Cинтез двовимірних діагностичних параметрів при коваріаційного аналізу тривимірних вейвлет-перетворень вібросигналів

Abstract

Results of the kovariatsionny analysis of continuous veyvlet-transformation of spectral and non-stationary vibrosignals are presented. or from scale. The model of piecewise and linear approximation of any of sections of the three-dimensional veyvlet-image by sequences of regressions with the casual coefficients which dispersive analysis allows to allocate Fischer's four criteria two-dimensional statisticians who depend either on shift, or on scale is offered. Reflected that as informative for problems of vibration diagnostics are F-statistics that depend on the scale. The resulting F-statistic, as random variables with known distribution law, allow you to go from a qualitative assessment to quantitative measurement characteristics and evaluate the probability of error diagnosis based on standard parametric models.