Vibration monitoring of rotating machines is probably the most established diagnostic method. The application of acoustic emission (AE) for rotating machine fault diagnosis is gained as a complementary tool; however, limitations in the successful application of the AE technique have been partly due to the difficulty in processing, interpreting and classifying the acquired data. The experimental investigation reported in this paper is centred on the application of the AE technique for identifying the seal rubbing on the rotor rig. An experimental test rig was designed to simulate the 200MW gas turbine rotor shafts. On the rig different degrees rubbing-impact on the seal is performed. The AE transducer and the vibration acceleration transducer are set on the bearing block. Comparisons between AE and vibration analysis over a range of speed and different degrees rubbing-impact are presented. In fact there are so many sources of AE that the successful identification of rubbing-impact signal is very important. Account for the characteristics of acoustic emission signals the wavelet transform is employed to analyze the AE signal. The wavelet transform can decompose the AE signals in time and wavelet scale domains, and catch the differences in these waves. It enables to distinguish the rubbing-impact from other sources. It is concluded that AE offers earlier fault detection and improved identification capabilities than vibration analysis, allowing the user to monitor the rubbing-impact degrees of the rotor system, unachievable with vibration analysis.

This content is only available via PDF.
You do not currently have access to this content.