Statisticals models of powder spreadability in powder-bed-based additive manufacturing

• A test rig is developed to assess powder spreadability in additive manufacturing. • Statistical analysis links powder properties to spreadability. • Multiple linear regression allows the prediction of powder layer density. • Powder layer density is highly correlated with static powder apparent density. • Powder free-flowability is not a relevant indicator of powder spreadability. Powder spreading is an essential and common step in Powder-Bed-Based Additive Manufacturing (PBBAM) processes. However, the effect of powder properties and spreading parameters on the powder layer quality is not well understood. This study gathers characterization data on 75 steel powders to study the correlations between particle size and shape, apparent density, free-flowability, and ultimately spreading behavior. To do so, a novel instrumented apparatus is developed to investigate both powder layer density and powder layer roughness in a representative environment. Then, a characteristic subset of 28 powders from the complete dataset (exhibiting both spherical and irregular shapes) was further tested with two different spreading tools. Based on the entire collected dataset, Multiple Polynomial Regression analyses were conducted on powder apparent density (PAD) and powder layer density (PLD) to achieve both explanatory and predictive modelling, yielding respective R 2 values of 82 % for PAD and 90% for PLD.