This paper presents an intelligent methodology for diagnostics of incipient faults in turbine machine. A fault diagnosis system is developed for turbine machine. In this system, wavelet transform techniques are used in combination with function approximation model to extract fault features used in the diagnosis of turbine machine faults. The neural networks is constituted. The main contributions of this paper are two aspects. A improvement method based on nonlinear adaptive algorithm has been developed for excitation function approximation of neural networks. In order to perform diagnosis using intelligent system, a preprocessing of singularity fault signal is required. The second contribution is the development of a neural networks classifier for identification of fault. The developed system is scalable to different turbine machine and it has been successfully demonstrated with a turbine generator unit.
[1]
Mo-Yuen Chow,et al.
A neural network approach to real-time condition monitoring of induction motors
,
1991
.
[2]
Michael Nikolaou,et al.
An approach to fault diagnosis of chemical processes via neural networks
,
1993
.
[3]
G. G. Stokes.
"J."
,
1890,
The New Yale Book of Quotations.
[4]
Mo-Yuen Chow,et al.
Neural-network-based motor rolling bearing fault diagnosis
,
2000,
IEEE Trans. Ind. Electron..
[5]
Chuei-Tin Chang,et al.
A simple design strategy for fault monitoring systems
,
1993
.
[6]
Ibrahim N. Tansel,et al.
Recognition of chatter with neural networks
,
1991
.
[7]
Moustafa M. Fahmy,et al.
Neural classifiers and statistical pattern recognition: applications for currently established links
,
1997,
IEEE Trans. Syst. Man Cybern. Part B.