PREDICTIVE MODEL FOR ASSESSING DIAGNOSTIC SIGNIFICANT PARAMETERS OF ACOUSTIC EMISSION: MACHINE LEARNING EVIDENCE
АКУСТИКАЛЫҚ ЭМИССИЯНЫҢ ДИАГНОСТИКАЛЫҚ МАҢЫЗДЫ ПАРАМЕТРЛЕРІН БАҒАЛАУДЫҢ ПРЕДИКТИВТІ МОДЕЛІ: МАШИНАЛЫҚ ОҚЫТУ ДЕРЕКТЕРІ
Altay Y.A. Dosbaev Z.M. Altayeva A.A. Rakhmetova P.M. Absadykov D.B.
2025National Academy of Sciences of the Republic of Kazakhstan
News of the National Academy of Sciences of the Republic of Kazakhstan, Series of Geology and Technical Sciences
2025#2025Issue 58 - 21 pp.
Relevance. Acoustic emission (AE) systems and complexes offer a sensitive method for determining acoustic stress in rocks and geological fields, and also detecting various defects during selective laser melting of heat-resistant alloys. However, when operating this system, the AE signal recording process is inevitably subject to interference, which can significantly reduce the accuracy of signal parameter measurements. Therefore, to improve the accuracy of AE signal measurement, filtering methods are used to isolate diagnostic parameters from noise. Objective. The aim of this study is to improve the accuracy of AE signal parameter estimation by developing a digital filtering method and a phenomenological model of the information component. Methods. The proposed bi-directional filtering method for the AE signal model based on Butterworth digital filter is considered. The informative components of the signal are extracted, the parameters are measured and the relative measurement error between the signal models iscalculated. Furthermore, at the output of the digital filtering method, the signalto-noise ratio is computed to determine the association between this indicator and the measurement accuracy of diagnostic AE parameters. The relationship between the indicators is approximated using the least squares method and visualized by a scatterplot, which displays the distribution of data as points along the x-y coordinate. Results and conclusions. The implementation of the method improves the accuracy of measuring AE parameters, while the relative error does not exceed 3% compared to the Daubechies wavelet filter of the selected order and decomposition level.
acoustic emission , acoustic inspection , Butterworth filter , measurement accuracy enhancement , noise , signal processing , signal-to-noise ratio , wavelet filter
Text of the article Перейти на текст статьи
Satbayev University, Almaty, Kazakhstan
International Educational Corporation, Almaty, Kazakhstan
Astana IT University, Astana, Kazakhstan
Satbayev University
International Educational Corporation
Astana IT University
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026