Yuzefovych R. Modeling and statistical analysis of joint periodically nonstationary vibration signals for the detection of mechanisms faults.

The spectral and correlation theory of joint periodically correlated random signals is developed, the features of representation of their probabilistic characteristics in the form of Fourier series are defined, the relation between these series and with characteristics of random processes, which describe amplitude and stochastic modulation of vibration signal main harmonics, is analyzed. Coherent and component methods of cross-correlation analysis are generalized for the case when the period of nonstationarity is unknown. The vibrodiagnostic system, algorithms and software for the analysis of stochastic oscillations with different origination on the base of real data, are created and verified.