A novel acoustic monitoring method of laser peening

The acoustic monitoring method of laser peening has aroused a great deal of interest for realizing the non-destructive detection and real-time monitoring. Considering the special characteristics of the acoustic signal generated in the laser peening, we have explored a new signal processing method for feature extraction, which is based on the intrinsic relationship between the adjacent data in the signal series. During the implementation, we measure the cosine similarity of neighboring subsequences for presenting that interrelation. After investigating the variation of the cosine similarity in the series, evident features could be observed visually, and useful characteristic value could be extracted to assist in identifying the working state. As a result, the effectiveness of the new method is well verified and its application potential is worthy of expectation.