A study on a vision sensor based laser welding system for bellows

Welded metal bellows are commonly manufactured by welding pairs of washer-s haped diaphragms of thin sheet metal. Two types of welding operations are r equired: inner edge welding to make pairs of diaphragms, and outer edge wel ding to form a bellows by welding the number of pairs. In this study, the inner and outer edges of bellows were welded by using a CW Nd:YAG laser. For the inner edge welding, the Nd:YAG laser beam was dire cted at an angle of 45 degrees to the edge because of the limited accessibi lity. In the outer edge welding, welded pairs were fixed on a jig and compr essed axially. The difference between the center of the bellows and that of a jig shaft causes an eccentricity, while the tolerance between a motor sh aft and a jig shaft causes a wobble-type motion. A vision sensor based on t he optical triangulation was used for seam tracking. Two image processing a lgorithms that can find a bellows edge were developed for inner and outer e dge welding. A geometrical modeling that describes the difference between t he position of the Nd:YAG laser spot on a bellows edge and that of the lase r diode stripe of the vision sensor was established. The seam tracking by a two-axis Cartesian robot was performed after the vision sensor determined the positional deviation in each sampling. Experiments were performed with 0.1 mm thick stainless steel bellows and 0.5 mm outer edges pitch. The seam tracking by the sensed positional deviation and geometrical modeling was p erformed successfully with an acceptable error.