Prediction of Melting Profile of Mild Steel Weld Metals Using Regression Analysis

Abstract

The microstructure of a weldment can be maintained by ensuring a steady state homogenous melting profile of the welding operation, which includes the deposition of optimum volumes of melted filler wire and substantial part of the heat affected zones of the parent metal to form the weld pool. The melting pattern of the entire welding process should be protected from atmospheric air, so as to enhance weldment quality. In this study, the melting profile of mild steel is investigated by looking at the parent metal angular distortion bead width and penetration volumes of deposited filler and the melting efficiency was determined. Predictive models were also developed to determine the above listed melting properties by applying the regression analysis. The result obtained showed that, there is almost a perfect fit between the calculated and predicted angular distortion, as well as between the calculated and predicted volume of filler wire melted. There is also a close correlation between the calculated and predicted melting efficiency. However, for the bead width, bead penetration and volume of weld metal deposited, there were variations of values and a heterogeneous correlation between the calculated, measured and predicted values. The effects of the process parameters on the obtained properties of the melting profile were investigated and optimum process parameters were determined.