Statistical Vector-Tensor Analysis of Vibrations of a Centrifuge with Developed Defect of the Rotating Unit

Abstract

The structure of estimations of the linear and quadratic invariants of the correlation tensor function of stochastic vibrations of a centrifuge and their amplitude spectra are analyzed by the methods of coherent and component analyses for vector periodically correlated random processes. It is shown that the use of invariants enables us to improve the efficiency of diagnostics of vibrations, to determine the space variations of vibration loads in the case of appearance of a defect, and to localize this defect.