# Surveillance des systèmes de production automatisés : détection et aide au diagnostic

Abstract : The goal of the thesis is to develop a new approach of monitoring discrete events systems. A method based on the monitoring of the service time was developed using timed automata as a modelling tool.
From a controlled system, the first objective is the monitoring of the different activities of this system in order to improve its availability. It is approved by the minimization of the number of stops which penalize the production; by following in real time, the operating condition of all sensors of the system. Thus, for each task, the monitoring of spending time between two events (the moment when the command is ordered the task and the moment when the sensor indicates the end of task execution) allows to detect possible failures. The second step corresponds to the diagnosis which is the determination of the causes of the problem observed.
Until now, the approaches of monitoring use a reference model based a priori on the knowledge of all the prohibited situations of the system. The proposed model of monitoring has a real advantage. Indeed, the detection of possible failures does not require an exhaustive list of all the possible failures of the system. We introduce a tolerance; depending of the category of the production system and the nature of the tasks to be supervised. Thus the concept of functioning on degraded mode' is introduced. Beyond this tolerance, the system will be in failure mode'. Various examples illustrate our method, which show the power of such an approach.
Then an optimisation algorithm allowing to calculate the fault latency is proposed, as well as the performance evaluation of the monitoring system.
Mots-clés :
Document type :
Theses
Domain :

Cited literature [6 references]

https://tel.archives-ouvertes.fr/tel-00169988
Contributor : Patricia Reynier <>
Submitted on : Monday, September 17, 2007 - 1:44:52 PM
Last modification on : Friday, November 6, 2020 - 3:46:08 AM
Long-term archiving on: : Friday, April 9, 2010 - 1:37:41 AM

### Identifiers

• HAL Id : tel-00169988, version 1

### Citation

Hassan Rayhane. Surveillance des systèmes de production automatisés : détection et aide au diagnostic. Automatique / Robotique. Institut National Polytechnique de Grenoble - INPG, 2004. Français. ⟨tel-00169988⟩

Record views