Dynamic System Based On Lmi Approach To Fault Tolerant Control

With economic demand for high plant availability, dependability is becoming an essential concern in industrial automation. Fault-tolerance, as an important way to achieve such an aim, is defined as the ability of a controlled system to maintain control objectives despite the occurrence of faults. There have been considerable amounts of researches in the field of fault-tolerant control, among which are reliable stabilization, simultaneous stabilization and integrity. The main contents of the present scientific dissertation can be outlined as follows:1. The design method of robust fault-tolerant control systems based on LMI for linear multivariable system is studied. Sufficient conditions for the existence of controller are analyzed. And effective controller design methods are provided via LMI approach. If the satisfying controller exists ,and then a region of the controller is given. In addition ,other performances can be considered and designed .2. The problem of robust and reliable H∞ fault design for linear uncertain systems with time-varying norm-bounded parameter uncertainty and also with actuator failures based on Linear Matrix Inequality(LMI). An H∞ -norm bound constraint on attenuation of augmented disturbances are introduced. Sufficient conditions for the existence of controller are derived. Ranges of consistent indices are analyzed. And effective controller design methods are provided via LMI approach.3. The satisfactory fault-tolerant control problem, which guarantees closed system to possess multiple performance requirements, is formulated in an LMI (Linear Matrix Inequality) setting. Consistency theory on circular pole index and steady variance index for linear stochastic systems under state feedback and satisfactory control thought are introduced to fault-tolerant control field. On the premise of system integrity, we analyze ranges of consistent indices, and provide effective controller design methods for system with constraints of consistent indices via LMI approach.