Study of the Characteristics of Acoustic Emission Signals.

Abstract

This report aims to study the characteristics of signals collected by Acoustic Emission Testing (AET) and categorize them into types to detect yield. A literature survey of existing knowledge of AET and Acoustic Emission (AE) signal processing was conducted. In addition, data was processed by utilizing and developing tools to visualize and analyze the signal waveform based on signal duration and its corresponding amplitude. The investigation was done by exporting existing data from a three-point bending test that was used as a foundation for executing the analysis. Signal was visualized, and parameters were optimized to identify and classify signal types. The data used in this work was collected from the previous project in spring 2019 (Dawood_Khaled, 2019).

Results show that due to the length of the hit data recorded, some limitations on classification were experienced. However, it is possible to isolate signal characteristics that could recognize different signal types. Laboratory experiments should be performed to validate whether it is still feasible to detect relevant signals in a noisy environment such as an offshore platform before setting up a full-scale test setup. Further research into algorithms are recommended to increase the reliability of signal categorization.