Discrete element reduced‐order modeling of dynamic particulate systems

Abstract

One of the key technical challenges associated with modeling particulate processes is the ongoing need to develop efficient and accurate predictive models. Often the models that best represent solids handling processes, like discrete element method (DEM) models, are computationally expensive to evaluate. In this work, a reduced-order modeling (ROM) methodology is proposed that can represent distributed parameter information, like particle velocity profiles, obtained from high-fidelity (DEM) simulations in a more computationally efficient fashion. The proposed methodology uses principal component analysis (PCA) to reduce the dimensionality of the distributed parameter information, and response surface modeling to map the distributed parameter data to process operating parameters. This PCA-based ROM approach has been used to model velocity trajectories in a continuous convective mixer, to demonstrate its applicability for pharmaceutical process modeling.

Perspectives on the continuous manufacturing of powder‐based pharmaceutical processes

“…An alternative approach involves the use of latent variables to reduce the dimensionality of data from DEM simulations combined with response surface modeling to map the process operating parameters to the distributed parameter information. The use of reduced‐order modeling in powder processing applications can enhance the accuracy of flowsheet models by incorporating information from high‐fidelity simulations such as DEM without incurring the corresponding computational costs (see Figure ).…”

“…Therefore, a limitation on the number of parameters considered is recommended before one conducts a variance‐based sensitivity analysis. This can be accomplished through screening methods (e.g., Morris screening) or a preliminary regression‐based sensitivity analysis (e.g., with the partial correlation coefficient or partial rank correlation coefficient) …”

Discrete element modelling (DEM) input parameters: understanding their impact on model predictions using statistical analysis

“…The most important parameters will have a strong contribution to the largest eigenvalues [37][38][39]. In DEM, the method was recently applied in the modelling of a continuous convective mixer [40]. Sensitivity analysis of model outputs with respect to model parameters can also be applied using the underpinning mechanistic equations which are solved in DEM simulations (e.g., Hertz-Mindlin equation).…”

Modeling and Optimization of Continuous Pharmaceutical Manufacturing Processes

“…DEM) simulations and equation-oriented process models. For instance, PCA-based reduced-order models have been shown to facilitate efficient and accurate representation of distributed parameter information from DEM simulations (Boukouvala, Gao, et al, 2013;Rogers & Ierapetritou, 2014). These reduced-order models evaluate in a matter of cpu-seconds (as compared with hours or even days for DEM) and are therefore much better suited to flowsheet modeling applications.…”