(377b) A Multi-Scale Hybrid Modeling Approach for Continuous Pharmaceutical Unit Operations | AIChE

(377b) A Multi-Scale Hybrid Modeling Approach for Continuous Pharmaceutical Unit Operations


Bhalode, P. - Presenter, Rutgers University
Ierapetritou, M., University of Delaware
A prominent research focus for systematic process development in the continuous pharmaceutical industry is the implementation of Quality-By-Design [1] (QbD) approach. This approach focuses on detailed system understanding from a mechanistic perspective, using high-fidelity multi-physics simulation tools like computational fluid dynamics (CFD) [2] and discrete element modeling (DEM) [3]. These tools can accurately replicate the process dynamics within a unit operation, providing insight into complex processes, such as powder mixing and flow. However, due to huge computational costs associated with such simulations, direct implementation of these tools is prohibitive for quick process evaluations. On the other hand, flowsheet modeling focuses on empirical, low-order models using simplified representations of process dynamics. Though it provides quick assessment, it fails to capture important process complexities, such as powder mechanics. These underlying differences create a need for a synergistic integration of high-fidelity simulation models for capturing complex process dynamics with predictive flowsheet models for fast evaluations.

In this study, the proposed integration strategy is demonstrated for continuous blending unit operation using a sequential hybridization approach [4]. Under the proposed approach, particle-scale information [5,6] from DEM simulations (using Simcenter STAR_CCM+ software) is obtained to develop a data-driven model. This data-driven model is then combined with a flowsheet model (developed in PSE gFORMULATE modeling platform [7]) to provide fast predictions for the entire production line. The combination of DEM simulation with flowsheet modeling via a data-driven approach is performed such that any process changes occurring in the process flowsheet are recorded and replicated in the DEM simulation. This allows DEM simulation to capture dynamic effects of process changes on particle mechanics, which is then transferred to update the data-driven model linked to the flowsheet. This approach allows the flowsheet to incorporate effects of real-time process dynamics on powder behavior. The proposed study focuses on the application of this approach for evaluation and prediction of cross-sectional blend uniformity within the blender. Given the difficulties encountered in experimental measurement of cross-sectional blend uniformity for continuous blenders [8,9], the proposed approach provides an alternative for accurate evaluation of blend uniformity with detailed insights into the effects of process design and operating parameters. Blend uniformity is a critical quality attribute for tablet manufacturing and its accurate evaluation can avoid non-conformity of the final product.

In conclusion, the proposed work focuses on incorporation of detailed particle-scale information obtained from high-fidelity simulations within flowsheet models using hybrid data-driven approaches. This increases the mechanistic understanding of powder systems while improving the predictive ability of flowsheet simulations. The integrated approach presented in this work for powder blending systems, can be extended for other unit operations for the construction of robust flowsheet models and allow the implementation of QbD for continuous pharmaceutical process development.


[1] L.X. Yu, G. Amidon, M.A. Khan, S.W. Hoag, J. Polli, G.K. Raju, et al., Understanding Pharmaceutical Quality by Design, Aaps J. 16 (2014) 771–783.

[2] J.D. Anderson, J. Wendt, Computational fluid dynamics, Springer Berlin Heidelberg, Berlin, Heidelberg, 1995.

[3] P.A. Cundall, O.D.L. Strack, A discrete numerical model for granular assemblies, Géotechnique. 30 (1980) 331–336.

[4] M. von Stosch, R. Oliveria, J. Peres, S.F. de Azevedo, Hybrid semi-parametric modeling in process systems engineering: Past, present and future, Computers & Chemical Engineering. 60 (2014) 86–101.

[5] N. Metta, M. Verstraeten, M. Ghijs, A. Kumar, E. Schafer, R. Singh, et al., Model development and prediction of particle size distribution, density and friability of a comilling operation in a continuous pharmaceutical manufacturing process, International Journal of Pharmaceutics. 549 (2018) 271–282.

[6] D. Barrasso, A. Tamrakar, R. Ramachandran, Model Order Reduction of a Multi-scale PBM-DEM Description of a Wet Granulation Process via ANN, Procedia Engineering. 102 (2015) 1295–1304.

[7] Z. Wang, M.S. Escotet-Espinoza, M. Ierapetritou, Process Analysis and optimization of continuous pharmaceutical manufacturing using flowsheet models, Computers and Chemical Engineering. 107 (2017) 77–91.

[8] P.M. Portillo, M. Ierapetritou, F.J. Muzzio, Characterization of continuous convective powder mixing processes, Powder Technology. 182 (2008) 368–378.

[9] A.U. Vanarase, F.J. Muzzio, Effect of operating conditions and design parameters in a continuous powder mixer, 208 (2011) 26–36.