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

Published: 17-09-2020| Version 1 | DOI: 10.17632/sxg3hg3txw.1
Joel Torres-Serra


In Torres-Serra et al, we propose a methodology based on cluster analysis of data sets including quantitative flow descriptors of a wide range of powders and grains, with application in the packaging industry. Our work has implications for objectifying the commonly qualitative process of selecting the most suitable bulk feeding technique in handling equipment design. The first data set DS1 (‘DS1.csv’) describes 174 materials characterised by 6 conventional material properties. The second data set DS2 (‘DS2.csv’) describes 11 representative materials, fully-characterised by 126 conventional and specialised variables. Numbering of the specialised variables in DS2 identifies average measurements of up to 20 new specialised material properties for 6 different test cases, as discussed in the associated article. The tables in ‘legends.pdf’ detail the legend of variables in the two data sets DS1 and DS2, comprising material property symbols and descriptions used in the associated article. The interactive MATLAB® figures ‘classes.fig’ and ‘clusters.fig’ show observations in DS1 projected into a reduced 3D space defined by PCA, corresponding respectively to figures Fig. 8a (feeder-type classification from industrial know-how) and Fig. 8b (flowability-based clustering) in the associated article.