Effect of heat treatment on the microhardness of Ti-6Al-4V alloy fabricated by laser powder bed fusion and its prediction using machine learning

Laser Powder Bed Fusion process of Additive manufacturing has evolved as a game-changing technology for creating highly sophisticated components for automotive, aerospace and medical applications. The impact as well as contribution of LPBF parameters, such as laser power as a source of energy, exposure time and hatch space in developing microhardness for Ti-6Al-4V alloy is explored using a Taguchi-based design of experiments. By employing an orthogonal array, the number of required experimental runs were reduced to nine by ensuring efficient parameter optimization. This study examines the post-processing effect of heat treatment on microhardness of Ti6Al4V alloy specimen, measured using Vickers hardness testing machine. Results obtained have demonstrated a notable hardness improvement after annealing which attributed to microstructural homogenization, reduced porosity and defect mitigation. Among the heat- treated samples, laser power emerged as the most influential parameter, with sample 4S (350 W laser power, 40 µs exposure time, 0.09 mm hatch spacing) achieving the highest microhardness of 543 HV, representing a 20.9% increase over as-built conditions. In the second phase, machine learning algorithms were implemented to develop predictive models of microhardness. The highest accuracy achieved by Random Forest with R2=0.91 and Cat boost with R2= 0.93 for as build and heat-treated samples respectively. Significant insights into process optimisation and property prediction of Ti-6Al-4V produced using LPBF are offered by this integrated experimental - computational approach.