Optimizations of friction stir welding of aluminum alloy by using genetically optimized neural network

Abstract

Genetically optimized neural network systems (GONNS) was developed to simulate the intelligent decision-making capability of human beings. After they are trained with experimental data or observations, GONNS use one or more artificial neural networks (ANN) to represent complex systems. The optimization is performed by one or more genetic algorithms (GA). In this study, the GONNS was used to estimate the optimal operating condition of the friction stir welding (FSW) process. Five separate ANNs represented the relationship between two identical input parameters and each one of the considered characteristics of the welding zone. GA searched for the optimized parameters to make one of the parameters maximum or minimum, while the other four are kept within the desired range. The GONNS was found as an excellent optimization tool for FSW.