Grey Clustering Analysis Based Classifier for Steam Turbine-Generator Fault Diagnosis

This paper proposes a method for steam turbine-generator fault diagnosis using grey clustering analysis (GCA). According to the field records, diagnostic information can be provided to monitor mechanical condition by the spectrum of the vibration signal. Frequency-based features are computed by fast Fourier transformation (FFT), the frequency ranges are <0.4f, 1f, 2f, 3f, and >3f. The maximum and minimum values of power spectrum indicate mechanical vibration fault at a particular frequency, and frequency patterns are applied to diagnose faults. For numerical tests with practical filed records, test results were conducted to show the proposed method demonstrates computational efficiency and high accuracy.