Identification and parity space techniques for failure detection in SISO systems including modelling errors

This paper deals with fault detection and isolation methods (FDI) in linear dynamic SISO systems. Two different approaches for the residuals generation are reviewed: parity space and parameters identification based approaches. These two methods are applied to a stochastic SISO model including modelling and measurement errors. This leads to an analytical formulation of the residuals from the two methods, and enables classification of the errors. First, a comparison between the residuals is performed, then statistical and geometrical interpretations of the result are proposed in this framework. Finally, some simulation results are developed in order to exemplify the similarity of the two approaches