Mode-based fault detection and isolation techniques: the parameter identification approach within the framework of the TOPMUSS CAE-system

The increase in international competition and the constant research for improving reliability and reducing costs have spurred the applications of the system monitoring to power plant; the monitoring functions are extremely useful for plant management purposes and fault diagnosis. In particular failure detection and isolation (FDI) strategies monitor system operations for degradation and, if detected, classify the failure source. In this paper, after a brief description of the different approaches at the FDI problem, we consider the model-based methods (analytical redundancy) and in particular the parameter identification of a system on the basis of observed input/output data. Some algorithms have been developed inside a European project (TOPMUSS) and the monitoring system realized by Elsag Bailey Energy and Industry Division can now add to the standard functions these tools for the fault detection routines. An illustrative example of an industrial application is given.<
>