Failure Diagnosis on Discrete Event Systems

Abstract

The modern technology advances to a point where it is possible and extensively desirable to improve reliability and the technical process safety. This is achieved by computer implanted FDI procedures (Fault Detection and Isolation). However, the malfunction of actuators, sensors and of the process components, as well as erroneous actions of human operators can have some disastrous consequences in high risk systems such as: Spatial engines (Astronomy), aircrafts (Aviation), nuclear reactors and chemical plants. Thus, each failure or fault can lead to shutdowns or a rupture of service and consequently a plant output reduction. There is an improvement of consciousness and attitude to possible disaster provoked by failures that could enable a failure tolerating system development. Such system must maintain a optimal performance during normal operating conditions and must handle encountered critical situations during which the system’s conditions are abnormal that is by performing of detection and diagnosis procedures and reconfiguration according to accurate software programs. In this study, we focus on the diagnosis of the flexible manufacturing systems which are described by a model based on the Petri nets. The basic idea consists of residuals generators resulting from the equation of marking evolution of the process and having appropriated structures to facilitate fault isolation.