Abstract

This article presents a method for the analysis of diagnostic acoustic signals for defect detection in composite materials. For this purpose, a multiple-scale decomposition of the acoustic signal and the determination of informational entropy for its components are used. It is established that, in the presence of small defects, the distribution of the power spectral density is little different from the power spectral density for a defect-free region. As a criterion for defect detection, a linear correlation coefficient is selected for vectors composed of informational entropies of multiple-scale components of the acoustic signal.

中文翻译：

---

用于复合材料诊断的声信号分析

摘要

本文介绍了一种分析诊断声学信号以检测复合材料中的缺陷的方法。为此，使用了声信号的多尺度分解和确定其分量的信息熵。可以确定的是，在存在少量缺陷的情况下，功率谱密度的分布与无缺陷区域的功率谱密度几乎没有差异。作为缺陷检测的标准，为由声信号的多尺度分量的信息熵组成的向量选择线性相关系数。