Dependability analysis of instrumented linear static systems based on their observability

ABSTRACT

In many applications, the ability of a system to operate correctly is related to the capacity to estimate its state variables. The quality of the estimated variables is strongly conditioned by sensors state, number and location. Availability and reliability are highly sought in industry where systems should continue operating even in the presence of some failures. This makes the problem of designing instrumentation systems very important in system control, diagnosis and reliability. In this paper, we propose an approach to assess the dependability of an instrumented system from its observability perspective. This approach is based on two analysis and computation stages. In the first one, the concept of analytical redundancy degree (ARD) for state estimation is discussed. The ARD reflects the process’ capacity degree to tolerate some sensors’ faults without affecting its observability. In the second stage, the classical definition of system's reliability will be extended to the estimation of its state variables. Based on the ARD resulting from the first stage analysis, a stochastic activity networks-based modelling approach and Monte Carlo simulation will be used to assess the minimal reliability and availability of the system's state variables based on their observability. This study concerns material/energy transportation systems where variables are related to each other by static linear equations.

KEYWORDS:

• Observability
• sensor failures
• dependability
• stochastic activity networks (SANs)
• analytical redundancy degree
• Monte Carlo simulation

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Samia Maza

Samia Maza is an associate professor at ENSEM, a school of Electrical and Mechanical Engineering in Nancy and member of the Centre de Recherche en Automatique de Nancy since 2006. She got her Ph.D degree in 2003 from the University of Nantes in the field of discrete event systems. Her current research interests are the development of integrated approaches for dependability analysis of automated and fault-tolerant systems as well as the modelling, simulation and control of discrete events systems.

José Ragot

Prof. Ragot is researcher in the "Centre de Recherche en Automatique de Nancy, France" (CRAN, CNRS UMR 7039) where he was the head of the group "process diagnosis" during 12 years. Presently he is Emeritus Professor at Lorraine University. His major research fields include data validation and reconciliation, process diagnosis, fault detection and isolation, fault tolerant control. A part of his activities is devoted to modeling and state estimation adapted to process diagnosis and mainly in the field of multi-models. He regularly contributes to international activities as a member of TC4 and TC6 of IFAC. He has published about 130 papers in international journals, 370 communications in international conferences and 4 books. Applications have been in various fields such as mineral and metallurgical processing, chemical engineering, water treatment, aerospace, environmental processes.