In this paper a neural network approach is used to diagnose faults in gas turbine engines. In particular, faults are required to be diagnosed from many engines within a fleet of aircraft. The problem of diagnosing faults across a fleet is difficult because any faults have to be identified using parameters which vary naturally from engine to engine within a healthy fleet. Such variations in no-fault performance arise due to tolerances in manufacture and construction of critical engine components. Any in-field technique which is applied to a number of engines needs to allow for or accommodate this variation in the no-fault performance but still recognise abnormal behaviour. For this study a low bypass ratio turbofan engine was tested with a number of faults implanted in the variable geometry schedules of the compressor and exhaust nozzle. A probabilistic neural network is employed to diagnose the faults with encouraging results. The method is shown to be capable of often diagnosing faults even when the parameter changes due to the fault are less than the no-fault engine-to-engine variation.
Fault diagnosis of fleet engines using neural networks
Fehlerdiagnose der Triebwerke einer Luftfahrzeugflotte mit Hilfe von neuronalen Netzen
1995
11 Seiten, 10 Bilder, 3 Tabellen, 14 Quellen
Conference paper
English
Fault Diagnosis of Fleet Engines Using Neural Networks
British Library Conference Proceedings | 1995
|Fault diagnosis of automotive engines using artificial neural networks
Automotive engineering | 1993
|Fault diagnosis of automotive engines using artificial neural networks
Tema Archive | 1993
|Fault Diagnosis in Combustion Engines Using Artificial Neural Networks
SAE Technical Papers | 2021
|Fault Diagnosis in Combustion Engines Using Artificial Neural Networks
British Library Conference Proceedings | 2020
|