Research On Weld Defect Based On LabVIEW Ultrasonic Testing System

In recent years,nondestructive testing technology has been widely used and developed in industrial testing.Among them,ultrasonic nondestructive testing technology in manufacturing,construction,metal processing,chemical industry and other fields that need defect detection and quality control has the advantages of simple and convenient operation,non-destructive,no damage to human body,and can detect various types of defects,It can be used for both laboratory research and field engineering testing.However,the traditional ultrasonic testing in the quantitative and qualitative aspects of defects is very dependent on the experience and judgment of testing personnel.For example,in the weld inspection of metal based materials,the inspectors can only judge the type of defects through the past experience,and can not make qualitative analysis,At the same time,in the process of inspection,the ambient noise will also have a great impact on the test results.Therefore,the traditional ultrasonic nondestructive testing technology is difficult to meet the development and requirements of modern testing technology.In this paper,a study of weld defects based on Lab VIEW ultrasonic testing system is proposed.Under the premise of realizing the traditional detection of weld defects,the ultrasonic echo signal can be collected and processed.The main work and innovation of this paper are as follows:(1)An ultrasonic weld nondestructive testing system based on virtual instrument Lab VIEW is proposed and built,which combines hardware with software.Compared with the traditional ultrasonic testing equipment,there are some shortcomings,such as high maintenance cost,single function and so on.The testing system based on virtual instrument can not only meet the traditional functions,but also adjust and improve according to different testing materials and testing environment.(2)The system adopts the combination of hardware and software,and uses the hardware equipment such as data acquisition card and DPR500 and Lab VIEW software to realize the acquisition,storage,processing and analysis of ultrasonic echo signal data,so as to improve the real-time and effectiveness of data acquisition,so as to facilitate the processing and analysis of subsequent data.After that,the whole system is tested and verified,The experiment shows that the system runs well and can provide a good data source for the detection and qualitative analysis of weld defects.(3)In the process of defect echo signal processing,We use db5 wavelet de-noising method to de-noising the echo signal of weld defects,and use MATLAB software to generate the noise signal.Through simulation analysis,we get the optimized wavelet de-noising method,so as to improve the signal-to-noise ratio of the system and achieve higher detection requirements.Then we select the appropriate wavelet basis function and decomposition scale,decompose the echo signal by wavelet packet,extract the feature value of defect energy after noise reduction,judge the feasibility of defect energy signal as feature information,analyze the weld defects,and provide the basis for BP neural network recognition of weld defects.(4)BP neural network is built to train and test the data in the existing database,The experimental results show that the overall recognition rate of three typical weld defects(porosity,slag inclusion and crack)is 82.05%,which indicates that it is feasible and recognizable to build BP neural network for qualitative analysis of weld defects.