Predicting the effect of tool configuration during friction stir welding by cellular automata finite element analyses

Abstract

The present work aims at selecting the pin geometry for single-sided friction stir welding of steel, and modelling the double-sided FSW process, both through cellular automata finite element analyses. Cellular automata cells are used to hold initial grain size of the base sheets. The stirring action during the process is not modelled physically. Instead the heat flux and strain-rate attained are incorporated by models. In the first objective, by developing heat flux models, strain-rate and strain models for different pin profiles, and the final grain size distribution across the weld zone is predicted. The final pin profile predicted from the present work is not the same used by authors in the existing literature, as observed from grain size prediction. In the second objective, during double-sided friction stir welding, the grain size predictions are agreeing well with the existing data. The efficiency of the approach for tool selection has been demonstrated successfully. [Submitted 8 December 2016; Accepted 5 December 2017]