A finite memory observer approach to the design of fault detection algorithms

Modelling uncertainty is inevitable for most systems. Residual robustness with respect to modelling uncertainty is one of the most important issues in any practical fault diagnosis scheme. A significant class of model-based fault detection and isolation structure uses observers to generate residuals. Infinite dynamic observer memory inflicts a phenomenon generally referred as divergence. To overcome this drawback, a structure which intrinsically has finite process memory is proposed. Observer residuals give a possibility to check up explicitly how the more recent measurements which are suspected to be provided by a faulty sensor fit the process history. This paper presents extensions and improvements on the finite memory observer. The main contribution is to incorporate the parity relation design and observer into a diagnosis scheme. The application of the finite memory observer to the sensor fault detection problem is illustrated by a numerical example.