
V Webinar 2021: Behavioral identification of Discrete Event Systems. Application to fault diagnosis
Published at : November 26, 2021
Behavioral identification of Discrete Event Systems. Application to fault diagnosis
Webinar promovido pela SBA com apoio do CT de Automação
Palestrante: Prof.Prof. Jean-Jacques Lesage
Resumo: Discrete Event Systems (DES) are dynamical systems whose changes of state are driven by occurrence of events. Modelling of a DES most often consists in translating this event driven dynamics by means of Finite Automata or Petri nets. Due to numerous factors (huge number of states, huge number of events, strong parallelism in behavior ...) the "by-knowledge" building of such a model is always difficult and time consuming, and most often becomes error prone when the size of the system to be modelled
increases. In the field of continuous systems, which are time driven systems, identification techniques, allowing the building of the behavioral model of an existing system by an experimental approach are used since long time. Identification technics are today very efficient and are often used in industry. In the field of DES, such identification approaches have only been developed since few years. In this webinar, different approaches allowing to build the behavioral model of a DES (under the form of Finite Automata or Petri nets) by identification will be presented. Afterwards, in order to show the efficiency of identification methods and the accuracy of the identified models, the use of such an identified model for fault diagnosis of DES will be presented.
Webinar promovido pela SBA com apoio do CT de Automação
Palestrante: Prof.Prof. Jean-Jacques Lesage
Resumo: Discrete Event Systems (DES) are dynamical systems whose changes of state are driven by occurrence of events. Modelling of a DES most often consists in translating this event driven dynamics by means of Finite Automata or Petri nets. Due to numerous factors (huge number of states, huge number of events, strong parallelism in behavior ...) the "by-knowledge" building of such a model is always difficult and time consuming, and most often becomes error prone when the size of the system to be modelled
increases. In the field of continuous systems, which are time driven systems, identification techniques, allowing the building of the behavioral model of an existing system by an experimental approach are used since long time. Identification technics are today very efficient and are often used in industry. In the field of DES, such identification approaches have only been developed since few years. In this webinar, different approaches allowing to build the behavioral model of a DES (under the form of Finite Automata or Petri nets) by identification will be presented. Afterwards, in order to show the efficiency of identification methods and the accuracy of the identified models, the use of such an identified model for fault diagnosis of DES will be presented.

Sociedade Brasileira de Automaticasbaquaternions