Study On Robust Fault Diagnosis Strategies Of Nonlinear Deterministic Systems

Fault detection and isolation (FDI) based on analytical models is one class of the most important approaches in fault diagnosis area. Due to the universal existence of nonlinearities and model uncertainties in practice, robust fault diagnosis of nonlinear systems has received much more attention in recent years. Since model uncertainties are unexpected dynamics of the system as well as faults, they constitute a source of false alarms which corrupt the performance of the FDI system to such an extent that it may even become totally useless. As is well-known, nonlinearity is a fundamental problem in mathematics, which causes many difficulties in the construction of FDI systems, such as the design of observers.This thesis investigates robust fault diagnosis of nonlinear deterministic systems. The main contributions are:1. A Survey on the robust fault diagnosis approaches of nonlinear systems is presented, which include two main approaches: one is based on unknown input observer and another one is based on adaptive learning. The difficulties and future developments of this area are also discussed.2. By introducing sliding mode observer with boundary layer control and a new adaptive law, we propose a fast robust fault diagnosis strategy of nonlinear systems based on Polycarpou's online approximator, which is proved theoretically and illustrated by simulations to be much faster than Polycarpou's approach.3. Inspired by the unknown input Kalman filter, we extend it to nonlinear cases as the EKF does. Using this as a nonlinear observer (called unknown input extended Kalman observer: UIEKO), we prove its convergence under some mild conditions. Simulation studies demonstrate the effectiveness of the robust fault diagnosis strategy based on the UIEKO.