Implementation of hybrid DEM-PBM approach to reduce the computational cost of powder mixing modeling

The high computational cost incurred for implementing a computationally complex model such as discrete element method (DEM) to simulate the microscale particle interactions renders it grossly inefficient while developing models for real scale geometries and for simulating actual time length of powder processing unit operations. In this study, we examine an alternative computational cost saving methodology that utilizes the hybrid framework of coupled discrete element method (DEM) and population balance method (PBM) to allow for an extended particulate system analysis by extrapolating a limited DEM simulation through PBM methodology. This work focuses on the capability of DEM-PBM framework to not only simulate the dynamics of powder mixing but also to reduce the time-scale of computationally taxing DEM simulations. Through case studies of powder blending, which is a critical unit operation in pharmaceutical industries, in both continuous (bladed mixer) as well as batch (tumbling bin blender) setting, this study investigates the optimal DEM simulation length required to retain accurate mixing dynamics. The mixing dynamics are quantitatively recorded in terms of powder composition through relative standard deviation (RSD) studies. The predictive accuracy of PBM models derived from limited DEM simulations of various time-scales are also compared against a full DEM simulation to show the reduction in computational cost achieved by utilizing a DEM-PBM hybrid.