Estimation of Multiple Transient Actuator Faults Using Augmented Error Technique

Abstract

A fault diagnosis scheme for nonlinear uncertain
dynamic systems with both abrupt and incipient faults is discussed.
An active fault approach is designed that utilizes adaptive laws of
augmented error technique in such a way that accounts for matching
uncertainties and the occurrence of actuator faults. The main idea is
designing the robust fault diagnosis scheme that guarantee stability of
the system in the presence of faults. Using the augmented error
technique from model reference adaptive control, an observation
error model is formulated to give an adaptive diagnostic algorithm
which produces the estimate of actuator faults. Changes in the system
due to faults are modelled as unknown nonlinear functions. An
occurred fault is isolated if the residual associated with the observer
remains below its corresponding adaptive threshold, while at least
one of the components of the residual associated with all other
estimators exceeds its threshold at some finite time. Unknown Input
Observer (UIO) is an estimator which is decoupled from the
unknown inputs (certain disturbances, or faults) that may be acting on
the system. A key design issue of the proposed fault isolation scheme
is the derivation of adaptive residual thresholds associated with
observer. The simulation result indicate that the proposed algorithm
is more realistic, in the sense that better decoupling properties can be
assured without knowledge about unknown inputs, and it is
potentially useful in the development of a fault-tolerant control
system.