【作者】 吴军生；

【导师】 田作华；

【作者基本信息】 上海交通大学 ， 控制理论与控制工程， 2009， 博士

【摘要】 容错控制技术是提高动态系统安全性、可靠性的一种极为重要的途径,因而深入研究容错控制技术,不但具有重要的理论意义,而且也具有巨大的实际应用价值。目前关于容错控制技术的研究引起了广泛关注,已经成为控制界的热门研究课题之一。在各类工业系统中,时滞现象是极其普遍的,如长管道进料或皮带传输、极缓慢的过程或复杂的在线分析仪等均存在时滞现象,这都可归结为时滞系统模型。时滞的存在使得系统的分析和综合变得更加复杂和困难,同时时滞的存在也往往是系统不稳定和系统性能变差的根源。正是在这一背景下,本文系统地研究了时滞系统的容错控制方法,主要内容及研究成果如下:研究了带有执行器故障的变时滞不确定系统的鲁棒可靠控制。给出了系统在正常运行和带有可预期执行器故障情况下,闭环系统保持二次稳定的时滞依赖充分条件。运用线性矩阵不等式方法给出了可靠状态反馈控制的综合方法。该方法相对于时滞独立条件下得到的结果,具有较小的保守性。研究了线性状态时滞不确定系统部分传感器和执行器同时失效的容错控制问题。给出了带有时滞的状态反馈控制的设计方法。采用该方法设计的容错控制系统能够在保证系统闭环稳定的同时满足给定的性能指标,所设计的状态反馈控制器含有时滞项,可以减小时滞对闭环系统带来的影响。问题解存在的条件以线性矩阵不等式的形式给出。H∞研究了多状态时滞系统的主动容错控制问题。基于广义内模控制,给出了两种容错控制的设计方法。第一种方法基于Youla控制器参数化,将闭环系统控制器设计纳入广义内模控制的框架;设计得到的控制器由两部分构成,分别针对系统的性能和鲁棒性。如果系统没有发生故障,只有针对性能设计的控制器起作用;而当系统发生故障,针对鲁棒性设计的控制器才会起作用。第二种方法在广义内模控制的结构下,利用互质分解方法设计了故障残差产生器,通过将故障诊断问题转化为H∞控制问题,从而引入了一类新的故障诊断性能指标;据此性能指标,给出了残差后滤波器的设计方法,该后滤波器可以实现故障信号的渐近估计。在故障估计的基础上,利用故障补偿实现了闭环系统的容错功能。该方法可视为故障检测与容错控制之间的桥梁,给出了故障残差与真实故障之间的对应关系。通过仿真研究验证了两种方法的有效性,并基于仿真结果对两种方法进行了对比分析。研究了带有不确定性的时变时滞系统的容错控制。基于滑模控制提出了一种容错控制设计方法。线性滑动平面存在的时滞依赖充分条件以线性矩阵不等式(LMI)可行性问题的形式给出,该滑动平面如果存在,即可以保证由原系统得到的降阶等价系统的轨迹在有限时间内收敛于该平面。基于这个滑动平面,设计了一种可以作为容错控制的到达控制。该控制可使得带有模型不确定性、扰动输入和执行器故障的闭环系统保持二次稳定,具有容错功能。研究了线性变参数时滞系统的主动容错控制问题。在参数的凸条件约束和系统参数矩阵的仿射依赖条件假设下,利用线性分式变换,将故障滤波器设计问题转化为H∞控制器设计问题,给出了该变参数控制器的存在条件和增益求解方法,并给出了故障补偿的设计方法。将故障诊断与容错控制设计问题变为H∞控制器设计问题的思想可以推广到其他类型的系统。采用一种基于Newton-Raphson算法的渐进式容错控制设计方法研究了带有参数故障的时滞系统的容错控制问题。该方法可明显降低由于控制增益计算和实施所带来的延迟而造成的性能损失和系统失稳的危险。该方法考虑非理想状况,即容错控制的实施不是在故障诊断结束之后立刻进行,因而实用性更强。更多还原

