Prediction of Single Weld Bead Size of TIG Welding Arc Additive Manufacturing

Orthogonal test of GH4169 nickel-based superalloy was carried out with the tungsten inert gas (TIG) welding arc additive manufacturing platform. Effects of welding current, welding speed, and wire feeding speed on weld width and reinforcement were studied. BP artificial neural network was established and optimized by genetic algorithm, the single weld bead shape and size prediction model was obtained, and a human interactive interface for size prediction was created by the graphical user interface (GUI) module of MATLAB. The results show that the weld width increased with the welding current, decreased with increasing welding speed, and first increased and then decreased with increasing wire feeding speed. The weld bead reinforcement was linearly positively correlated with the wire feeding speed, and negatively correlated with welding current and welding speed. The prediction relative error of the weld bead size prediction model was within 6%, which could predict the shape and size of a single weld bead more accurately.