Study Of The On-line Paraxial Monitoring During Fiber Laser Welding

High-speed train industry has been developed rapidly in recent years. The market has a huge demand for stainless steel body because of its beautiful appearance, corrosion resistance and low cost. Welding is the main process method in the manufacturing process of the high-speed train. Compared with traditional spot welding, laser welding of stainless steel body has obvious advantages in welding precision, deformation, welding quality and welding efficiency. However, the welding quality is affected by various factors, and the change of penetration in stainless steel body lap welding could cause vehicle quality problems. So it is extremely necessary to monitor the weld quality on line in laser welding.First of all, a paraxial fiber laser welding process monitoring system has been built up. Yellow-Green Emission(500-600nm) and Near Infrared Emission(900-1000nm) emitted during the fiber laser welding process have been picked up as detected parameters. The signals can be sampled with high speed and storage automatic. System can provide a variety of signal analysis means to achieve the extraction of signal characteristics. The real-time monitoring system has high stability and reliability. It can also realize the alarm to the welding quality fluctuation and the statistics of the welding defect.The influence of laser power, welding speed and defocusing amount on monitoring signal had been studied systematically, and the dynamic behavior of plume had also been investigated with the aid of high-speed cameras. It had been found that the increase of laser power and defocusing amount could lead to amplitude increase of Yellow-Green emission and Near Infrared emission in time domain. But the increase of welding speed had an opposite effect. In frequency domain, laser power had a little influence on signal? frequency components. Moreover, the decrease of welding speed and the increase of defocusing amount could make signal energy gathering at low frequency. According to the high-speed video images, the increase of welding speed has nothing impact on the volume of plume. Instead, the increase of defocusing amount could lead to volume increase of plume, and the dynamic behavior of plume became more stable.Finally, signal features of Yellow-Green Emission and Near Infrared Emission had been extracted when welding defects appeared, such as misalignment and gap. It had been found that welding defects could decrease the heat radiation intensity of plume which made the amplitude of Yellow-Green Emission and Near Infrared Emission reducing rapidly in the time domain. The Yellow-Green signal? low frequency could reflect the existence of misalignment, but it was not sensitive enough to identify butt gap in the frequency domain by using wavelet analysis. Its signal component(near 1600Hz) could also reflect the size change of lap joint gap. Near Infrared signal could clearly reflect all kinds of quality problems because its signal components in the frequency domain and time domain had good consistency.