Generalized Recurrence Quantitative Analysis And Its Application In Gear Fault Diagnosis

Recurrence plot is a classical method to characterize time series recursion qualitatively through binary image.Recurrence Quantitative Analysis(RQA)quantitatively studies recurrence plot by extracting distribution statistics of recursion points in recurrence plot.However,the traditional recurrence quantitative analysis only extracts the horizontal,vertical and diagonal distribution rules of the recurrence point,and does not describe the whole and local characteristics of the recurrence plot.In this paper,a generalized recurrence quantitative analysis method based on principal singular value and local detail spectrum is proposed.The Principal Singular Value Proportion(PSVP)is used to describe the overall recurrence plot in the core part of the amount of information,Local Detail Spectrum(LDS)used to depict recurrence figure limited range recursion point in Local distribution.In this paper,the generalized recurrence quantitative analysis used for gear fault diagnosis,the experiments show that compared with the traditional recurrence quantitative analysis,generalized recurrence quantitative reliability for gear signal classification recognition rate is higher,stronger.