Printing parameters optimization assisted by machine learning and sintering behavior of binder jetting 3D printed 2024Al alloy

This study presents the results of printing parameters optimization of Binder Jetting 3D printed 2024Al alloy assisted by machine learning (ML). The effects of layer thickness and resolution on the printed 2024Al green parts including relative density, dimensional change and compressive strength were analyzed. The green parts printed with the optimized printing parameters were sintered, and their sintering behavior was analyzed. The results show the effect of layer thickness on the relative density and compressive strength of the green body is greater than that of resolution. However, the influence of resolution on the dimensional accuracy is higher than that of layer thickness. Based on the experimental results, the optimized printing parameters can be determined to be a layer thickness of 0.06 mm, a resolution of 1000*800, and a binder saturation range of 30%-70%, the relative density is 58.4%. For the few shot dataset, ANN is the optimized model to predict the printing process with the highest value of R2 above 0.92, and the calculated results agree well with the experimental results. The relative density of green part sintered in N2 atmosphere increases from 58.4% to 90%, and the optimized sintering temperature for 2024Al is 600 degrees C. The microstructure of sintered part is composed of alpha-Al, Al2Cu and a few of AlN, and the microhardness above 100 HV. AlN is more likely to form on the surface of the advancing liquid during pore filling, which is helpful for the densification.