Classification of Elastic Wave for Non-Destructive Inspections Based on Self-Organizing Map

An arrival time of an elastic wave is the important parameter to visualize locations failures and/or velocity distributions in field non-destructive testing (NDT). The detection conducted generally using automatic picking algorithms a measured time-history waveform. According algorithms, it expected that detected from low S/N signals has accuracy if are measurements. Thus, order accurately detect for NDT, classification waves required. However, based on not been extensively conducted. In this study, method self-organizing maps (SOMs) applied classify waves. SOMs relation data wherein number classes unknown. Therefore, SOM selects high and adequately validated model tests pencil lead breaks (PLBs), was confirmed successfully consisted signal. Moreover, classified were source localization noteworthy localized sources more accurate comparison with all