AIChE Journal, Vol.63, No.9, 3692-3705, 2017
DEM-PBM Modeling of Impact Dominated Ribbon Milling
Ribbon milling is a critical step in dry granulation using roll compaction as it determines the properties of granules, and subsequently the properties of final products. During ribbon milling, fragmentation of ribbons or flakes (i.e., compressed agglomerates from dry powders) are induced by either impact or abrasion. Understanding these fragmentation mechanisms is critical in optimizing ribbon milling processes. In the current study, the discrete element method (DEM) was used to model fragmentation at the microscopic level, providing a detailed insight into the underlying breakage mechanism. In DEM modeling, virtual ribbons were created by introducing an appropriate interfacial energy using the cohesive particle model based on the JKR theory. A set of three-dimensional parallelepiped ribbons with solid fraction phi=0.7422 and surface energies ranging from gamma=0.03 J / m(2) and gamma=2 J / m(2) were created and then fractured during impacts with a plane at various impact velocities, to model impact dominated milling. The fragmentation rate, and the number and size of fragments (i.e., granules) resulting from the breakage of a ribbon during the impact were determined. The DEM simulations showed that the granules size distribution had a bimodal pattern and there was a strong correlation between the size of fines generated from fragmentation during impact and the size of the feed powder (i.e., the size of the primary particles in this study), which was consistent with the observation from physical experiments. Two quantities were calculated from the DEM simulations: the number of fragments p and the fraction of fines z for each breakage event which were then used as input parameters for population balance models (PBM) to develop a DEM-PBM modeling framework. Comparision with published experimental data shows that the developed DEM-PBM model is a promising tool for analysing ribbon milling, but all breakage mechanisms involved need to to considered in order to achieve an accurate prediction. (C) 2017 American Institute of Chemical Engineers