A Research On Welding Parameters In Narrow-Gap Submerged-Arc Welding

Narrow-Gap Submerged-Arc Welding(NG-SAW), due to its higher melting rate, smaller Heat-Affected Zone and better mechanical properties of its welding metal, has been expanding its application domain and scope. After what welding materials to use is decided, the parameters in the welding process have decisive effects on the shape of the weld metal, its chemical composition and microscopic structure. In this research, the effects of the important process parameters selected from NG-SAW on the shape of the weld metal have been studied. At first, four important factors have been selected from the parameters that affect the shape of the weld metal:arc voltage(V)、feed speed(F)、 welding speed(S) and the nozzle to plate distance(N). In this experiment, the above four factors have been studied. After analyzing the welding suitability of the welding material and several trial runs, the scope of the factors have been decided. Using the Response Surface Methodology(RSM) in the Design of Experiments(DOE), a experiment with four factors and five levels in each factor was designed. And the experiments were performed using the NG-SAW welding machine adapted from traditional SAW machine in our laboratory. The results of the experiments were analyzed with the help of computer application named Design Expert8.0. The effects of process parameters on the shape of weld metal and the mathematic model to predict weld metal shape were got. Neural networks were also used to establish a model to predict weld metal shape. It used process parameters as inputs, the shape of the weld metal as outputs. With software Matlab7.0a Back Propagation(BP) neural network was established and its efficiency to predict the shape of the weld metal has been tested。 The results showed that the BP neural network has relatively good self-studying ability and predicting ability. It is of some value in practical application of predicting the shape of weld metal in NG-SAW.