No abstract

Surface tension as an important characteristic in much scientific and technological research is a function of liquid materials' physical properties. Thus, it is desirable to have an accurate correlation between effective parameters and surface tension. This study investigates the applicability of artificial neural networks as an efficient tool for the prediction of pure organic compounds' surface tensions for a wide range of temperatures. The experimental data gathered for training and verification of the network are related to a wide variety of materials such as alkanes, alkenes, aromatics, and sulfur, chlorine, fluorine, and nitrogen containing compounds. The most accurate network among several constructed configurations has one hidden layer with 20 neurons. The average absolute deviation percentage obtained for 1048 data points related to 82 compounds is 1.57%. The results demonstrate that the multilayer perceptron network could be an appropriate lookup table for the determination of surface tension as a function of physical properties.

Ionic liquids (ILs) have been of particular interest to researchers in recent years. Viscosity is considered as one of the most important thermophysical properties of ILs. This paper uses a genetic programming approach to propose a simple equation, with only one parameter to adjust, for predicting the viscosity of pure ILs and their binary mixtures as a function of temperature. The adjustable parameter of the equation was calculated for 103 pure ILs. The average absolute relative deviation of 4.60% obtained for 1374 data points confirms the good accuracy of the model. Also, an acceptable average absolute relative deviation of 12.83% at temperatures outside of the temperature range used to develop the model confirms the reliability of the model in the extrapolation of the temperature effect. To predict the viscosity of the binary mixture of ionic liquids by the proposed equation, the volumetric average of the parameter was applied to the proposed equation without introducing any new adjustable parameter. The average absolute relative deviation obtained for 954 data points comprised of 10 binary mixture of ionic liquids is 3.44%, which shows the good accuracy of the model in predicting the viscosity of a binary mixture of ionic liquids.

No abstract

No abstract