An Information-Statistical Approach to Analyzing Acoustic Emission Signals

Abstract

A new information-statistical approach is proposed that makes it possible to consider acoustic emission (AE) from the standpoint of synergetics based on representation of processes that govern the state of the object under study, a statistical model (image) in the form of a probability density function. It is shown that the multidimensional Dirichlet distribution has a set of properties that can be used to determine the integral metric for estimating the plastic deformation process from observed AE signals. As a quantitative measure in analyzing the plastic deformation process using AE signals, it is proposed to use the self-organization parameter. Regular mechanical tests of structural carbon steel 20 with a pearlite–ferrite structure were carried out to demonstrate that the information-statistical self-organization parameter is the most informative when describing AE-related processes.