Neural Network and Genetic Algorithm Based Modeling and Optimization of Tensile Properties in FSW of AA 5052 to AISI 304 Dissimilar Joints

摘要

A prominent benefit of friction stir welding (FSW) process is to join sheets with dissimilar material. In such condition, the mechanical properties of dissimilar joints are highly affected by FSW parameters. In the present work, an attempt is made to find optimal parameter setting of tool rotary speed, welding speed and tool offset regarding maximum tensile strength and elongation for AA 5052 and AISI 304 dissimilar joints. For this purpose, firstly an intelligent correlation between mentioned factors and tensile properties was developed by using neural network. Then, the developed network was integrated with genetic algorithm to find optimal solutions to achieve desirable mechanical properties. Furthermore, the obtained result is verified by conducting confirmatory experiment. Results indicated that settings of 500 RPM tool rotational speed, 80 mm/min traverse speed and 2 mm tool offset causes maximization of both tensile strength and elongation. Also, this result was then discussed based on FSW process mechanism.