Coherent and component methods for mean and covariance function estimation are analyzed using linear filtration theory.
The coherent estimators of probabilistic characteristics for periodically correlated random processes in the case of unknown period are analyzed. Shown that these estimators are asymptotically unbiased and consistence.
A new approach for the fault detection of the turbo-set friction bearings, based on the investigation the periodically non-stationary properties of vibration signals, is analyzed.
Method for separation of the periodically correlated random process into harmonic series representation is presented. The correlation and the spectral properties of the stationary harmonics are investigated. The expressions for precision of separated harmonics are worked out. Simulation results for the amplitude-phase modulated non-stationary signal are… (More)
Theoretical and experimental results of time series modeling are summarized. Special attention is given to periodically correlated random processes (PCRP). It is suggested to use the parametric modeling methods on the basis of PCRP decomposition on stationary processes for construction of the process model. Proposed parametrical model is helpful for… (More)