The spectrum analysis technology and feature extraction of rotor system steady and speeding-change process

Abstract

High-precision spectral analysis techniques are the important means to carry out research of rotor system stability. It can be achieved feature extraction parameters about rotor system steady-state and spectrum of vibration signals during speeding-change process. For a big calculation error problem existed in commercial software come with spectral analysis in the status, It is proposed that the spectrum analysis of proportional interpolation correction method and feature extraction technology combined with methods of time-space domain transformation and time-domain refinement, which used to resolve the distortion problem between vibration spectrum analysis results and the real spectrum during speeding-change process, and to obtain high precision spectral feature parameters. The results show that: the proportion of interpolation method can accurately extract the spectral feature parameters of the rotor system steady-state process, the maximum frequency error can be controlled at 0.06%, the maximum amplitude error can be controlled at 0.24%, the maximum phase error can be controlled within 1 degree; Especially, It can significantly improve the identification accuracy of spectral feature parameter for speeding-change process using high-precision spectral analysis techniques based proportional interpolation method, and combining with methods of time-space domain transformation and time-domain refinement into the sampling, such as the maximum amplitude error can be controlled at less than 15%.