Predicting the Mechanical Properties of Friction-Stir-Welded AA7068-T6 Joints Using an Adaptive Neuro-Fuzzy Inference System

Soft computing techniques are now frequently employed in the industrial sector for process parameter modelling and optimization. Soft computing approaches produce good projected values that match the experimental data. In this study, a prediction model for the mechanical parameters of friction-stir-welded AA7068-T6 was established, including microhardness and ultimate tensile strength at the weld nugget. The models were created using the adaptive fuzzy inference system approach. For microhardness and ultimate tensile strength of weld nuggets, the generalized bell membership function yields minimal training g errors of 2.0494 and 0.84327, respectively. To test the expected adaptive fuzzy inference system output, the validation experiment was carried out at a tool rotation speed of 1400 rpm and a welding speed of 60 mm/min. The observed values for ultimate tensile strength and microhardness of the weld nugget obtained after the validation experiment were closer to the expected adaptive fuzzy inference system output.