The University of Edinburgh -
Division of Informatics
Forrest Hill & 80 South Bridge

MSc Thesis #9671

Title:Tmdoctor: a Fuzzy Rule- and Case-Based Expert System for Turbomachinery Diagnosis
Date: 1996
Abstract:Turbomachinery can be found in many places from small to large scale plants and their success depends on the continued and safe operation of their machinery. There is a real need for a method, such as an expert system, to provide quick and accurate diagnoses for machine vibration problems. This thesis presents an expert system implemented in FuzzyCLIPS for the diagnosis of vibration problems in turbomachinery. The system, which works by incremental forward chaining, employs both fuzzy logic based approximate reasoning and traditional certainty factor techniques to deal with uncertainty. The former is used to model uncertainty associated with vague knowledge while the latter is mainly used to rank possible vibration causes using incremental forward chaining. Because human experts find past experience helpful in diagnosis, a simple case-based reasoning component was incorporated into the system to provide more accurate diagnoses when similar past experience can be applied. The system has been tested against real cases provided by Intelligent Applications Ltd., which initiated this project, and according to the experiments we have done so far, the system can identify the actual underlying causes of typical vibration problems. A paper about part of this project work has been accepted for publication at the 1st online workshop on soft computing (WSC1).

[Search These Pages] [DAI Home Page] [Comment]