An Approach To Fault Diagnosis With Sensor Faults Based On Finite Frequency And Parameter Estimation

In order to satisfy the dynamic control system performance indexes such as reliability, safety and effectiveness, fault diagnosis (FD) technology got rapid development over decades. FD has been one of the hot spots in current research due to the increasing demand for reliability and safety in aerospace, astronautics areas, and industrial processes. And the model-based FD approach has been accepted as a widely used effective approach. Recently, H∞ techniques have been introduced in model-based FD approach to improve the disturbance robustness. Many relevant approaches have been presented. However, there are still some problems to be developed associated with the existing approaches. For example, the frequencies of stuck faults are zeros. For these cases, designing a filter or observer in the full frequency domain may introduce conservatism to some extent.Based on previous works of others, this thesis presents an approach to deal with the inaccurateness of the frequency weights by the help of the generalized Kalman-Yakubovich-Popov (GKYP) lemma. The chapter 3 studies the problem of FD of continuous linear system with polytopic-type uncertainties because the uncertainty of actual system. What is more, the content of chapter 3 is further applied to the uncertain networked control systems with random packet loss in chapter 4. Parts of the results of chapter 3 and chapter 4 are applied to the fault detection filter or observer designs of F404 aircraft engine system and NASA High Alpha Research Vehicle aircraft system by simulations.Chapters 1-2 summarize the development, and overseas and domestic research status in FD literature, and give some relevant preliminaries about the considered problem.For a class of uncertain continuous linear systems, chapter 3 investigates the FD problem for sensor failure fault. Additionally, we take the frequency characteristic of servo inputs and disturbances into consideration, and a frequency domain inequality problem is converted into a linear matrix inequality (LMI) problem by the help of GKYP lemma. Finally, LMI-based optimization problem is given to design the filter. A simulation example of F-404 aircraft engine is given in the end of this chapter, and which illustrates the effectiveness of the proposed approaches.For a class of networked control systems with uncertainties and random packet dropouts, chapter 4 investigates the FD problem for sensor failure fault, and the content of this chapter is extended by chapter 3. A simulation example of NASA High Alpha Research Vehicle is given in the end of this chapter, and which illustrates the effectiveness of the proposed approaches.Finally, the results of the dissertation are summarized and further research topics are pointed out.