Optimization of FSW Parameters Using SA Algorithm and ANFIS-Based Models to Maximize Mechanical Properties of AZ80A Mg Alloy Joints

Abstract

This paper deals with the experimental research, modeling and parametric-based optimization of the mechanical properties of the friction stir welded AZ80A Mg alloy. A four-factor, five-level-based central composite design matrix was employed to minimize the experimental runs. Adaptive neuro-fuzzy inference system (i.e., ANFIS) was employed to map the relationship amid the parameters of FSW process (namely tool pin geometry, traverse speed, axial force, and rotational speed) and mechanical properties (including yield strength, tensile strength and hardness) of the joints. Later, the formulated ANFIS model was used along with simulated annealing (SA) algorithm determining the optimized parameters of FSW process so as to attain flaw free AZ80A Mg alloy joints. Formulated ANFIS model-SA algorithm anticipated that the friction stir welded AZ80A Mg alloy joints will possess a tensile strength of 240.52 MPa during the single-response optimization scenario and a tensile strength of 240.522 MPa during the multiple-response optimization scenario. Experimental results announced that the FSW process parameter combination of tool rotational speed of 1250 rpm, tool traverse speed of 1.75 mm/sec, axial force of 3 kN and tool possessing a threaded cylindrical pin geometry have contributed for attainment of largest values of mechanical properties during both the single-response and multiple-response optimization scenarios. During the confirmatory experimental work, the flaw free friction stir welded AZ80A Mg alloy joints exhibited a tensile strength of 242.16 MPa and the results of confirmatory experiment revealed that the ANFIS-SA system had exhibited superiority in modeling and optimization of the FSW process during joining of AZ80A Mg alloys.