Statistical modelling of the narrow gap gas metal arc welding process

The J-laying technique for the construction of offshore pipelines requires a fast welding process that can produce sound welds in the horizontal-vertical position. The suitability of narrow gap gas metal arc welding (NG-GMA W) process for this application was previously demonstrated. The present programme studied the influence of process parameters on the fusion characteristics of NG-GMA welding in a range of different shielding gas compositions and welding positions. Statistical techniques were employed for both designing the experimental programme and to process the data generated. A partial factorial design scheme was used to investigate the influence of input variables and their interaction in determining weld bead shape. Modelling equations were developed by multiple linear regression to represent different characteristics of the weld bead. Transformation of the response variable based on the Cox-Box method was commonly used to simplify the model format. Modelling results were analysed by graphical techniques including surface plots and a multiplot approach was developed in order to graphically assess the influence of up to four input variables on the bead shape. Conditions for acceptable bead formation were determined and the process sensitivity to minor changes in input parameters assessed. Asymmetrical base metal fusion in horizontalvertical welding is discussed and techniques to improve fusion presented. At the same time, the interaction between the power supply output characteristic and the bead geometry was studied for narrow gap joints and the effect of shielding gas composition on both process stability and fusion of the base metal was assessed. An arc instability mode that is strongly influenced by arc length, power supply characteristic and shielding gas composition was demonstrated and its properties investigated. An optimized shielding gas composition for narrow gap process was suggested.