Optimization of process parameters for direct energy deposited Ti-6Al-4V alloy using neural networks

Abstract

Direct energy deposition (DED) is a highly applicable additive manufacturing (AM) method and, therefore, widely employed in industrial repair-based applications to fabricate defect-free and high degree precision components. To obtain high-quality products by using DED, it is necessary to understand the influence of the process parameters on the product quality. The optimization of such processing parameters provides several advantages such as minimization of the loss of material and time. However, the optimization of the complex relationship between the process parameters-geometry-properties of the fabricated sample is difficult to realize and requires significant experimentation. Herein, a computational model based on artificial neural networks was developed to optimize process parameters for DED-processed Ti-6Al-4V alloy. The model was developed to estimate the density and build height (the actual build height realized after fabrication) of the sample as a function of power, scan speed, powder feed rate, and layer thickness. The optimum model with high-accuracy predictions was employed to construct the process maps for the DED-processed Ti-6Al-4V. The relative importance indices of process parameters on the build height and density were investigated. Further, the effect of power and scan speed on the microstructure of Ti-6Al-4V alloy was discussed. Finally, based on the obtained results, the optimum fabrication conditions for the DED-processed Ti-6Al-4V alloy were determined.