Neural Computing and Applications

, Volume 18, Issue 1, pp 15–24 | Cite as

Artificial neural networks for modeling in reaction process systems

  • Cristina Oliveira
  • Petia Georgieva
  • Fernando Rocha
  • Sebastião Feyo de Azevedo
IJCNN 2007

Abstract

This work is focused on developing a more efficient computational scheme for estimation of process reaction rates based on NN models. Two scenarios are considered: (1) the kinetics coefficients of the process are completely known and the process states are partly known (measured); (2) the kinetics coefficients and the states of the process are partly known. The contribution of the paper is twofold. From one side we formulate a hybrid (ANN and mechanistic) model that outperforms the traditional reaction rate estimation approaches. From other side, a new procedure for NN supervised training is proposed when target outputs are not available. The two scenarios are successfully tested for two benchmark problems, estimation of the precipitation rate of calcium phosphate and estimation of sugar crystallization growth rate.

Keywords

Neural network computational models Reaction rate estimation State observer 

References

  1. 1.
    Bastin G, Dochain D (1990) On-line estimation and adaptive control of bioreactors. Elsevier Science Publishers, AmsterdamGoogle Scholar
  2. 2.
    Lubenova V, Rocha I, Ferreira EC (2003) Estimation of multiple biomass growth rates and biomass concentration in a class of bioprocesses. Bioprocess Biosyst Eng 25(6):395–406. doi:10.1007/s00449-003-0325-1 CrossRefGoogle Scholar
  3. 3.
    Walter E, Pronzato L (1997) Identification of parametric models from experimental data. Springer, UKMATHGoogle Scholar
  4. 4.
    Noykove N, Muller TG, Gylenberg M, Timmer J (2002) Quantitative analysis of anaerobic wastewater treatment processes: identifiably and parameter estimation. Biotechnol Bioeng 78(1):91–103Google Scholar
  5. 5.
    Galvanauskas V, Georgieva P, Feyo de Azevedo S (2006) Dynamic optimisation of industrial sugar crystallization process based on a hybrid (mechanistic + ANN) model. In: IEEE world congress on computational intelligence, Vancouver, 16–21 July 2006Google Scholar
  6. 6.
    Georgieva P, Oliveira C, Rocha F, Feyo de Azevedo S (2007) Process modelling strategy combining analytical and data based techniques—I. NN identification of reaction rates. In: International joint conference on neural networks (IJCNN), Orlando, 12–17 August 2007Google Scholar
  7. 7.
    Haykin S (1999) Neural networks: a comprehensive foundation. Prentice Hall, NJMATHGoogle Scholar
  8. 8.
    Chen L, Bastin G (1996) Structural identifiability of the yeals coefficients in bioprocess models when the reaction rates are unknown. Math Biosci 132:35–67. doi:10.1016/0025-5564(95)00048-8 MATHCrossRefGoogle Scholar
  9. 9.
    Hagan Martin T, Mehnaj Mohammad B (1994) Training feed forward networks with the Marquardt algorithm. IEEE Trans Neural Netw 5(6):989–993. doi:10.1109/72.329697 CrossRefGoogle Scholar
  10. 10.
    Sorensen JS, Lundager Madsen HE (2000) The influence of magnetism on precipitation of calcium phosphate. J Crystal Growth 216:399–408. doi:10.1016/S0022-0248(00)00449-8 CrossRefGoogle Scholar
  11. 11.
    Ferreira A, Oliveira C, Rocha F (2003) The different phases in the precipitation of dicalcium phosphate dehydrate. J Crystal Growth 252:599. doi:10.1016/S0022-0248(03)00899-6 CrossRefGoogle Scholar
  12. 12.
    Georgieva P, Meireles MJ, Feyo de Azevedo S (2003) Knowledge based hybrid modeling of a batch crystallization when accounting for nucleation, growth and agglomeration phenomena. Chem Eng Science 58:3699–3707CrossRefGoogle Scholar

Copyright information

© Springer-Verlag London Limited 2008

Authors and Affiliations

  • Cristina Oliveira
    • 1
  • Petia Georgieva
    • 2
  • Fernando Rocha
    • 1
  • Sebastião Feyo de Azevedo
    • 1
  1. 1.Department of Chemical Engineering, Faculty of EngineeringUniversity of PortoPortoPortugal
  2. 2.Department of Telecommunications Electronics and Informatics /IEETAUniversity of AveiroAveiroPortugal

Personalised recommendations