Statistical Signal Processing and Sorting for Acoustic Emission Monitoring of High-Temperature Pressure Components

• Main Author: Hrairi, Meftah
• Format: Article
• Language: English
• Published: Wiley-Blackwell 2009
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://irep.iium.edu.my/6536/1/EXT418.pdf; http://irep.iium.edu.my/6536/; http://onlinelibrary.wiley.com/journal/10.1111/%28ISSN%291747-1567

A study was conducted to demonstrate feasibility of designing and developing a reliable and real-time monitoring methodology based on statistical pattern recognition for early detection of defects in small, low-alloy steel vessels. These low-alloy steel vessels were pressurized at high temperature with an aqueous hydrogen sulfide solution. A portable acoustic emission (AE) system was used to capture emission signals. An in-house-developed FORTRAN program calculated 37 features, including 18 from the time domain signal and 19 from the frequency domain. SAS statistical software package was used to find a correlation between these features, leading to the classification of AE signals according to type and discover relationships between emissions and the findings of the investigation. The specimens used in the study were made of low-alloy steel with similar characteristics to those used by petroleum refineries.