Acoustic Emission Testing
Continuous On-line Monitoring
Acoustic emission testing is usually conducted as a short term, well-defined event involving a planned change in pressure or temperature. However, there are cases where this is impractical or undesirable for one reason or another. In such cases, acoustic emission monitoring during normal operation is an option for detecting process-related damage, as it occurs.
The benefit of what we might call a “standard” acoustic emission test is that the additional loading stimulates acoustic emission activity from any defects and concentrates it over a defined time span. This simplifies the analysis process and is a major factor in the success of the current practices and standards for testing of metal and composite tanks and vessels.
On-line monitoring is generally conducted over days, weeks or months and does not require load or temperature changes although these are beneficial. Apart from avoiding the need to interrupt the process, a major benefit of on-line monitoring is that it can show any association between crack growth and operating conditions. Acoustic emission data, therefore, can help with optimizing the operating conditions to minimize equipment damage.
Until recent years, on-line monitoring was confined to few, specialized applications. Costs were high due to the need for AE personnel to remain on-site and data analysis was a difficult and time-consuming task. The advent of better computers and software, high-speed modems and more reliable instrumentation has made remote monitoring a reality. The AE specialist can now operate the AE system and evaluate on-line data remotely and report results by e-mail.
Even with the advent of faster computers, evaluation of
on-line data remains a challenge. The
majority of
recorded data is background noise and identification of significant
AE activity from the often-large body of noise takes skill and some new analysis
tools. Matrix is meeting this
challenge with innovative analysis techniques built around pattern recognition
technology. This methodology
separates different data types and, over time, produces a classifier or template
to identify the different AE sources. This
allows rapid separation of background noise from signals of interest. Also, any new sources or types of AE activity then stand out and are
evaluated in depth for any characteristics of cracking or other damage.
This technique not only speeds up the analysis process after the initial learning cycle but also makes the data analysis more consistent and focused. Our experience is that each monitoring set up has unique characteristics. By building classifiers from the data itself rather than using pre-set patterns we reduce the chances of false calls and make sure any new patterns of activity are recognized and evaluated.
Whole vessels may be monitored in this way but typical
applications so far are for areas of interest and involve
comparatively few
channels (8-16) of AE instrumentation. Entire
vessels or long lines may still be tested by moving the sensors from section to
section. This is because
eliminating the one-time loading sequence also removes the need to instrument
the entire vessel or piping system.
On-line monitoring lends itself to a wide range of process equipment but it is probably best suited to those where there is some cycling of temperature or pressure or an aggressive chemical environment. We see this technology as another option for keeping process equipment running and will be pleased to discuss your particular application.
Last updated:
September 06, 2002