Robust Fault Detection and Fault-tolerant Controlfor Uncertain Systems with Nonlinear Faults

摘要

In this note, a novel robust fault diagnosis and fault-tolerance control scheme for uncertainsystems with nonlinear faults, is proposed. The uncertainties are required to satisfy some boundedconditions. Firstly, a robust fault detection and diagnosis algorithm is developed by using a fault detectionand approximation estimator in which an online approximator in the form of B-splines neural network isused to approximate the nonlinear fault dynamics. Once a fault is detected, a time-delay-optimizedreconstruction model based on scroll optimization methods for the fault system is obtained by using neuralnetwork. Moreover, on the basis of time-delay-optimized reconstruction model, we reconfigure a novelrobust fault-tolerant controller to the time-delay-optimized model by adopting the control approach totime-delay systems. To demonstrate the proposed fault detection and fault-tolerant control scheme, asimulation example is considered. Simulation results illustrate the applicability and effectiveness.