부산대학교 도서관

Monitoring equipment in real time has become increasingly important, mainly when operational safety is targeted. Acoustic emission (AE) testing has been applied to the inspection of various types of equipment, particularly rigid pipes. This paper presents a study on the use of acoustic emission to detect defect propagation in pressurised rigid pipes. The resulting AE signals were classified as no propagation (NP), stable propagation (SP) and unstable propagation (UP) and used as inputs in the implementation of non-linear classifiers by error back-propagation. The correct classification results reached close to 91%, proving the efficiency of the method in the conditions tested in this study. The methodologies used for the construction of the probability of propagation (POP) curve are presented, which are a great innovation in this research field and the focus of an international patent. [ABSTRACT FROM AUTHOR]