Diagnosis for loose blades in gas turbines using wavelet analysis

Abstract

The application of wavelet analysis to diagnose loose blades condition in gas turbines is examined in this paper. Experimental studies were undertaken to simulate loose blades condition occurring in gas turbines in an attempt to understand vibration response associated with loose blades under different operating conditions. Results showed that loose blades were undetectable under steady state operating condition. During turbine coast down, a loose blade could be detected based on the impactic signals induced by the loose blades on the rotor and thus excited the natural frequencies of the rotor assembly. Results from the coast down condition showed that wavelet analysis was more sensitive and effective than Fourier analysis for loose blade diagnosis. The severity, the number, and the configuration of the loose blades could be potentially estimated based on the pattern of the coast down wavelet map.