Vers le concept de redondance algorithmique pour le diagnostic : le projet DIALOGS

Abstract : This thesis treats about fault diagnosis and decision support systems on industrial plants. The framework for this research was offered by the European project EUREKA MAINE DIALOGS EU996. The results presented here contributed to the software implementation of the diagnostic assistant system generator called DIALOGS. The document is organized in two main parts. The first part emphasizes the role and the place of the diagnostic task in the manmachine interface, and focuses on the structure of a diagnostic assistant system. The basic concepts of the DIALOGS project are presented next. The importance of using complementary diagnostic methods is shown and the new concept of algorithmic redundancy is introduced. In the second part, an overview of some real-time statistical methods is made. There are presented a steady-state detection method, three limit checking tests, and an outliers detection and elimination method. The concept of algorithmic redundancy is developed and an original diagnostic method based on balance equations is given as example. The results are then used to generate a fuzzy alarm indicator. This alarm is interpreted using a qualitative reasoning and can be used either directly by the human operator or indirectly by another decision maker at a higher hierarchical level.
Document type :
Theses
File URL :
http://docnum.univ-lorraine.fr/prive/SCD_T_1997_0004_GHETIE.pdf
Complete list of metadatas

https://hal.univ-lorraine.fr/tel-01747375
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 10:58:12 AM
Last modification on : Friday, March 30, 2018 - 1:28:52 AM

Identifiers

  • HAL Id : tel-01747375, version 1

Collections

Citation

Marius Alexandru Ghetie. Vers le concept de redondance algorithmique pour le diagnostic : le projet DIALOGS. Autre. Université Henri Poincaré - Nancy 1, 1997. Français. ⟨NNT : 1997NAN10004⟩. ⟨tel-01747375⟩

Share

Metrics

Record views

6