This paper shows two different approaches for improving the fault diagnosis
of gas turbines. They are based on the use of the information embedded wit
hin the control system. The first approach consists of the automatic develo
pment of a troubleshooting system which can be automatically built using th
e deterministic knowledge embedded within the Programmable Logic Controller
(PLC). This system tries to overcome diagnosis deficiencies present especi
ally in older gas turbines PLC controllers, The second approach is based on
the analysis of the digital control system to detect the Faults in the fee
dback systems (actuators, components or sensors), such is the case of gas t
urbines. The interaction between the system and the control system allows t
o use the information of the control system to detect faults in the system.
An example of the second stage nozzles of the turbines illustrates this ap
proach. (C) 2000 Elsevier Science Ltd. All rights reserved.