Artificial neural network modeling for evaluating of epitaxial growth of Ti6Al4V weldment

The effect of epitaxial growth on microstructure of Ti–6Al–4V alloy weldment was examined by artificial neural networks (ANNs). The microplasma arc welding (MPAW) procedure was performed at different currents, welding speeds and flow rates of shielding and plasma gas. Microstructural characterizations were studied by optical and scanning electron microscopy (SEM). Finally, an artificial neural network was developed to predict grain size of fusion zone (FZ) at different currents and welding speeds. sults showed that a coarse primary β phase develops in the fusion zone as a result of epitaxial nucleation on coarsened β grains near the heat affected zone (NHAZ) which grow competitively into the molten weld pool. Based on ANNs analyses, a map of current and welding speed for α → β transformation in the HAZ can be constructed. For a lower energy input, grain growth of β phase in the HAZ could be restricted by α phase. The presence of small quantities of this phase at high peak temperatures in the weld cycle is sufficient to prevent the grain growth of β phase in HAZ and FZ.