Prediction of heat generation effect on force torque and mechanical properties at varying tool rotational speed in friction stir welding using Artificial Neural Network

Abstract

Friction stir welding (FSW) has played a significant role in joining aerospace alloys. During this process, the tool rotational (TRS) speed has been found to significantly affect heat generation compared to other parameters. Therefore, the study has investigated the effect of heat generation on force-torque and mechanical properties at different tool rotational speeds (TRS) in the FSW process through experimentation followed by Artificial Neural Network (ANN) technique. Further, the influence of different TRS ranging between 600 and 1800 rpm with an increment of 400 rpm on considered responses; namely thermal weld cycle, microstructure, and grain distribution in nugget zone (NZ) for 2050-T84 Al-Cu-Li alloy plates, welded using FSW were also investigated. It is observed that the vertically downward force (Z-force), longitudinal force (X-force), and spindle torque (Sp. T) decrease with increasing TRS. It is also observed an increasing (up to 1400 rpm) and then decreasing trend for tensile strength and hardness of welded samples. Moreover, the generation of frictional heat and grain size in NZ is increased with increasing TRS from 600 to 1800 rpm. However, the scanning electron microscope (SEM) micrographs of all-welded samples revealed a ductile mode of tensile fracture. Furthermore, the obtained experimental results were validated using the ANN technique. A quite better agreement has been established among the predicted outcomes from ANN with experimental results.