saving, and is arbitrary in the actuator. In the case of failure failure, from the aspects of system modeling, the design and analysis of the observer and controller, the cooperative design of the DETCS and the controller, the problem of NNCS main passive hybrid robust fault-tolerant control is studied under DETCS. The main research contents include the following aspects: 1) the closed-loop fault NNCS modeling under DETCS is introduced in this paper, which is closely related to the state of the system. DETCS, at the same time, taking into account the actual existence of network induced delay and the external finite energy disturbance, based on the T-S fuzzy model, the paper establishes the nominal, uncertain and uncertain NNCS model.2 of the closed loop fault NNCS model, and the research on the NNCS master passive hybrid fault-tolerant control and the DETCS collaborative design is based on the above model. The parallel distribution compensation scheme (Parallel Distributed Compensation, PDC) and the construction of appropriate Lyapunov-Krasovskii functional methods have proved the time delay / event dependent sufficient conditions for NNCS with integrity, robust fault tolerance and robust H_ infinity fault-tolerant under DETCS; at the same time, linear matrix inequalities (Linear Matrix Inequality, LMI) form are also given. The method of solving the cooperative design of the master passive hybrid fault-tolerant controller and DETCS is given, and the fault detection observer (Fault Detection Observer, FDO) is designed based on the H_ control idea to realize the on-line detection of the fault. Then, the PFTC and AFTC are designed to enable the system to maintain its own stability when the generator is given a known fault type. In the initial stage of the unknown fault type, the speed of the system performance degradation can be reduced by the role of PFTC, which will win valuable time for the FDD subsystem to detect accurate fault information and the reconfiguration / reconfiguration of AFTC. Once the accurate fault information is obtained, AFTC will immediately restructure the new controller to compensate the unknown fault to the system. The impact that can bring, and then maintain the stability of the system.3) under DETCS, the uncertain NNCS passive hybrid robust H_ infinity robust fault-tolerant controller without impact switching is considered to affect the performance of the controller because of the mismatch of the control signals, thus affecting the security of the system. Therefore, in NNCS The smooth switching problem of the reconfigurable controller is considered in the design of the master passive hybrid robust H_ fault-tolerant controller. By introducing a suitable smooth switching function, the matching of the controller signals is realized, the jitter and the jump phenomenon are reduced, and the system is guaranteed under the arbitrary failure of the actuator. Security.4) the design and implementation of the NNCS main passive hybrid fault-tolerant control simulation program under DETCS is designed and implemented. The simulation results show that the fault information can be detected accurately by the FDO designed in this paper. The hybrid fault-tolerant controller designed in this paper is found by comparing the effect of AFTC, PFTC and the main passive hybrid fault-tolerant controller. Performance is better than any one of the two: fault tolerance for known / unknown fault types can be effectively fault-tolerant, and the proper selection of smooth switching functions realizes "no impact" switching between controllers; in addition, the introduction of event triggering conditions greatly saves the limited network resources and improves the utilization efficiency of network resources.

【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】