A first arrival identification system of acoustic emission (AE) signals by means of a high-order statistics approach

One of the fundamental problems in non-destructive testing and geophysics is the precise determination of the first arrival of acoustic emission ( AE) signals ( or seismic signals) recorded by multi-channel systems. The knowledge of this time is very important, mainly in the case of automatic localization of individual AE events. Several approaches are routinely used in practice such as crossing of the threshold level, analysis of the LTA/STA ( long-time average/short-time average), etc. In our paper, an approach based on high-order statistics ( HOS) that is able to carry out precise arrival time determination without human intervention is presented. The approach was tested on real AE data recorded by an eight-channel recording system. This simple, accurate and quite fast method is predetermined to be used in automatic processing of transient waveform data from acoustic emission, seismic signals, ultrasonic sounding, etc.