摘要

Vibration-based condition monitoring of rotary machinery develops toward application for highly non stationary operational conditions, mainly rotational speed and load. If speed and load vary during the acquisition of vibration signal, both, instantaneous amplitudes and instantaneous frequencies of signal components vary accordingly. This scenario causes a strong demand for novel signal processing approach suitable for analysis of such processes. In this paper, angular-temporal spectrum is introduced as a novel and practical solution for analysis of such signals. The tool is based on well-established principles of angular-temporal determinism, and allows for representation of energy of analyzed signal on bi-frequency plane related to angular and temporal properties of the signal. Additionally, proposed approach includes amplitude normalization technique in order to give more interpretable measures of amplitude related to varying speed and load. The paper gives overall description of the proposed tool as well as its intuitive explanation from the perspective of signal generation mechanism. The outcome of the angular-temporal spectrum is presented using generated signal, vibration signal measured on lab test-rig operating with damaged bearing during run-up process, and finally on industrial data from a wind turbine.

  • 出版日期2016-8