Parameter Transient Behavior Analysis on Fault Tolerant Control System

In a fault tolerant control (FTC) system, a parameter varying FTC law is reconguredbased on fault parameters estimated by fault detection and isolation (FDI) modules.FDI modules require some time to detect fault occurrences in aero-vehicle dynamics.This paper illustrates analysis of a FTC system based on estimated fault parametertransient behavior which may include false fault detections during a short time interval.Using Lyapunov function analysis, the upper bound of an induced- L 2 norm of theFTC system performance is calculated as a function of a fault detection time and theexponential decay rate of the Lyapunov function. 1 INTRODUCTIONIn the past decades, there has been interest in a FTC system which has ability to detect ac-tuator/sensor faults automatically, and to prevent faults from developing into system failure.Especially, in designing a ight control system, it has been researched for achieving singleaircraft accident prevention [1, 2, 3, 4, 5]. An active FTC system requires its control lawto react to actuator/sensor faults through reconguration and fault detection and isolation(FDI) modules to detect actuator/sensor fault occurrences. FTC law synthesis for aerospacevehicles has been studied since 1980’s based on possible faults on actuators and sensors [2].Recently, using linear matrix inequality (LMI) optimization solutions [3, 4, 5], linear pa-rameter varying (LPV) FTC laws are designed, based on an open-loop system modeled asa function of fault parameters. In the LPV-FTC synthesis procedure, it is assumed to beimminently identied by separate fault detection and isolation (FDI) modules. FDI modulesare designed, based on only an open-loop system and applied in a closed-loop system [5, 6, 7].In general, FDI modules and FTC laws are individually designed, without considering theother dynamics [4, 5].After individually designing an LPV-FTC law and FDI modules, analysis of a FTC systemis required to validate the FTC laws and FDI modules together. A typical way of analyzinga FTC system is nonlinear simulation with the pre-dened command inputs (not all possiblecommand inputs), for possible fault scenario. It is hard to provide a generic criteria of aFTC system via nonlinear simulations. A FTC system analysis frame work is required toprovide a certain criteria of a FTC system related with characteristics of FDI modules.