Spatial green density variation and its effect on distortion prediction in binder jet additive manufacturing

Binder Jet Additive Manufacturing (BJAM) often encounters significant geometric distortions with part shrinking by over 15%. Advanced compensation models have been developed to mitigate these distortions, reducing final deviations to within 2-3% of target dimensions. A critical factor in these models is the green density of the printed part, which varies based on printing parameters and part location within the build. This study delves into the three-dimensional spatial variation of green density in BJAM, examining two prevalent powder spreading mechanisms: a fixed powder feedstock container with a vertically movable platform and cylindrical roller spreader, and a horizontally moving hopper for powder deposition. Experiments were conducted on three directionally scaled samples to assess the influence of geometric size on green density distribution. Analysis reveals at least 5% green density variations across different vertical positions ('part layers'), particularly in systems with extensive spreading distances due to the powder spreading and compaction process. Additionally, density discrepancies exceeding 10% were observed within parts of the same build in larger printing systems, underscoring the impact of spatial variation on the final properties of the manufactured parts. Incorporating spatial green density variations into finite element sintering models dramatically enhances prediction accuracy, reducing errors to within 1% (approximately 0.5 mm) of experimental results, representing a 75% improvement over uniform density assumption. This research highlights the necessity of accounting for spatial green density variations to achieve precise control over the final geometry of BJAM parts.