TIGER With model based diagnosis: initial deployment

Citation
R. Milne et al., TIGER With model based diagnosis: initial deployment, KNOWL-BAS S, 14(3-4), 2001, pp. 213-222
Citations number
8
Categorie Soggetti
AI Robotics and Automatic Control
Journal title
KNOWLEDGE-BASED SYSTEMS
ISSN journal
09507051 → ACNP
Volume
14
Issue
3-4
Year of publication
2001
Pages
213 - 222
Database
ISI
SICI code
0950-7051(200106)14:3-4<213:TWMBDI>2.0.ZU;2-3
Abstract
The TIGER system is in continuous use on over 20 gas turbines across four c ontinents, performing fault diagnosis of industrial gas turbines used for p ower generation TIGER receives several hundred data inputs at a once per se cond interval and performs several layers of fault detection and diagnosis using expert system techniques in a temporal reasoning rulebase. Development of the system has now extended for more than six years and more than 80 man-years. The diagnostics are supported by an extensive user inte rface, data archive, trending and display system. Across the 20 + current l ocations, TIGER has helped to identify and diagnose over 700 incidents repr esenting 200 different fault categories. Because of the high value of gas t urbines and the high expenses in operating them, TIGER's benefit is conside rable and has been documented at $150,000 or more for most of the locations . This paper gives a summary as TIGER is today, the benefits to many of the s ites and an overview of the recently added model based diagnosis module, Ca -En. (C) 2001 Elsevier Science B.V. All rights reserved.