Failure Detection and Isolation Methodology

This paper presents a methodology for the design and implementation of model-based failure detection and isolation systems for industrial applications. The methodology addresses various issues, including: failure detection in the presence of modeling errors and noise, failure type classification, sensor failure isolation and system failure isolation. The methodology is illustrated on a monitoring system for a turbine engine.