Artificial neural network model for predicting the tensile strength of friction stir welded aluminium alloy AA1100

Abstract

Friction stir welding (FSW) is a solid state welding technique, in which high strength weldswith minimal defects, can be obtained even with materialsthat are hardly weldable by conventional techniques. FSW is influenced by a number of process parameters. Some of the highly influential process parameters that determine the quality of the welds in FSW are tool rotation speed, welding speed, shoulder diameter and pin diameter of tool. In this study, FSW trials were conducted on AA1100 as per central composite design, with four parameters varied at five levels. The tensile strength of the joints were measured using a computerized tensile testing machine and these results were used to develop an artificial neural network model. The input parameters to the model were tool rotation speed, welding speed, shoulder diameter and pin diameter and the output was tensile strength of the joints. Levenberg Marquardt algorithm was used to establish the relationship between the process parameters and the output. The feed forward model was trained using 80% of the experimental data and the remaining 20% of the data was used for validation and testing of the model. The R2 valuesfor validation data and testing datawere found to be 0.80and0.99respectively, displaying the closeness between the experimental and predicted data. The results indicate that the generated model has high efficacy in predicting the tensile strength of friction stir welded aluminium alloy AA1100 joints.