Metallurgical Phenomena Modeling in Friction Stir Welding of Aluminium Alloys: Analytical Versus Neural Network Based Approaches
Abstract
In this paper, the metallurgical phenomena occurring in friction stir welding processes of AA6082-T6 and AA7075-T6 aluminum alloys are investigated. In particular, to predict the local values of the average grain size, either a simple analytical expression depending on a few material constants or a properly trained neural network is linked to the finite element model of the process. The utilized tools, which take as inputs the local values of strain, strain rate, and temperature, were developed starting from experimental data and numerical results.