REPRESENTATION OF IONIC LIQUID VISCOSITY-TEMPERATURE DATA BY GENERALIZED CORRELATIONS AND AN ARTIFICIAL NEURAL NETWORK (ANN) MODEL

• Published in: Chemical engineering communications 2013-12, Vol.200 (12), p.1600-1622
• Main Authors: Dutt, N. V. K., Ravikumar, Y. V. L., Rani, K. Yamuna
• Format: Article
• Language: English
• Subjects: Arrhenius model; Artificial neural network; Artificial neural networks; Atmospheric pressure; Correlation analysis; Generalized correlations; Image contrast; Ionic liquids; Learning theory; Neural networks; Temperature; Training; Viscosity; VTF model
• ISSN: 0098-6445; 1563-5201
• DOI: 10.1080/00986445.2012.756396

Ionic liquid (IL) viscosity (η)-temperature (T) data (at atmospheric pressure) are represented by an artificial neural network (ANN) model and the reduced forms of the Arrhenius, Vogel-Tamman-Fulcher (VTF), Arrhenius + higher order T-terms, and VTF + higher order T-term models. Training the data of 73 liquids (comprised of 5 cations and 32 anions) at 654 data points by a 2-7-1 neural network model (with inverse reduced temperature with respect to 323.15 K and logarithm of viscosity at 323.15 K as input neurons) yielded an overall percent deviation ( ) of 6.6% and a coefficient of determination (R 2 ) equal to 0.9. The other four models, in contrast, yielded values of and R 2 ranging between (21.2% and 21.7%) and (0.80 and 0.81), respectively. Testing the models with the data on an additional set of 8 ILs at 82 points yielded values of and R 2 comparable to those in the case of training. The results obtained show that the neural network scheme provides reasonably accurate viscosity estimates. On the other hand, the other four models could be used for design purposes, providing for approximate calculations. [Supplementary materials are available for this article. Go to the publisher's online edition of Chemical Engineering Communications for the following free supplemental resources: supplementary data tables.]