Dynamic Modeling of the Essential Oil Extraction Based On Artificial Neural Networks

Supercritical oil extraction is a separation technique in chemical and food industries which exploit the solvent properties of oil near the critical point. Modeling of the yield and solubility of materials is an essential issue in supercritical oil extraction. Mathematical models of this process are very difficult because of the highly nonlinear relations between process variables and solubility. In this paper, an experimental flow-type apparatus has been designed for the extraction of essential oil from spearmint leaves with supercritical carbon dioxide. Three classes of artificial neural networks were developed for the simulation of the supercritical fluid extraction of spearmint oil based on laboratory data. Simulation results show the advantages of employing artificial neural networks in modeling of the nonlinear chemical process. Finally, a short comparison between the models of these three classes of neural network was presented.