Friction Stir Welding Flaw Evaluation Of Aluminum Alloy By Laser Ultrasonic

Friction stir welding is a rapidly developing non-melting weldingmethod in recent years which is characterized by simple process,non-pollution and excellent mechanical properties. In practicalapplications, the unreasonable process parameters of welding speed,stirring head shape can cause typical friction stir welding defects, such ashole size, root defect and internal "Zigzag" and so on. These defects arealways subtle, sticky with complex orientation. The traditional detectionmethod has the limitations of low sensitivity, high quality requirementsof surface, which influenced the reliability of detection results.Laser ultrasonic testing is a new non-destructive testing method thatrapidly developed in recent years, which involves many subjects such asacoustic, optical, thermal, material science and has advantages of widefrequency band, non-contact, various wave excitation, which can detecttiny and be suitable for harsh environments.The main purpose of this paper is to detect the actual defects offriction stir welding of aluminum alloy using laser ultrasonic, explore thequalitative and positioning method for different defects. Based on studyof FSW defects forming mechanism, three kinds of typical defects suchas aluminum alloy2219tunnels, non-welded root,"Zigzag" wereobtained through different welding parameters. The transient sound fieldmodel of laser generated ultrasonic waves in materials containing different defects were established through using simulation software. Andthen the influence of experimental methods, different defects size,position and different laser ultrasonic methods on defect signals isobtained. Combined with the ultrasonic detection platform, differenttesting methods are firstly employed to detect and analyze the defects andthe corresponding positions. Secondly, perform the detection signalamplitude analysis in time domain and frequency domain, using powerspectrum density for analysis and feature extraction. And do theclassification using support vector machine to provide methods fordetection and classification of defects in comprehensive evaluation.