Extended hybrid statistical tools ANFIS- GA to optimize underwater friction stir welding process parameters for ultimate tensile strength amelioration

Abstract

The qualities of functional parts produced by underwater friction stir welding (UWFSW) with additive water are significantly reliant on standard FSW process parameters. To improve the goal function, hybrid statistical tools can be used to optimize operation parameters. This work investigates the tensile strength(σUTS) of tests ASTM D638-14 specified parts manufactured using UWFSW by Al AA 6063-T6 material. Three parameters were varied in the fabrication of test specimens: speed of rotation from 1000 to 1800 rpm, speed of traveling from 4 to 10 mm/s, and shoulder diameter from 10 to 20 mm. Using a Hybrid artificial neural network- genetic algorithm (ANN-GA) and Hybrid artificial neural network-fuzzy-genetic algorithm (ANFIS-GA). The ANFIS-GA achieved the highest precision of 98.99 %, resulting in optimum parameters like rotational speed 1800 rpm, travelling speed 4 mm/s, and shoulder diameter 15 mm to produce a maximum tensile strength of 266 MPa. The hybrid models developed could be used to predict and maximize specific process parameters and impacts for a variety of industrial situations.