Adaptive neuro-fuzzy inference systems (ANFIS) modeling of reactive distillation process


Giwa A., Giwa S. O., HAPOĞLU H.

ARPN Journal of Engineering and Applied Sciences, cilt.8, sa.7, ss.473-479, 2013 (Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 8 Sayı: 7
  • Basım Tarihi: 2013
  • Dergi Adı: ARPN Journal of Engineering and Applied Sciences
  • Derginin Tarandığı İndeksler: Scopus
  • Sayfa Sayıları: ss.473-479
  • Anahtar Kelimeler: ANFIS, Aspen HYSYS, Modeling, Reactive distillation, Simulation
  • Ankara Üniversitesi Adresli: Evet

Özet

This work has been carried out to develop ANFIS models for the reactive distillation process used for the production of isopropyl alcohol from the hydration reaction of propylene. The data used for the development of the models were generated from the Aspen HYSYS system of the process that comprised two feed streams - the upper feed stream from where the less volatile feed, water, was fed and the lower feed stream from where the more volatile feed, propylene, was fed into the column. The hydration reaction of the process was a reversible type occurring in liquid phase in the reaction sections of the column. The ANFIS models were trained, tested and simulated with the aid of MATLAB. The inputs of the models were the reflux ratio and the reboiler duty while the outputs were the top segment and the bottom segment temperatures. The high fit values and the low means of absolute errors obtained respectively from the training and the testing of the ANFIS models developed for the top segment and the bottom segment of the reactive distillation column used for the production of isopropyl alcohol have revealed that the developed ANFIS models represented the reactive distillation process in a very good manner. © 2006-2013 Asian Research Publishing Network (ARPN).