A comprehensive characterization of metal powders for the quality assurance of additive manufac-turing processes is crucial for the industrialization of technologies like Laser Powder Bed Fusion or Laser Metal Deposition. The powder properties, e.g. the morphology, the rheology and the chemical composition, are of crucial importance for the quality of the manufactured components. The analysis of these three main powder property categories involves many different measurement methods and results in a large number of variables. A final assessment of the powder quality based on its properties remains very challenging. Especially when trying to understand the powder-process-product relationship, a multivariate analysis of the input variables is required.

In this approach, a multivariate data analysis is used to elaborate key variables which mainly attribute to the powder quality and to identify interconnections of the properties on multiple levels of the process chain. Tools like data mining methodologies such as machine learning can help to predict and optimize the process performance following the powder characteristics.

This contribution will introduces a multivariate analysis approach of metal powder properties for AM processes and shows potential application cases for novel powder qualification technologies via hyper spectral imaging and the process-specific characterization of powders with a challenging rheology using the example of microscale powders.