Prediction of Weld Bead Geometry of Mild Steel Using Taguchi Technique and Multiple Regression Analysis

William E. Odinikuku *

Department of Mechanical Engineering, Petroleum Training Institute, Warri, Nigeria.

Joseph E. Udumebraye

Department of Mechanical Engineering, Petroleum Training Institute, Warri, Nigeria.

David Atadious

Department of Mechanical Engineering, Petroleum Training Institute, Warri, Nigeria.

*Author to whom correspondence should be addressed.


Abstract

The weld bead geometry is very important in predicting the quality of weld as cooling rate of the weld depends on it. For this purpose, the Taguchi technique was applied to determine optimum process parameters of weld bead geometry in submerged arc welding. The study involves using Taguchi’s L9 orthogonal arrays to conduct nine experiments on a 6 mm plate of IS2062 grade mild steel by using SKU MIL-SubArc AC/DC submerged arc welding machine with constant voltage. Three-levels of the four process parameters- arc voltage, welding current, welding speed and electrode stick out were considered and their effect on weld bead geometry−bead width, depth of penetration and weld reinforcement was observed. The signal to noise ratios was computed to determine the optimum parameters. From the results obtained, optimum process parameters of ,  and  was suggested for weld bead width, weld penetration and weld reinforcement respectively. Regression analysis is done to establish the relationship between the input parameters and geometrical parameters of weld bead. The proposed mathematical model can be used to predict bead width, weld penetration and weld reinforcement values for any given SAW welding conditions.

Keywords: Submerged arc welding, taguchi technique, multiple regression analysis, weld bead geometry.


How to Cite

Odinikuku, William E., Joseph E. Udumebraye, and David Atadious. 2020. “Prediction of Weld Bead Geometry of Mild Steel Using Taguchi Technique and Multiple Regression Analysis”. Journal of Engineering Research and Reports 13 (4):31-46. https://doi.org/10.9734/jerr/2020/v13i417111.

Downloads

Download data is not yet available.