Design Of Stress Wave Nondestructive Testing System Based On LabVIEW

This paper designes the stress wave nondestructive testing system which combines the advantages of high speed acquisition of CTS04-PC acquisition card、the advantages of wavelet packet transform and ensemble empirical mode decomposition analysis for nonstationary signals、the method of feature slection baseing on the combinating inter class distance with floating sequence forward selection、the flexibility of support vector machine classification and semi supervised learning and the superiority of LabVIEW programming software.The main research contents are as follows:The first a model of defect feature representation and feature selection is built.To maximize the use of defect information and fully reflect the characteristics of the nature of the defect,we extract the structure characteristics on time domain and frequency domain、transformation characteristics on WPT and EEMD for stress wave.A feature selection method is proposed based on the combination of the average distance between classes and the forward selection.The second based on the research of defect recognition and semi supervised incremental learning model,the defect recognition model based on supervised learning is constructed,and the known defects are classified.And in view of the traditional SVM need to label the training data,and how to use a small number of labeled samples and a large number of unlabeled examples to improve the learning performance of the semi supervisor learning model was studied.The last we built the stress wave nondestructive testing system and carry on the flaw on-line recognition experiment.Finally the experiment results verify the reliability of the system.