Study On Laser Ultrasonic Defect Signal Processing Method Based On Wavelet Packet Analysis

Laser ultrasonic non-destructive testing technology is one of the current research hotspots.It is suitable for detecting in harsh environment and has been widely used in various industrial fields.In order to evaluate the defects of the objects,the post-processing and analyzing of the laser ultrasonic signals is becoming an important research content in laser ultrasonic surface defect detection field.In this paper,we mainly study the laser ultrasonic defect signal based on the wavelet packet analysis method..Firstly,we summarized excitation mechanism and surface acoustic wave of laser ultrasonic and introduced the composition and working principle of laser ultrasonic detection system,then described the experiment flow in detail and several laser ultrasonic surface defect signal processing methods.Secondly,we also introduced the principle and steps of the wavelet denoising method and the wavelet packet denoising method while also discussed about the selection of wavelet parameters.By comparing the denoising effects of the two methods on the analog signal,the wavelet packet denoising method has a good self-adaptability,and it recognizes and separate the signal and noise components effectively.The SNR improvement factor is 5.0042 dB about,which improves the wavelet denoising method by 0.6683 dB and enhances the performance by 13.37%.At the same time,we analyzed the reflected signal in frequency domain based on the wavelet packet denoising.By extracting the feature quantity and comparing with the simulation signal,the result shows that the frequency domain characteristics of the experimental signal coincides with the finite element simulation signal,and the validity of the experimental detection defect and the effectiveness of the wavelet packet denoising are also verified,which lays the theoretical foundation for the later research.Finally,basic theories of wavelet analysis,multiresolution analysis and wavelet packet analysis are briefly introduced.Based on these theories and the algorithm of energy feature extraction in different frequency bands,we extracted the energy distribution characteristics of ultrasonic defect signal of different defect crack depths.The results show that the defect depth of 0.5mm as the dividing line,and the signal energy distribution on both sides is different,which provides a reference for laser ultrasonic surface defect detection.