Fault Tolerant Control For Several Classes Of Intermittent Faults

Because practical systems such as production, transportation and communications become increasingly large-scale and complicated, security issue has attracted common attention of the whole society. Fault-tolerant control is a kind of control theories and methods, which could automatically accommodate failures in system and maintain the system with acceptable performance such as stability and reliability. Over the last four decades, fault-tolerant control has drawn extensive attention from both industrial and academic communities, and it has achieved substantial progresses.Faults are inevitable in actual systems. It can induce unexpected behaviors of the system, endanger system stability, and thus become the focus of fault-tolerant control. In terms of the lasting duration, faults can be divided into permanent faults and intermittent faults. Currently, the most faults considered in fault-tolerant control research and application are permanent faults, and only a few publications investigate the fault-tolerant control for intermittent faults. However, in practical systems, there widely exists intermittent faults, which have serious effects on performance and safety of the system and cannot be handled by the existing fault-tolerant control methods designed for permanent faults. Therefore, this dissertation systematically studies the fault-tolerant control problem for intermittent faults. The considered faults in the dissertation cover intermittent faults that not only act in different function forms (additive and multiplicative) but also occur on different parts of the system (sensors and actuators). The main contributions of the dissertation are as follows:1. Chapter 2 studies the fault-tolerant control problem for a class of linear systems subject to additive intermittent faults. Three fault scenarios, only-sensor-fault, only-actuator-fault, and both-sensor-and-actuator-fault, are considered respectively. The additive intermittent faults are described by random variables obeying the Bernoulli distribution. The fault-tolerant performance is evaluated by a given H∞ performance index. The sufficient condition for the existence of fault-tolerant controller can be expressed as a linear matrix inequality (LMI). Through solving this LMI, the corresponding dynamic output feedback controller can be obtained. The simulation results demonstrate the effectiveness of the proposed method.2. Chapter 3 investigates the problem of fault-tolerant control for a class of nonlinear systems with multiple additive intermittent faults. Similarly to the chapter 2, the above-mentioned three fault scenarios are considered. Furthermore, this part takes nonlinear dynamics and modelling uncertainties into account. The multiple additive intermittent faults are described by a set of independent random variables which obey Bernoulli distribution. Similarly to chapter 2, H∞ index is employed to evaluate the performance of the designed fault-tolerant control system. The sufficient condition for the existence of fault-tolerant controller is also expressed as a LMI, and then the corresponding dynamic output feedback controller can be obtained. The simulation on an aircraft engine system illustrates the effectiveness of the proposed method.3. The forth chapter of the dissertation studies the fault-tolerant control problem for a class of nonlinear systems with multiplicative intermittent faults. This chapter considers three fault scenarios as well. The multiplicative intermittent faults are depicted by random variables obeying Markov chain. The control objective of this study is given in the form of H∞ performance index. The sufficient condition for the existence of fault-tolerant controller is described by a set of matrix inequalities. Solving the matrix inequalities could obtain the corresponding dynamic output feedback controllers. The simulation results demonstrate the effectiveness of the proposed control strategy.4. Chapter 5 investigates the active fault-tolerant control problem for a class of Wiener systems with intermittent sensor faults, where the considered Wiener systems take the clinical anaesthesia in medical treatment as a background. In this chapter, the internal model control algorithm is introduced to construct a closed-loop anaesthesia system. For the potential intermittent sensor faults, the extended state observer is employed to generate residuals and to detect faults. Based on the output of the extended state observer, the fault-tolerant control system is established through a switching strategy. Finally, the simulation results on anaesthesia simulation platform illustrate the effectiveness of the proposed fault tolerant control method.