Study On Welding Cracks Detection By Acoustic Emission

Welding has become an indispensable process in manufacturing various kinds of mechanical equipment as well as in building important engineering structures. However, in some circumstances, there might be some problems with welding, i.e., cracks may occur in weld. The reasons for weld cracking are related to many factors, e.g., the base metal used, the welding method, the parameters of welding process, the shape and the thickness of the welding component,etc. The nature of acquiring real time process phenomena relies on successful process control and quality detect.Acoustic emission is a non-destructive testing used to represent essence of physics phenomenon. The course of weld's arc burning, torch transition, metal distortion, phase change, crystal shape, freezing, motion of dislocation and crack produce and spread can release energy and create acoustic emission accompanying in parts. So through analysis the collected acoustic signals we can inspect weld quality. It suits to forecast accident of weld structure, judge the degree of detect, entire inspect quickly, reduce the cost and etc. AE detecting technology is being widely applied to real time analyze and inspect of material distortion and rupture.Based on a lots of technology literature at home and abroad, the paper concludes the theory and application of acoustic emission, practically researches the weld crack inspect with AE technology. Research stressed on acoustic emission produce and spread principle, and studied the wavelet decomposing and wavelet reconstruction, and designed the hardware and software of the dynamic weld crack monitoring system which based on virtual instrument. The hardware of the system constituted by micro-computer, AE sensor, magnify and filter circuit, A/D data acquisition card. The application software system built by LabVIEW virtual instrument software and engineering math software Matlab. Contemporary utilized the modernized signal process methods and technology.The system realized module functions of signal acquisition, frequency analyse, wavelet decomposition, wavelet reconstruction, power spectrum, AE characteristic parameter extracting. And through collecting and comparing the AE signals of welding cracks of cast iron and 45 steel under different restraint welding conditions, found different crack signal characteristics. Validated the validity of weld crack signal frequency mainly in the range of 110kHz～190kHz and the filter noise by wavelet technic. Primary found some phenomena and rules help to go deep to study the weld crack, and explore a path to locale engineering application.