The present research investigates the application of artificial intelligence tool for modelling and multi-objective optimization of friction stir welding parameters of dissimilar AA5083-O–AA6063-T6 aluminium alloys. The experiments have been conducted according to a well-designed L27 orthogonal array. The experimental results obtained from L27 experiments were used for developing artificial neural network-based mathematical models for tensile strength, microhardness and grain size. A hybrid approach consisting of artificial neural network and genetic algorithm has been used for multi-objective optimization. The developed artificial neural network-based models for tensile strength, microhardness and grain size have been found adequate and reliable with average percentage prediction errors of 0.053714, 0.182092 and 0.006283%, respectively. The confirmation results at optimum parameters showed considerable improvement in the performance of each response.
Artificial Neural Network Modeling of Surface Roughness in Magnetic Abrasive Finishing Process
