Sensor Network Design for Failure Detection and Isolation

Abstract

The model-based approach to fault detection and isolation in automated process has received considerable attention during the last two decades. One of the approaches to facilitate fault isolation is to design a structured residual set. The term structured here means that each residual is designed to be sensitive to a certain group of faults and insensitive to others. However, despite this issue has been recognised for a long time, the problem of designing such residuals is not completely solved. In the context of linear static systems, this paper proposes a new method for sensor network design respecting some sensor failure detectability and isolability requirements.

Diagnosability Analysis Based on Component Supported Analytical Redundancy Relations

“…The available sensors are given by Sa={S(CV), S(X), S(F), S(P1), S(P2), S(Pz), S(T), S(KV3)} 12 . The mathematical equations describing the components behaviour come from [17].…”

“…analysing properties such as diagnosability and characterising the instrumentation system in terms of the number of sensors and their placement are highly valuable. There is a significant amount of work dealing with diagnosability and sensor placement both in the AI Model Based Diagnosis (DX) community [5], [20], [8] and in the Control Model Based Diagnosis (FDI) community [2], [12], [10], [11].…”

“…Let us call M-ARRs the ARRs that are directly obtained from a given complete matching M. The number of M-ARRs is the same for any existing complete matching, as it only depends on the redundancy degree of the observed system [12]. Given the set of M-ARRs, also called primary ARRs, additional ARRs can be obtained by combining M-ARRs, i.e.…”

A general-purpose method for Pareto optimal placement of flow rate and concentration sensors in networked systems – With application to wastewater treatment plants

“…In this and prior work, we have only considered observability and redundancy objectives. However, redundancy only ensures that one can recognize the presence of sensor faults (detection), not necessarily identify them (isolation) (e.g., Luong et al, 1997;Commault and Dion, 2007;Pérez et al, 2011;Sarrate et al, 2014;Palleti et al, 2016). To do so, the value of surplus redundancy should be studied in more detail.…”

Diagnosability of Networks of Hybrid Systems

“…Fault detection and isolation methodologies have sought to ensure the robust operation of dynamic systems in the presence of faults, modeling uncertainties as well as process and measurement noise through the use of control and system theoretic techniques. A good overview of these methods can be found in Gissinger et al (2000) Hou and Muller (1994), Frank (1996) and Luong et al (1997). This area of work has generally focussed on specific examples or classes of systems and has typically used observers and residual generators to detect deviations of the observed system trajectory away from some nominal operating condition.…”