Neural - Wiener Model for Multivariable Methyl Tertiary Butyl Ether (MTBE) Reactive Distillation

Abstract:

MTBE is a chemical that can be used as anti-knock additive to replace lead additive (tetra ethyl lead) which can be efficiently produced using reactive distillation process. It has been established in the literature that MTBE reactive distillation poses a highly nonlinear behavior due to the combination of reaction and separation processes. A reliable model for predicting the behavior is required especially for the control purposes. In this work, a Neural Wiener model which is one of the available types of oriented block model was utilised to develop the MTBE reactive distillation model. The required data for the Neural Wiener model were generated using a validated Aspen dynamics model for the MTBE reactive distillation process. It is found that the Neural Wiener model is capable to predict the MTBE purity and isobutene conversion with accuracy of 98.55% and 96.95%, respectively. Those values are quantitatively better in comparison to the state space model which gives lower values for prediction accuracy of 87.86% and 82.90%, respectively.

• Periodical: Applied Mechanics and Materials (Volumes 284-287)
• Pages: 409-413
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.284-287.409
• Online since: January 2013
• Authors: Sudibyo Sudibyo, Muhamad Nazri Murat, Norashid Aziz
• Keywords: Neural Network Model, Neural Wiener Model, Reactive Distillation

References

[1] N. Sharma and K.Singh: Int. J. Chemical Reactor Eng. Vol. 8 (2010), R5

[2] E. VanSonnenberg, A. F. Hofmann, J. Neoptolemus, G. R. Wittich, R.A. Princenthal, S.W. Willson: AJR Vol. 146 (1986),p.865

DOI: 10.2214/ajr.146.4.865

[3] J. Labovsky, Z. Svandova, J. Markos, L. Jelemensky: J. Loss Prevention in the Process Industries Vol. 20 (2007), p.230–237

[4] A. P. Higler, R. Taylor and R. Krishna: Chem. Eng. Science, 54( 1999),p.2873

[5] M. M. Arefi, A. Montazeri, J. Poshtan and M. R. J. Motlagh: Chem. Eng. J. Vol. 138 (2008), p.274

[6] M. Lawry´nczuk: Neurocomputing Vol. 74 (2010), p.401

[7] G. Shafiee, A. A. Arefi, M. R. J. Motlagh and A. A. Jalali: Chem. Eng. J. Vol. 143 (2008), p.282

[8] M. Lawry´nczuk: ISMIS 2011, LNAI 6804 (2011), p.663

[9] M. Lawry´nczuk and P.Tatjewski: ICAISC 2006, LNAI 4029 (2006), p.76

[10] M. Lawry´nczuk: AIMSA 2008, LNAI 5253 (2008), p.410

[11] R.K. Al Seyab, Y. Cao: J. Process Contr.16 (2006), p.795

[12] R. Jacobs and R. Krishna: Ind. and Eng. Chemistry Res. 32 (1993), p.1706

[13] W. L. Luyben: J. Process Contr. Vol. 16 (2006), p.115

[14] M. Lawry´nczuk: ICANNGA 2011, Part I, LNCS 6593 (2011), p.230