Research of Acoustic Signals Digital Processing Methods Application Efficiency for the Electromechanical System Functional Diagnostics

Abstract

The article is devoted to researching of application efficiency of acoustic signals digital processing methods for the electromechanical systems (EMS) functional diagnostics in real time. The steps according to which it is necessary to carry out the analysis of the acoustic signals generated during operation of EMS for construction and adjustment of functional diagnostics systems are offered. The primary task is to study the statistical properties of acoustic signals. The next step is spectral analysis of acoustic signals. In the process of acoustic signal conversion after its hardware processing, the limits of the informative part of the signal are determined on the basis of spectral analysis based on Fourier transform of the signal. It is proposed to use logic-time processing based on the estimation of the normalized energy spectrum and spectral entropy to determine the frequency range of the informative part of the signal. The expediency of using the autoregressive model of the moving average as an acoustic signal model is substantiated. The problem of building mathematical models on the basis of which it is possible to adequately identify the intensity of work and the state of the equipment is solved. The procedure for identifying the parameters of the signal model by the recurrent least squares method is given, which allows to analyze the state of the equipment in real time. It is proposed to use multiplescale analysis to speed up the process of signal analysis and reduce the amount of calculations.

Method of the Ship Main Engine Condition Operational Diagnostics

“…are the actual normalized values of the time series, and as a result of sigma calculations according to formula (7) belong to the range satisfies the accuracy requirements.The next important task is the development of an algorithm for the diagnosis of the ship main engine based on the assessment of the normalized energy spectrum of the measured signals and spectral entropy. Diagnostics based on spectral entropy consists of a sequence of the following stages[11][12][13].ISSN 2313-4763 дд 139…”