Fault Detection and Isolation for Industrial Systems

Abstract

The chapter analyzes a fault detection and isolation approach for efficient condition monitoring of industrial systems. As shown, two main issues in statistical methods for fault diagnosis are residuals generation and fault threshold selection. For residuals generation an accurate model of the system in the fault-free condition is needed. Such models can be obtained through nonlinear identification techniques or through nonlinear state estimation and filtering methods. On the other hand the fault threshold should enable both diagnosis of incipient faults and minimization of the false alarms rate.