Fault Detection and Identification Using Parameter Estimation Techniques

This paper focuses on the use of parameter estimation techniques for the implementation of real-time Fault Detection and Diagnosis schemes. A detailed analysis of the nonrecursive and recursive Least Squares methods is given in the context of the system diagnosis problem, and a procedure for performing fault detection and identification for multivariable systems is proposed. An application example from the field of aircraft control is considered, which illustrates the suitability of the Recursive Least Squares method with exponential weighting and constant forgetting factor for Fault Detection and Fault Identification / Estimation within complex applications.