Фізичне моделювання та діагностика попадання сторонніх предметів в обертову систему

Abstract

The paper analyzes the vibroacoustic signals obtained by physical modeling of the rotating system, for example, an aircraft gas turbine engine, in the conditions of steady-state and non-steady-state modes. An air starter (supercharger) is used as a physical model of a rotating system, which is driven by a DC motor. The measuring system uses a dynamic microphone with an amplifier, a tachometer, a two-channel digital oscilloscope, a personal computer with technological and special software. The simulation of the ingress of foreign objects into the rotating system is performed by throwing paper balls during the rotation. The multilevel processing based on sequential application of methods of frequency-time analysis, multispectral analysis, and fractal analysis is proposed and substantiated for processing of measured vibroacoustic signals. The results of the frequency-time analysis showed that at the time of throwing the balls the intensity of the components at higher frequencies increases. For fragments of signal realization without throwing and with the throwing of balls the multispectral analysis is carried out and estimates of the bispectrum modulus are received in the form of contour images. At the third level of signal processing, the Minkowski dimension of the contour images of the bispectrum module estimates is determined. The Minkowski dimension is an integral quantitative indicator of the geometry of isolines and differs in value for the selected fragments of the vibroacoustic signal. So it can be used as a diagnostic sign of a foreign object entering the rotating system at the final level of processing. The obtained results can be used to improve the systems of condition monitoring of complex rotating systems, increase sensitivity, expand functionality and provide multi-class diagnostics in the event of damage and violation of normal operating modes