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

MSc Thesis #9838

Title:Automated Fault Isolation for Gas Turbine Alarms
Date: 1998
Abstract:The principal goal of this project is to develop an automatic fault isolation mechanism for gas turbine alarms based on the work done for the TIGER (TM) system which is a real time, knowledge based fault detection and diagnosis system.TIGER uses a variety of AI techniques to detect unusual situations and diagnose their cause automatically; this project aims to extend the portion that deals with the alarms from the gas turbine controller. Currently when there is a major incident, such as s trip, there can be many interrelated sources of malfunction. Knowing the inputs and internal logic processing for most of the alarms and even having a large portion of relevant information it will be shown that it is feasible to identify which of several possible causes, actually provoked the alarm.What is going to be illustrated is an automated fault isolation mechanism that can combine the inputs available from the gas turbine with the alarm descriptions to identify the root cause of an alarm; a real industrial problem.

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