Optimization and prediction of process parameters in laser powder bed fusion of Ti-6Al-4V alloy using hybrid Taguchi-grey relational analysis and artificial neural network

This study investigates the optimization and prediction of arithmetic average roughness (Ra) and hole dimensional accuracy of implant plates fabricated from Ti-6Al-4V alloy using laser powder bed fusion (L-PBF). Grey Relational Analysis (GRA) and an Artificial Neural Network (ANN) were employed to optimize the process parameters and predict the corresponding responses. The investigated process parameters were laser power (175, 185 and 195 W (W)), scanning speed (600, 900 and 1200 mm per second (mm/s)) and hatch distance (70, 90 and 110 μm (μm)). A Taguchi L9 orthogonal array was used to design the experiments. The responses considered were top flat surface roughness (TFSR), internal circular curve roughness (ICCR), external hole diameter (EHD), internal hole diameter (IHD) and volumetric energy density (VED). The GRA identified the optimum combination as a laser power of 175 W, scanning speed of 1200 mm/s, and hatch distance of 70 μm. An experiment conducted using these optimum parameters produced response values of 3.25 μm, 15.58 μm, 0.2767 mm (mm), 0.3093 mm and 53.26 J per cubic millimeter (J/mm 3 ), respectively. The ANN predicted corresponding values of 3.32 μm, 16.12 μm, 0.2834 mm, 0.3176 mm and 55.08 J/mm 3 , respectively. The prediction errors were below 3.5%, demonstrating good agreement between the experimental and predicted results. The findings confirm that the hybrid GRA and ANN approach is effective for optimizing process parameters and predicting Ra and hole dimensional characteristics in L-PBF of Ti-6Al-4V alloy.