Prediction of bead geometry in cold metal transfer welding using back propagation neural network

Abstract

This paper presents the development of a back propagation neural network model for the forecasting the weld bead geometry (bead height and width) and penetration (depth and area) in cold metal transfer (CMT) welding process. Several welding parameters seem to influence the bead geometry and penetration. Typically, it is observed that high welding speeds or low heat inputs normally produced poor fusion. For this study, the model is based on experimental data. The input parameters considered are peak welding current, welding speed and heat input. The bead height and width, penetration depth and dilution area are taken as output parameters to design the framework of the model. These networks have achieved good agreement with the training data and have yielded satisfactory module. Neural network may effectively be implemented for estimating the weld bead and penetration geometric parameters. The results from the experiments indicate a small error percentage between the predicted and experimental values.