Observability, redundancy, reliability and integrated design of measurement systems

Marie Luong (1) , Didier Maquin (1) , Chang Trung Huynh (2) , José Ragot (1)

1 CRAN - Centre de Recherche en Automatique de Nancy
2 Elf Aquitaine

The problem of sensor placement is considered. The quality of diagnosis depends on the number and location of measurements. We first deal with this dependence by analysing the observability. The concepts of degree of redundancy is introduced. The criterion of reliability is useful for designer to enhance the reliability of the measurement system. When some variables are required for the control of the process, the reliability of the measurement system may be computed taking into account the reliability of all sensors. In order to improve the diagnosis, we propose an original and computational tool for designer to conceive the automatic and optimal sensor placement. Our objective is to design the measurement system meeting the following constraints: observability of variables required for the controlling and the maintenance, fulfilment of different degrees of redundancy imposed for some variables and fulfilment of criterion of sensor cost or/and reliability.

Mots clés: Degree of redundancy; observability; reliability; canonical matrix; cycle analysis

Domaines: Automatique / Robotique

Marie Luong, Didier Maquin, Chang Trung Huynh, José Ragot. Observability, redundancy, reliability and integrated design of measurement systems. 2nd IFAC Symposium on Intelligent Components and Instruments for Control Applications, SICICA'94, Jun 1994, Budapest, Hungary. ⟨hal-00302975⟩