Diagnosis of anisotropy and asymmetry in rotating machinery via directional spectral analysis

Abstract

The design of rotating machines of these days has a tendency to high capacity, light weight, and high operating speed so that the machines become more sophisticated than ever. As a result, unexpected excessive vibrations may be produced and eventually may result in faulty components. Therefore, early stage preventive diagnosis for the system faults should be required. In general, a rotor system consists of two subsystems: rotor and stator, which may have some degree of non-axisymmetric mechanical properties. The non-axisymmetric properties of stator and rotor are referred to as anisotropy and asymmetry and have significant influences upon the dynamic characteristics such as unbalance response, critical speeds and stability of rotor systems. Consequently, the initial strength and changes of anisotropy and asymmertry in rotor systems should be diagnosed. Any changes in system parameters will result in changes in the frequency characteristic. Based on this premise, a great deal of vibration monitoring and analysis techniques have been developed and applied to rotating machinery in industrial fields. Conventional diagnostic techniques treat the rotor vibrations as real quantity so that their frequency spectra lose their own meaning of directivity (i.e. forward and backward). On the other hand, in order to utilize the directional information in rotor vibrations, use of complex coordinates has received increasing attentions for the rotor dynamic analysis. Recently, the complex modal testing technique has been developed, whose key idea is to define the two-sided directional frequency response functions (dFRFs) between complex inputs and outputs. The complex modal testing allows clear physical insights into the backwared and forward modes, and also enables the separation of these modes in the frequency domain so that effective modal parameter identification is possible. The primary objective of this work lies in the development of frequency domain diagnostic methods of...