FCAW process optimization using the multivariate mean square error

Abstract

Optimization of welding processes is not a trivial task, mainly due to the great number of required and desirable characteristics that must be analysed. Moreover, the optimization of a welding process with multiple characteristics without considering the variance–covariance structure may lead to inadequate optimum. To help with this task, a method of multi-objective optimization based on the multivariate mean square error applied in the study of multiple correlated characteristics of a flux-cored arc welding process is presented. This method characterized by a combined approach based on the response surface methodology, design of experiments, and principal components analysis consisted of an attempt to achieve the nearest values to specific targets, for each characteristic (penetration, deposition rate, deposition efficiency, convexity index of the weld bead, and dilution), considering the welding variables expressed as a result of welding voltage (V), wire feed speed (Va), and contact tip to workpiece distance (d). The results point, to a good adequacy of the proposed method.

Keywords:

• multivariate mean square error
• FCAW
• design of experiments
• response surface methodology

Acknowledgements

The authors would like to thank the support of Capes, CNPq, and FAPEMIG while carrying out this research.

Notes

^†Published with permission from ABCM – Brazilian Association of Engineering and Mechanical Sciences – Article originally presented at the 5th COBEF, Belo Horizonte, State of Minas Gerais (MG), April 2009