Time-frequency domain analysis of vibration signals for machinery diagnostics (II) The weighted Wigner-Ville distribution

This report examines the application of sliding weighting functions to the Wigner-Ville distribution. The effect of the weighing function on the time-frequency distribution is discussed and it is shown that the level of the interference terms, the production of aliasing errors and the repetition terms in the weighted Wigner-Ville distribution are significantly reduced. It is suggested that the Gaussian function provides a suitable form of weighting function. A digital computer program implementing the discrete weighted Wigner-Ville distribution is described and its performance is demonstrated by the analysis in the time-frequency domain of a series of numerically generated test signals. It has been shown previously that the Wigner-Ville distribution may be applied to the analysis in the time-frequency domain of experimentally-measured time domain averages of the vibration of gears in industrial and helicopter gearboxes, in order to detect early signs of impending mechanical failure. In this report it is shown that the damage to the gears can be detected by visual inspection of the changes which occur in the pattern of the weighed Wigner-Ville distribution more easily than they could with the original unweighted Wigner-Ville distribution.