Augmented Proportional and Integral Observer Design for Fault Estimation in Discrete-time Systems With Applications to Wind Turbine Systems and Electro-mechanical Servo Systems

In this paper, a fault estimation technique is addressed to simultaneously estimate system states, actuator, and sensor faults for discrete-time dynamic systems. Specifically, an augmented system is constructed by defining an extended state vector composed of original system states and actuator faults. For this augmented system, an augmented proportional and integral observer is addressed to simultaneously estimate system states, actuator faults as well sensor faults for a discrete-time linear system. A robust augmented proportional and integral observer is designed for Lipschitz nonlinear systems subjected to unknown input uncertainties. The proposed approaches are applied to wind turbine drive train system and electro-mechanical servo system for the validation, which have shown satisfactory estimation performance.