【Abstract】 Fault-tolerant control is an important approach to improving the safety and reliability for dynamic systems. Hence, the study on fault-tolerant control technology has both theoretical and practical importances. At present, it has drawn wide attention, and has been one of the main topics in control science. Time delay is commonly encountered in various engineering systems, such as chemical processes, long transmission lines in pneumatic, hydraulic and rolling mill systems, furthermore, these can be regarded as time delay systems. Time delay usually results in unsatisfactory performances and is frequently a source of instability. The methods of fault-tolerant control for time delay systems are studied systematically in this dissertation. The main contributions of this dissertation can be summarized as follows:Robust reliable control design problem for uncertain systems with time-varying state delay as well as actuator failures in the input channels is discussed. A sufficient condition is established such that the quadratic stability of the closed-loop system is ensured for both the nominal system and the system with possible actuator failures. The condition is a delay-dependent one. A reliable state feedback controller is synthesized within the framework of linear matrix inequalities(LMIs). As compared with the delay-independent conditions, the result proposed has the merit of being less conservative.The fault tolerant control design problem for uncertain time-delay systems with both sensor and actuator failures is discussed. A memoryless and memory state feedback control, which can decrease the influence of time delay, is designed to guarantee the integrity of the system together with the prescribed H_∞performance index. The sufficient condition for the existence of the solution is formulated in terms of LMIs.Active fault-tolerant control problem of multi-state delay system is studied. Based on generalized internal model control (GIMC) two fault-tolerant control design methods are proposed. In the first method, based on Youla controller parameterization and GIMC, the feedback controller architecture includes two parts: one part for performance and the other part for robustness. The feedback control system will be solely controlled by the performance controller when there is no fault and the robust controller will be active when there are faults. In the second method, based on GIMC, firstly, the residual signal is generated by means of a co-prime factorization method. Secondly, a new performance index is proposed by converting the fault diagnosis (FD) problem to a H∞robust control problem. And then a post-filter is designed as a robust controller which can make the filtered residual signal close to the fault signal so that the fault estimation purpose can be fulfilled simply. This method may be regarded as a H∞bridge between FD and fault tolerant control, the relationship of residual signal and the real fault signal is established. Some illustrative design examples are used to demonstrate the validity and applicability of the proposed approaches. Some comparisons are made based on the simulation results.Fault tolerant control for uncertain systems with time varying state-delay is studied. Based on sliding mode controller design, a fault tolerant control method is proposed. By means of the feasibility of some LMIs, delay dependent sufficient condition is derived for the existence of a linear sliding surface which guarantees quadratic stability of the reduced-order equivalent system restricted to the sliding surface. A reaching motion controller, which can be seen as a fault tolerant controller and can retain the stability of the closed loop system in the present of uncertainties, disturbances and actuator fault, is designed. A numerical simulation shows the effectiveness of the approach.An active fault-tolerant control scheme of linear parameter varying time-delay system (LPVTD) is proposed. Under the assumption that the parameters are convex and the system matrices can be of affine parameter dependence, by a linear fractional transformation (LFT), the design of a fault filter can be equal to the design of a H∞controller. The existence condition and the method to get the gain of the controller are proposed, based on which fault compensation can be realized. The idea that transform the fault diagnosis and fault tolerant control problem to a H∞control problem can be easily extended to other kinds of systems.The fault accommodation problem for time-delay systems with parametric faults is studied. The progressive accommodation strategy which based on the Newton-Raphson scheme is adopted to solve this problem. This accommodation scheme significantly reduces the degradation of performance and the risk associated with system instability resulted from the time delay induced by fault accommodation algorithms to provide a solution. Consider that in the real situations, it takes some time for the controller to compute and apply the fault tolerant control after the fault been detected. So the delays should be considered and thus the method proposed is more applicable in practice.更多还原