Application Research On Fault Diagnosis Of Turbine Rotor With Wavelet Analysis

Turbine is essential equipment in industrial rotating equipment, and the rotor of turbine, as the most important part of turbine, has been paid more and more attention on its reliability, availability, maintainability and security issues, which puts forward higher requirements on a failure mechanism and diagnostic techniques. The vibration signals of the equipment contain important features which could reflect the working status of the rotating equipment. Through analysis of the characteristics of the vibration signals,we could judge whether working status of the equipment is proper, which has an important significance on protection of personal safety and improving production effective.In this paper, firstly, we analyzed the fault types and corresponding main characteristics of the turbine rotor, which may occur during the operation and reviewed the methodology and principles of fault diagnosis of the rotor rotating equipment. We also grasped how to reach operating mode by the study of the vibration signals and picking up characteristic signals. We applied ENTEK integrated condition monitoring (ICM) to monitor and analyze turbine rotor failure in one oil refinery and pointed out some limitations in Fourier methods in analyzing the vibration signals.Secondly, to address the error diagnostics when the vibration feature is not obvious because of being buried in the noise signals. We adopted the wavelet analysis which was Cutting-edge technology and we mastered the technique of wavelet analysis, which included decomposion, filter, reconstruction and analysis by referring to a large of literature. Finally, we adopted wavelet analysis to carry out fault diagnosis using Matlab. We also compared wavelet analysis with Fourier and concluded that when the characteristics of the vibration signals were not prominent, the wavelet analysis has superiority.