Research On Health Inspection Of Composite Plate Structure Based On Ultrasonic Guided Wave Technology

Composites are widely used for their excellent performance and are prone to defects during the manufacturing and service phases due to various environmental factors.To avoid human casualties and economic losses due to potential defects,health inspection for composite materials has important practical significance.Ultrasonic guided wave has the characteristics of long propagation distance,sensitivity to defects,and high detection efficiency,and is an inspection tool with great potential for development.In this paper,structural health inspection research based on ultrasonic guided wave inspection technology is conducted for composite plates,and the main research works are as follows.1.a finite element model of the composite plate is built using ABAQUS simulation software,and a sensing array is set up on the surface of the plate structure,and a guided wave excited by a Hanning window modulated signal with a center frequency of 160 khz is loaded for numerical simulation.Delamination and crack defects with changing geometric parameters are preset on the constructed simulation model,respectively,and the detection signals are collected to form a signal sample library of multiple defect features.2.The data dimensionality reduction methods of principal component analysis and linear discriminant analysis are introduced to extract the signal features to construct the defect feature matrix as input,and the classification models of support vector machine and Light GBM after parametric optimization and training are realized to identify the types of defects,delamination and major geometric dimensions of cracks.Based on this,the performance of each discrimination model is evaluated using accuracy,precision,recall,and F1 score.3.The basic principles of the probabilistic damage imaging method are described in detail.Based on the results of the machine learning models for the identification of two defect geometric parameters,an improved algorithm for probabilistic damage incorporating defect features is proposed.The improved algorithm mainly fine-tunes the beta value in the probabilistic damage algorithm and sets weight factors for the sensing paths associated with the defect geometric parameters to realize the feature characterization and localization imaging of defects.4.The basic principle of ellipse imaging method is introduced in detail.The damage factor of each path is analyzed and calculated by Pearson correlation analysis,and the paths are screened for multi-path fusion imaging based on the damage factor to narrow the area where the defects exist in the plate.Based on the screened paths,a Bayesian approach is used to fuse the two imaging methods to achieve the prediction of the horizontal and vertical coordinates of the defects and the defect localization imaging.