ACOUSTIC EMISSION PATTERN RECOGNITION ANALYSIS APPLIED TO THE OVER-STRAINED PIPES IN A POLYETHYLENE REACTOR

Techniques of acoustic emission (AE) were used during the over-straining applied on straight pipes and elbows of a polyethylene reactor operating in a factory. This reactor has several dozen pipes, straight and bent, of 10-m length. Over-straining is done before the pipes were put in operation and also periodically during service. AE enables us to monitor the stability of the deformation process and indicates the zones, which will require further detailed examinations by other NDT techniques; it should also improve safety during this process through early detection of critical damage. After a few series of tests and using evaluation criteria with AE parameters, we examined next 45 pipes with recorded waveforms for additional analysis. Vallen VisualClass pattern recognition analysis program enabled us to separate the signals due to the plastic deformation of materials during over-straining process from the signals generated by cracking damage. The presence of damages in three groups of over-strained pipes was identified and confirmed by other NDT techniques.