Content area
Full Text
Abstract
The architecture design of instrumentation of a system is crucial for improving the monitoring of industrial processes. Indeed, the position of the sensors and their numbers play a very important role in the monitoring of a system. This paper presents a new approach to sensor placement in order to monitor element failures defined in specifications. It relies on computing a linking in tripartite graph. The latter is obtained from Bond-graph model, which in turn represents an industrial physical plant.
Keywords: Bond-graph, tripartite graph, monitoring, linking, sensor placement, observability degree
Received: September, 22, 2015
1. Introduction
An automation system in engineering processes contributes significantly to industry development. The size and complexity of these systems lay difficulties in the design and implementation of various monitoring methods. To ensure the proper functioning of these systems, it is necessary to use the techniques of control and verification, which are monitoring techniques.
The detection and isolation of faults industrial plant are essential task in the monitoring process.
The monitoring of systems (possibility of detection and fault isolation) mainly depends on the architecture of instrumentation implementation [1] [2].
Improving the safety process is essentially based on detection and isolation procedures (FDI for Fault Detection and Isolation procedure). FDI algorithms consist of comparing the current behaviour of the system and the behaviour of the reference model that describes the normal behaviour mode.
However, the essential problem of fault detection is to infer the existence of a defect in the structure from measurements taken by sensors located on the structure. It would be necessary in practice to optimize the number and placement of sensors to minimize the cost and increase the reliability of the system [3].
According to holding that there is an effective procedure for fault detection, the problem is how to place sensors and where to place them reliable efficiency. Sensor placement decisions depend on expected objectives. For example, in the theory of command, the sensor placement is used to provide sufficient information to the control system.
In [4], authors proposed a sensor placement method that treats linear systems. This method allows the use of Gauss-Jordan elimination to find a minimum set of variables to be measured. This ensures observability variable while minimizing the cost of sensors. In...