Research On Process And Control Method Based On Visual Sensing Of The MAG Weld During One-Side Welding With Back Free Formation

In this paper the clear pool images of MAG were obtained by passive visual sensing technology, the process of one-side welding with back free formation of 6mm Q235 was the core of this subject, and visual features and electrical parameter changes of MAG welding pool of different penetration were analyzed. A "Normalization- Discretization penetration model" of weld pool tail contours of one-side welding with back free formation was summarized, and a fuzzy control method based on this penetration model was put forward, this can provide related data and models for online control of welding quality.For 6mmQ235 steel plate one-side welding with back free formation, groups of experiments were designed, the experimental results showed that:in addition to the welding parameters, the accumulation of heat input had a lot of influence in the forming, so during welding the parameters should be changed constantly to limit the heat input.The geometry parameters of weld pool in different penetration state were summarized as follows:whenα>90°andβ>120°with the weld pool generally seemed like oblate, the weld seam non-penetrated;α<90°andβ<110°with the pool seemed like peach shape or oval,the weld seam penetrated; ifα<70°andβ<80°and the pool showed more slight, the collapsed happened.LabVIEW2010 was used as a software development platform to acquired and analyzed the electrical parameters during welding process. The conclusion:stable welding, its current and voltage waveform was consistent,αandβstability, theΔα< 5°; when welding penetration changed, the waveform of welding parameters and the welding instantaneous energy changed, too; for the test in this paper, a welding voltage peak> 40 V appeared after continuous short circuit means collapse would happen.The standardization curve of penetrated weld pool tail contours S was regarded as the model. Using "Normalization- Discretization" method to get the weld pool tail contours in experiments and the mean deviation e was obtained:e∈[-0.28,-0.13], collapse happened; e∈[-0.12,-0.08], the welding seam belongs to excessive penetration; e∈[-0.07,0.07],the weld seam penetrated; e∈[0.08,0.30], half penetration, e∈[0.31,0.6], non-penetrated. On this basis, a fuzzy control system for weld penetration formation was established, the mean deviation e as the input parameter while the neighboring 100 ms mean welding currentΔI as output parameter, which can promote the developing steps of welding quality online control.