A process optimization strategy of a pulsed-spray fluidized bed granulation process based on predictive three-stage population balance model

Abstract

In this work, a three-stage population balance model (TSPBM) was developed for a pulsed top-spray fluidized bed granulation (FBG) process. The batch top-spray FBG process was divided into three stages based on the analysis of granule properties evolution and different granulation mechanisms were considered in each stage of the TSPBM. In each stage, population balance model (PBM) describes the evolution of granule size distribution (GSD), and partial least square (PLS) regressions describes the relationship between the operating variables and kernel parameters in PBM. By fitting coefficients of PLS regressions using experimental data, the developed TSPBM establishes a predictive relationship between the manipulated binder spray parameters of pulsed frequency, binder flow rate and atomization pressure and granule critical quality attributes (CQAs). A model-based multi-stage optimization strategy was proposed to improve the granule quality of the pulsed-spray FBG. The optimization strategy reduced the process error caused by mismatch between developed model and actual system. In the optimization strategy, volume average granule diameter is only measured online at the end of each stage. Based on the online measurement, the optimization strategy adjusts process operating variables to remedy any drift between measured and predicted GSD. Validation experiments and simulation tests were carried out to validate the effectiveness of the proposed TSPBM and optimization strategy. The developed TSPBM is shown to accurately predict experimental GSD and shows high accuracy comparing to the one-stage PBM. The proposed optimization strategy can improve the prediction capability by >50% compared to an offline optimization.