Development of an Artificial Neural Network Model to Predict the Tensile Strength of Friction Stir Welding of Dissimilar Materials Using Cryogenic Processes

Abstract

The objective of this study was to develop an artificial neural network (ANN) model for predicting the tensile strength of friction stir welding (FSW) joints between dissimilar materials, with a particular focus on aluminum and copper, using cryogenic processes. The research addresses the challenges posed by differences in material properties and the complex nature of FSW, where traditional experimental methods are time-consuming and costly. FSW experiments were conducted under a variety of conditions, and the resulting temperature data were utilized as input for a heat transfer analysis. The maximum temperature and temperature gradient obtained from the analysis were employed as input variables for training the ANN. The ANN was optimized using the Hyperband tuner and validated against experimental results. The model successfully predicted tensile strength with an average error of 5.4%, demonstrating its potential for predicting mechanical properties under different welding conditions. This approach offers a more efficient and accurate method for optimizing FSW processes.

Keywords

• friction stir welding
• dissimilar materials
• aluminum alloy
• copper alloy
• finite element method
• tensile strength