Multivariate modeling to evaluate effect of fabricating process on product profile of high shear granulation

Abstract

ObjectiveTo investigate the feasibility of high shear granulation (HSG) for agglomeration of Crataegi Folium extract (Chinese edible herbal), and explore the effect of process variables on granule critical quality attributes (CQAs) by three different models, i.e., response surface methodology (RSM), multilayer perceptron neural networks (MLP), and partial least squares method (PLS). MethodsRSM, MLP, and PLS complementary to design of experiment (DoE) were utilized to investigate the influence of granulation time, impeller speed, and binder amount on the products. Crataegi Folium extract was employed as granulation powder. RSM was further carried out to develop design space of HSG. ResultsThe results indicated that RSM, MLP, and PLS modeling techniques enhanced the understanding and controlling of granules produced via HSG. The granule CQAs were mainly influenced by granulation time, impeller speed, and binder amount. Overlay plots of the RSM indicated that design space for the operating ranges of impeller speed and binder amount at high levels of granulation time was the smallest. ConclusionThe study showed that these models were useful to characterize the granulation process, and was particularly important to understand the process.