Identification of Natural Frequencies in Bearing Mechanical Vibrations by Wavelet Decomposition

In this paper is shown the results of a research study regarding the identification of natural frequencies in bearing mechanical vibrations by wavelet decomposition. This research work is intended to show clearly and didactically the use of the free software GNU Octave in order to provide with not only a theoretically background of the problem to be solved. The intention is to verify the effectiveness on how to program code in GNU Octave to implement wavelets efficiently in order to find the natural frequencies and damping coefficient of the mechanical vibrations of a bearing. In this research study are used some datasets about the vibration of a bearing in order to find the natural frequencies and damping of the mechanical vibrations yielded by it. The procedure is to obtain the wavelet transform of the data and to obtain the required bandwidth in order to extract the damping and natural frequencies of each dataset. It is important to mention that the wavelet decomposition is useful for this purposes by implementing a minimal post-processing of the results. It is worthy to mention that besides the wavelet implementation is performed a Fourier and Bode analysis in order to verify and validate the results presented in this research study. Some numerical experiments are performed with the respective conclusions and discussions.