Viscosity prediction of ternary mixtures containing ILs using multi-layer perceptron artificial neural network

Lashkarblooki, M ; Sharif University of Technology | 2012

1061 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.fluid.2012.04.017
  3. Publisher: Elsevier , 2012
  4. Abstract:
  5. Ionic liquids (ILs) have been considered as a good candidate to be replaced by the conventional solvent in recent years due to their potential consumptions and unique properties. In the present study, artificial neural network was used to predict the ternary viscosity of mixtures containing ILs. A collection of 729 experimental data points were gathered from the previously public shed literatures. Different topologies of a multilayer feed forward artificial neural network (MFFANN) were examined and optimum architecture was determined. Ternary viscosity data from the literature for 5 ILs with 547 data points have been used to train the network. In addition, to differentiate dissimilar substances, the molecular mass and boiling point temperature of the three components and two compositions of the non-ILs components were considered as input variables. It must be mentioned that due to the high boiling temperature of the ILs, most of them decomposes before achieving their boiling point. Therefore, Valderrama group contribution method was utilized to obtain the boiling points of the ionic liquids used for this study. Finally, the capability of the designed network was tested by predicting ternary viscosity of mixtures not considered during the training process of the network (182 ternary viscosity data points for 5 ILs). The results demonstrated that the proposed network was able to well predict the ternary viscosity data points even by using the predicted values of boiling temperatures of ionic liquids
  6. Keywords:
  7. ANN model ; Ionic liquids ternary viscosity ; Solutions ; Boiling temperature ; Boiling-point temperature ; Data points ; Experimental data ; Group contribution method ; High boiling temperatures ; Input variables ; Multi layer perceptron ; Multi-layer feed forward ; Ternary mixtures ; Three component ; Training process ; Viscosity data ; Viscosity prediction ; Forecasting ; Ionic liquids ; Mass transfer ; Neural networks ; Numerical analysis ; Optimization ; Viscometers ; Boiling point
  8. Source: Fluid Phase Equilibria ; Volume 326 , 2012 , Pages 15-20 ; 03783812 (ISSN)
  9. URL: http://www.sciencedirect.com/science/article/pii/S037838121200177X