Sensor classification for observability and diagnosis under sensor failure

International audience; This paper is concerned with the study of property preservation under sensor failure in a structured framework. We focus our attention on the observability preservation as well as the solvability preservation of Fault Detection and Isolation (FDI) problem under sensor failure. More precisely we will characterize among the sensors those which are critical \emph{i.e.} which failure leads to property loss, those which are useless to hold the property and the set of those which are useful. In this paper we use parameterized models which are well fitted to describe structural properties of dynamical systems. Using a graph approach we classify the sensors with respect to their importance for the preservation of observability and solvability of FDI problem under sensor failure. The sensor classification for observability and FDI problem is performed on a distillation column example.