Classification of granular materials via flowability-based clustering with application to bulk feeding

Abstract Feeder selection impacts the performance of bagging machinery throughout its life cycle, and yet it is usually based on qualitative assessments of flowability. We propose a data analysis methodology aimed at verifying the feeder-type classification of powders and grains by cluster analysis on their material properties. Results for a first data set of conventional properties show the granular materials clustered into as many groups as main bulk feeding systems. Mismatch between feeder classes and flowability-based clusters is explained by common industrial practice and incomplete material characterisation. For this reason, we introduce a set of specialised properties measured with the granular flow tester we have recently developed. Results for principal component analysis on a second extended property data set show that similarly flowing granular materials are better detected considering the specialised properties. This research contributes to objectify the decision-making process of bulk feeder selection from the quantitative description of granular flow.