In order to reduce pollution of the marine environment and energy, aluminum ships is increase. In particular, FRP ships have been replaced by the aluminum ships. Aluminium alloy has high strength, low density, high corrosion resistance property at low temperature. In particular, Al 5083 alloy is mainly used high-speed ships and LNG tank.
In this paper, butt welding(6mm) type I using Plasma-GMA welding was carried out. And we were to evaluate Plasma-GMA hybrid welding characteristics of the Al 5083 aluminium alloy. The orthogonal experimental design was used to study the influence of plasma-GMA welding parameters. In order to evaluate factors(Plasma current, wire feeding rate, GMA welding voltage and welding speed) that effect bead geometry, we conducted ANOVA. Then we conducted evaluation of contribution for process parameters. Experimental results show that four bead dimensions(Bead width, Reinforcement height, Penetration Depth, Back bead width) in the Plasma-GMA welding process for a I-groove joint are affected by wire feeding speed, welding voltage and welding speed.
Three regression models and neural network model to estimate tensile strength was proposed. Estimation performance was evaluated through the coefficient of determination and average error rate and it was found that neural network model had the most excellent performance.
Optimal process variables were searched to maximize the fitness function which was defined with respect to tensile strength and GMA welding current which represented the weldablilty and welding speed which represented the productivity. As the optimization methodology, genetic algorithm were used and optimal welding condition, which meant plasma welding current, GMA welding voltage, wire feed rate, welding speed, was obtained for Plasma-GMA welding parameters.