Classification Of Characteristics Of Plume And Spatters Based On SVM During Laser Welding

Laser welding technology originated from the United States in the1970s, which has characteristics such as high-quality, high-precision, high efficiency, high-speed and low distortion. With the successful development of high power CO2and YAG laser, the application of laser welding technology are more and more widely. It is generally used in industry fields in industrialized countries. High-power disk laser welding was one of the advanced laser welding technology. Compared with other welding technology, it has advantages such as high laser power, excellent beam quality, high weld seam depth-to-width ratio and high laser utilization.During high-power laser deep penetration welding, there exist the plume and spatters which are the results of the ejection of materials from the area of the weld due to strong laser heating. The characteristics of plume and spatters are related to the stability and quality of welding and the utilization of the laser power. An approach of analyzing the characteristics of plume and spatters during high-power disk laser bead-on-plate welding of Type304austenitic stainless steel plates at a continuous wave laser power of10kW is proposed. An extra violet and visible sensitive high-speed color camera was used to capture the plume dynamic images. Plume area, laser beam path through plume, swing angle, distance between laser beam focus and plume image centroid, abscissa of plume centroid, spatter numbers are calculated by image processing, and defined as the eigenvalues. The characteristic parameters were pre-processed at first, and then principal component analysis was used to reduce the dimension of the characteristic parameters. To get the optimal parameters, particle swarm optimization was applied. Support vector machine was used to classify the plume and spatters. The fluctuation of weld bead width was used to evaluate the stability of welding quality. Training set and prediction got high classification accuracy. The rules and relationship between the characteristics of plume and spatters and the welding quality have been researched. Experimental results showed that the high-power disk laser welding stability could be evaluated by using the support vector machine to classify the characteristics of plume and spatters.