- Journal of Artificial Intelligence and Data Science
- Volume:2 Issue:1
- A Novel Approach to Improve Tensile Strength of Al/Mg Hybrid Friction Stir welding Joint by Stochast...
A Novel Approach to Improve Tensile Strength of Al/Mg Hybrid Friction Stir welding Joint by Stochastic Optimization
Authors : Onur AYDIN, Elif ARAS GÜLTÜRK
Pages : 31-42
View : 14 | Download : 6
Publication Date : 2022-06-30
Article Type : Research Paper
Abstract :Ultrasonic-stationary shoulder-assisted friction stir welding is a novel hybrid welding technique that reveals promising prospects in joining Al/Mg dissimilar alloys. This study aims to develop a design procedure for optimizing the mechanical property of the Al/Mg hybrid friction stir welding joint. For this purpose, firstly, different nonlinear neuro-regression analysis has been performed in order to overcome insufficient approaches for modeling, designing, and optimizing mechanical property in Friction stir welding joint. Then, stochastic optimization methods were performed to model the friction stir welding process. Ultrasonic Power, Welding Speed, and Rotational Velocity are the three most essential criteria that have been used as indicators of process performance. The response characteristic can be predicted as ultimate tensile strength. After calculating the R_training^2, R_testing^2, 〖and R〗_validation^2 values, the limits of the nonlinear models are examined to see whether the model is acceptable for optimization. The best approach model was the second-order trigonometric multiple nonlinear insert ignore into journalissuearticles values(SOTN); model. In the optimization step, four different Modified Stochastic Optimization Algorithms, including Random Search insert ignore into journalissuearticles values(MRS);, Simulated Annealing insert ignore into journalissuearticles values(MSA);, Nelder Mead insert ignore into journalissuearticles values(MNM);, and differential equations insert ignore into journalissuearticles values(MDE); methods, were used. It has been observed that the different scenario types and the constraints chosen for the design variables are effective in the optimization results obtained using three different scenarios. Results showed that the maximum tensile strength was 182.301 MPa when ultrasonic power was selected as 186.938 W, 40.6854 mm/min for welding speed, and 1075.34 rpm for rotation speed.Keywords : Friction stir welding, neuroregression analysis, tensile strenght, stochastic optimization