Observations and process parameter sensitivities in fluid-bed granulation

A numerical approach using mechanistic modeling (population balance) for fluid-bed agglomeration is documented elsewhere. Here important factors for a pilot-scale fluid-bed granulater as determined by a 1/2 fraction factorial statistical design of experiment (DOE), Monte Carlo analysis of the mechanistic model, and comparison of experimental observations to model predictions, are summarized. DOE results indicate sensitive primary process parameters affecting granule size include the bed bowl charge, binder spray rate, air flow rate, and input air temperature. Sensitivity analysis of model results suggest the binder spray rate accounts for approximately 65% of the predicted variance, with the next most sensitive parameter the binder droplet diameter (similar to11%). Population balance modeling specifically for fluid beds, coupled with a DOE can be combined so the behavior of agglomeration of wettable powders within a fluid bed can be understood and predicted. Good agreement between experiment and simulation results is found for particles grown to sizes of industrial importance.