Multi-scale modeling approaches for a spray coating process in paddle mixers
Abstract
As a common unit operation in the manufacturing industry, the spray coating processes involve many fundamental physical phenomena such as the mixing and segregation of particulate systems, spray dynamics and liquid contact spreading. This thesis work focused on modeling the particulate flow dynamics by developing a multi-scale approach to predict the coating uniformity of polydisperse particles sprayed in a Forberg style paddle mixer/coater. The approach consisted of two principal models: the discrete element method (DEM) for modeling the general flow features, and the population balance (PB) for simulating the multidimensional particle property distribution. First principles analysis based on DEM simulations and spray post-processing was performed to characterize the complex flow and mixing kinetics, especially through the spray region where particles receive the coating mass. The number of spray zone visits and surface area exposure percentage were found to significantly affect the coating mass distribution such that small particles receive more spray than expected. This trend was the opposite of that observed from fluidized bed or spouted bed coating systems, in which large particles were preferentially sprayed. The spatial component analysis at the mixing steady state showed the size segregation of large particles in certain regions in the bed was the main cause of the spray preference. It was also found that the mixing intensity had little impact on the mixing kinetics, but the extent of segregation could be reduced by increasing the paddle rotational speed. To study the coating mass distribution, a new two dimensional PB model was developed based on a compartment model scheme which incorporated the flow heterogeneity information from the DEM study. The results showed the coating mass coefficient of variance (CoV) ideally evolved with the inverse square root of the spray time for size independence growth, and deviated from the trend for size dependence growth. Increasing the seed particle size CoV0 or growth rate size dependence both caused an increase in the coating mass CoV beyond that predicted by one dimensional models. Sensitivity analysis on compartment model parameters showed significant influences from the bed characteristic time, bed recycle ratio and shortcut circuit fractions. Further parametric study also demonstrated an increase trend of coating mass CoV increased as Froude number Fr increased for Fr < 2. Finally, lab scale spray coating experiments were conducted to validate the model. From the unimodal series experiments, it was observed a similar power law relation of coating mass CoV with the spray time, and an increasing trend of coating mass CoV with the Fr value to what predicted from the DEM-PB model. The results of bimodal series experiments demonstrated a preferential spraying towards the relative small particle population, again as predicted by the model. The DEM-PB model is a powerful tool for design and optimization of spray coating in a Forberg style paddle mixer. The multi-scale approach taken is a template for more general application to other coaters and granulators.
Degree
Ph.D.
Advisors
Litster, Purdue University.
Subject Area
Chemical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.