Recognition of Molten Pool Characteristics during High-power Fiber Laser Welding

Laser welding is an important processing techniques in industry field and facilitates the achievements of fast and deep welding in comparison with other traditional welding techniques pale in processing. Laser welding has so relatively high depth-width ratio that it enhances the quality of the welding seam joint.Because laser welding is not limited by special equipment requirements, so it can realize high automation in mass production. Optic fiber laser enjoys conspicuous advantages in terms of efficiency,light beam ability, weight and volume compared with traditional laser welding.During laser welding, two pieces of metals items absorb laser beam which have high energy density, consequently the weldment are melted and become joinable. The configuration characteristics of molten pool is closely related with the welding quality, through real-time monitoring the molten pool to determine the current state of welding. Because the spot diameter of laser beam is very small and it senses to the welding gap, so it is very important to monitoring laser beam deviation with welding seam accurately. The laser beam deviation with welding seam is closely related with the configurations of molten pool, the configurations of molten pool can reflect the deviation state of the welding seam indirectly. This paper recognized the object of high-power (10kW) fiber laser welding and observed and analyzed the changing characteristics of molten pool on the conditions of the laser beam was on the right position or deviated with welding seam center,which helped to determine the quality of the welding. For the features of high-power fiber laser such as strong metal vapor, high power density,rapidity,the observation of the features changes of molten pool image in case of welding seam deviation are based upon a sequence of dynamic molten pool infrared images that captured by high-speed visor sensor, which are combined with the analysis of image frequency and space domain. The configurations of a molten pool and a keyhole andseam width variation were analyzed through image processing techniques such as automatic segmentation by dynamic threshold and mathematical morphology and fourier transform and frequency domain filtering and Canny edge detection. The molten pool width, keyhole area, keyhole width and average gray of molten pool spectrum were defined as the feature parameters, and the welded seam width was defined as the parameter reflecting the welding quality. The laser welding experimental results and the analysis of the correlations among these parameters showed that the quality of a high power fiber laser welding process could be monitored by the defined configuration parameters of the molten pool,molten image features have the interrelation with welding seam deviation, which helps to lay the theoretical and experimental grounds for the high-power fiber welding. It provides a great significance for the automation and products quality improvement of laser welding.