FAULT DIAGNOSIS AND FAULT-TOLERANT CONTROL IN NONLINEAR SYSTEMS
Fault-tolerance is an essential property of many modern intelligent control systems. This dissertation presents a general framework for fault diagnosis and fault-tolerant control in nonlinear dynamical systems in the presence of possibly unstructured modeling uncertainty. The overall architecture is based on a learning approach, where the unknown fault is estimated using adaptive and on-line approximation techniques. First, the problem of fault detection and isolation in nonlinear uncertain systems is investigated. A novel fault isolation scheme is presented with its robustness and sensitivity properties enhanced by the use of adaptive thresholds in the residual evaluation stage. The fault isolation scheme is rigorously analyzed for its fault isolability condition and fault isolation time. Then we integrate the fault diagnosis (fault detection and isolation) scheme with fault-tolerant control design. Based on the fault information obtained during the diagnosis procedure, the system controller is reconfigured after fault detection and fault isolation, respectively, to compensate the effects of the fault. The closed-loop stability of the integrated fault-tolerant control system is established for different modes of the controlled plant. The effectiveness of the proposed fault diagnosis and fault-tolerant control scheme is illustrated via simulations in the three-tank system, a rigid-link robotic manipulator and the van der Pol oscillator system.
School:University of Cincinnati
School Location:USA - Ohio
Source Type:Master's Thesis
Keywords:fault detection isolation tolerant control accommodation nonlinear systems
Date of Publication:01/01/2002