Diagnostic based on estimation using linear programming for partially observable petri nets with indistinguishable events
| Autor: | Atef Khedher, Philippe Declerck, Anas Kamoun, Amira Chouchane |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
0209 industrial biotechnology
021103 operations research Information Systems and Management Linear programming Computer science 0211 other engineering and technologies Process (computing) Observable 02 engineering and technology Management Science and Operations Research Petri net Unobservable Fault detection and isolation Management Information Systems Polyhedron 020901 industrial engineering & automation Algebraic number Algorithm Computer Science::Databases Information Systems |
| Zdroj: | International Journal of Systems Science: Operations & Logistics. 7:192-205 |
| ISSN: | 2330-2682 2330-2674 |
| DOI: | 10.1080/23302674.2018.1554169 |
| Popis: | In this paper, we design a diagnostic technique for a partially observed labelled Petri net where the faults of the system are modelled by unobservable transitions. The fault detection and isolation uses an on-line count vector estimation associated with the firing of unobservable transitions exploiting the observation of firing occurrences of some observable transitions. The support of the approach is an algebraic description of the process under the form of a polyhedron developed on a receding horizon. We show that a diagnostic can be made despite that different transitions can share the same label and that the unobservable part of the Petri net can contain circuits. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|