## Abstract

Fixed frequency content (i.e. component or structural resonances) in spectra obtained from rotating equipment can be masked by the strong sources at harmonics (order) of the shaft running speed. The ratio of the fixed frequency components to the order components can be greater than 60dB making interpretation of resonances in the spectra difficult. Hence the order components are viewed as a corrupting phenomenon. An approach to remove the order components from the spectra, without affecting the remaining frequency domain information is presented in this work. The technique utilises a sequence of data sampling and transformations, between the time, order and frequency domains as follows:(1) Vibration data is sampled using a constant time basis (Δt).(2) The times corresponding to a constant angular basis (Δθ) are determined.(3) The vibration data is interpolated to a constant angular basis (Δθ).(4) The constant angle sampled data is transformed via the FFT to the order domain.(5) The high amplitude order components are now exactly bin centred and can be removed from the spectra.(6) An inverse FFT is applied to return to a constant angular increment sampled (Δθ) array, sans order content.(7) The constant angular increment sampled (Δθ) array is interpolated to an array sampled with constant time basis (Δt).(8) A FFT is applied and then standard spectral estimation procedures used to compute the vibration spectra with the high-level orders removed. The theoretical and implementation details of the double resampling approach are discussed. The approach is applied to experimental torsional vibration data acquired from a laboratory test rig designed to simulate a turbine rotor. The test results show that the method can recover fixed frequency components (i.e. turbine blade natural frequencies) in the presence of order components 50dB higher.

Original language | English (US) |
---|---|

Pages (from-to) | 483-500 |

Number of pages | 18 |

Journal | Mechanical Systems and Signal Processing |

Volume | 19 |

Issue number | 3 |

DOIs | |

State | Published - May 1 2005 |

## All Science Journal Classification (ASJC) codes

- Control and Systems Engineering
- Signal Processing
- Civil and Structural Engineering
- Aerospace Engineering
- Mechanical Engineering
- Computer Science Applications