A new hybrid meta-heuristic optimization method for predicting UTS for FSW of Al/Cu dissimilar materials

Abstract

Joining of aluminum (Al) and copper (Cu) materials used in aerospace, automotive, chemistry, nuclear, electrical-electronic and energy sector with traditional fusion welding methods has some negative consequences. One of the major negativities is the serious decrease in the mechanical properties of the materials as a result of the high temperature that occurs during welding. Therefore, Friction Stir Welding (FSW) is of great importance in joining materials which are dissimilar and have low melting temperatures. In this study, a hybrid artificial neural network (ANN) model has been developed to predict the relationship between FSW parameters and ultimate tensile strength and to increase the prediction performance of ANN using heuristic algorithms by welding Al and Cu plates with FSW. The used hybrid approach consists of ANN and Particle Swarm Optimization/Teaching Learning Based Optimization techniques. The welding applications conducted with three different welding parameters. These parameters, rotation speed (525, 1025, 1525 rpm), welding speed (50, 75, 100 mm/min) and tool offset (0-0.75-1.5 mm) were determined as input parameters of the developed ANN model. Tensile and microhardness tests were applied to determine the mechanical properties in the experimental studies after welding process. In addition, Scanning Electron Microscope analysis and X-ray diffraction were used for microstructure and characterization of the fracture surface and the results obtained were interpreted and examined.