Affiliation of Author(s):机电学院
Journal:WCCM - World Congr. Cond. Monit.
Abstract:The detection of cracks in rotor blades is an important problem from concern about the safety and reliability of gas compressor. Blade tip-Timing(BTT) vibration measurement technique has been widely used in rotating blades vibration monitoring. The sampling frequency of BTT, which is just a few hundred Hertz, is once per revolution and changes with different speed, far less than vibration frequencies of blades which can reach up to several thousand Hertz. The measured data by blade tip-Timing is nonuniform and under-sampling time series. The synchronous frequencies are integral multiples of rotating speed. Some of them are sensitive to formation of cracks in the blade. In this paper we present the identification method of blade synchronous frequencies based on the nonuniform Fourier transform of multi sampling data. Firstly the multi sensors data is interpolation in order to increase the Nyquist frequency. Then the Nonuniform Extended Discrete Fourier Transform(NEDFT) is performed and the blade vibration frequencies can be achieved by observing that the synchronous vibration frequencies are necessarily present in all spectrum diagrams based on their characteristics. Both of the simulation and experimental data applied to this method have verified it correctly.
Translation or Not:no
Date of Publication:2017-01-01
Co-author:l
Correspondence Author:Yue Lin
Date of Publication:2017-01-01