Artificial neural network modelling of continuous wet granulation using a twin-screw extruder

Transparent predictive modelling of the twin screw granulation process using a compensated interval type-2 fuzzy system

“…Since the granulation process can determine the 'fate' of the produced tablets, several studies have been devoted to understand and model and simulate such a process (Paavola et al, 2013;Saleh et al, 2015;Shirazian et al, 2017). Many research papers have provided a comprehensive understanding of the factors that affect the process mechanisms and the properties of the produced granules (Benali et al, 2009, Dhenge et al, 2012Kumar et al, 2014;Christian and Markus, 2014;Tu et al, 2013;Vanhoorne et al, 2016).…”

“…an artificial neural network) or physically-based (e.g. population balance) models (Paavola et al, 2013;Shirazian et al, 2017). Population balance models (PBMs) have been implemented to follow the evolution of the granules with time (Sanders et al, 2003;Ramkrishna, 2000).…”

“…Therefore, these models have been integrated with various modelling approaches such as Monte Carlo, in order to model a greater number of particle properties, and the discrete element method, in order to model the effect of equipment dynamics (Barrasso et al, 2014;Braumann et al, 2007;Shirazian et al, 2017). Data based paradigms have also been utilized to model the granulation processes by mapping the process inputs to the outputs (i.e.…”

“…the properties of the granules) (AlAlaween et al, 2016;Mansaet al, 2008). For example, for the TSG process, an artificial neural network was presented to predict the granule size, which was represented by the three diameters: D10, D50 and D90 (Shirazian et al, 2017). In addition, modelling frameworks that integrate data-driven models with physical based ones have also been presented in the related literature (Barrasso et al, 2014).…”

A high-dimensional, stochastic model for twin-screw granulation Part 2: Numerical methodology

“…These agglomerates may have a size, porosity and liquid distribution that allow them to be utilised within the pharmaceutical and food industries, among others [1]. The granulation process is typically modelled mechanistically using population balance models (PBM) [2] (though non-mechanistic pure neural network approaches have also been investigated [3,4]). Using PBMs, the particle ensemble is transformed through processes such as nucleation, coagulation, breakage, consolidation, layering and wetting.…”