A novel Bayesian approach to acoustic emission data analysis

Area of Science:

• Materials Science
• Non-Destructive Testing
• Statistical Signal Processing

Background:

• Acoustic emission (AE) is a valuable technique for materials characterization and non-destructive testing.
• AE signals originate from the stochastic motion of defects, making them inherently random time series.
• Identifying trend changes in fluctuating AE data is a significant analytical challenge.

Purpose of the Study:

• To propose a general Bayesian framework for analyzing AE time series.
• To develop an automated method for detecting breakpoints in AE data.
• To identify crossovers in the dynamics of underlying AE sources.

Main Methods:

• Development of a general Bayesian framework.
• Application to AE time series analysis.
• Automated breakpoint detection algorithms.

Main Results:

• The proposed framework enables automated identification of breakpoints in AE time series.
• Successfully detects changes in the dynamics of AE sources.
• Provides a robust method for analyzing complex AE data.

Conclusions:

• The Bayesian framework offers an effective solution for analyzing AE time series.
• Automated breakpoint detection enhances non-destructive testing capabilities.
• This approach improves the characterization of material defect behavior.