Journal of Pharmaceutical Innovation

, Volume 5, Issue 4, pp 161–168 | Cite as

Population Balance Model-Based Hybrid Neural Network for a Pharmaceutical Milling Process

  • Pavan Kumar Akkisetty
  • Ung Lee
  • Gintaras V. Reklaitis
  • Venkat Venkatasubramanian
Research Article



Population balances are generally used to predict the particle size distribution resulting from the processing of a particulate material in a milling unit. The key component of such a model is the breakage function. In this work we present an approach to model breakage functions that has utility for situations in which determination of the breakage function from first principles is difficult. Traditionally, heuristic models have been used in those situations but the unstructured nature of such models limits their applicability and reliability.


To address this gap, we propose a semi-empirical hybrid model that integrates first principles knowledge with a data-driven methodology that takes into account the material properties, mill characteristics, and operating conditions. The hybrid model combines a discrete form of population balance model with a neural network model that predicts the milled particle size distribution given material and mill information.


We demonstrate the usefulness of this approach for compacted API ribbons milled in a lab scale Quadra conical mill for different materials and mill conditions. Comparisons are also given to the predictions obtained via a purely neural network model and a population balance model with a linear breakage kernel.


Milling Modeling Pharmaceutical processing Breakage function Neural net Hybrid model Population balance 



The authors would like to thank Ryan McCann from Industrial and Physical Pharmacy department at Purdue for providing the necessary material and training. The authors would also like to thank NSF-ERC SOPS (Engineering Research Center for Structured Organic Particulate Systems) and Indiana Twenty-First Century Research and Technology Fund for their financial support


  1. 1.
    Ramkrishna D. The status of population balances. Rev Chem Eng. 1985;3(1):49–95.Google Scholar
  2. 2.
    Bilgili, E. and B. Scarlett (2003) Estimation of the selection and breakage parameters from batch grinding: a novel full numerical scheme. In Proceedings of AICHE Annual Meeting, San FranciscoGoogle Scholar
  3. 3.
    Bilgili E, Scarlett B. Population balance modeling of non-linear effects in milling processes. Powder Technol. 2005;153(1):59–71.CrossRefGoogle Scholar
  4. 4.
    Bilgili E, Yepes J, Scarlett B. Formulation of a non-linear framework for population balance modeling of batch grinding: Beyond first-order kinetics. Chem Eng Sci. 2006;61(1):33–44.CrossRefGoogle Scholar
  5. 5.
    Chen Y, Ding Y, Papadopoulos DG, Ghadiri M. Energy-based analysis of milling. J Pharm Sci. 2004;93(4):886–95.CrossRefPubMedGoogle Scholar
  6. 6.
    de Vegt O, Vromans Herman, Faassen F, van der Voort Maarschalk K. Milling of organic solids in a jet mill. Part 1: determination of the selection function and related mechanical material properties. Part Part Syst Charact. 2005;22(2):133–40.CrossRefGoogle Scholar
  7. 7.
    de Vegt O, Vromans H, Faassen F, van der Voort Maarschalk K. Milling of organic solids in a jet mill. Part 2: checking the validity of the predicted rate of breakage function. Part Part Syst Charact. 2005;22(4):261–7.CrossRefGoogle Scholar
  8. 8.
    Laursen SÖ, Webb D, Ramirez WF. Dynamic hybrid neural network model of an industrial fed-batch fermentation process to produce foreign protein. Comput Chem Eng. 2007;31(3):163–70.CrossRefGoogle Scholar
  9. 9.
    Salman AD, Ghadiri M, Hounslow MJ. Particle breakage, vol. 12. Handbook of powder technology. Amsterdam, The Netherlands: Elsevier; 2007.Google Scholar
  10. 10.
    Oliveira R. Combining first principles modeling and artificial neural networks: a general framework. Comput Chem Eng. 2004;28(5):755–66.CrossRefGoogle Scholar
  11. 11.
    Piuri V, Alippi C. Artificial neural networks. J Syst Arch. 1998;44(8):565–7.CrossRefGoogle Scholar
  12. 12.
    Shuiping L et al. Nonlinear comminution process modeling based on GA-FNN in the computational comminution system. J Mater Process Technol. 2002;120(1–3):84–9.CrossRefGoogle Scholar
  13. 13.
    Sundaram A et al. Design of fuel additives using neural networks and evolutionary algorithms. AlChE J. 2001;47(6):1387–406.Google Scholar
  14. 14.
    Prasher CL. Crushing and grinding process handbook. New York: Wiley; 1987.Google Scholar
  15. 15.
    Kavuri S, Venkatasubramanian V. Combining pattern classification and assumption-based techniques for process fault diagnosis. Comput Chem Eng. 1992;16(4):299–312.CrossRefGoogle Scholar
  16. 16.
    Kavuri S, Venkatasubramanian V. Representing bounded fault classes using neural networks with ellipsoidal activation functions. Comput Chem Eng. 1993;17(2):139–63.CrossRefGoogle Scholar
  17. 17.
    Kavuri S, Venkatasubramanian V. Using fuzzy clustering with ellipsoidal units in neural networks for robust fault classification. Comput Chem Eng. 1993;17(8):765–84.CrossRefGoogle Scholar
  18. 18.
    Kavuri S, Venkatasubramanian V (1994) “Neural network decomposition strategies for large scale fault diagnosis” (Morari and Morris (ed.)). Special Issue of Intl. J. Control 59(3): 767–792Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2010

Authors and Affiliations

  • Pavan Kumar Akkisetty
    • 1
  • Ung Lee
    • 1
  • Gintaras V. Reklaitis
    • 1
  • Venkat Venkatasubramanian
    • 1
  1. 1.Laboratory for Intelligent Systems, School of Chemical EngineeringPurdue UniversityWest LafayetteUSA

Personalised recommendations