Artificial Neural Network Modeling of Solubilities of 21 Commonly Used Industrial Solid Compounds in Supercritical Carbon Dioxide

In this communication, a feed-forward artificial neural network algorithm has been applied to calculate/predict the solubilities of 21 of the commonly used industrial solid compounds in supercritical carbon dioxide. An optimized three-layer feed-forward neural network using critical properties of solute and operating temperature and pressure is presented. Application of the model for 795 data points of 21 compounds gives a squared correlation coefficient of 0.9533 and an average absolute deviation of about 14% from the experimental values.