New method of detecting abrupt information based on the singularity value decomposition and its application

A method of detecting abrupt information in the signal is presented, which uses the singularity value decomposition about the track matrix of attractor reconstructed by time series. Using this method, to detect is performed according to a numerical signal and a numerical simulating signal of the early rub-impact between the rotor and stator of rotating machinery. Through comparing its detecting results with ones gotten by using wavelet transform, the conclusion can be gained that this method is feasible and can be used to detect signal's singularity. Then a further investigation is done about the effect of the noise signal on the results of the presented method and the wavelet transform. Finally, this method is used to detect the testing signal of early rub-impact between the rotor and the stator of rotating machinery, and then the satisfying results are obtained.