Predicting the Recovery of Aromatics in the Process of Pervaporation using Neural Network Modeling

Abstract

In this research, modeling for the process of separating penetrating aromas from a solution is performed using a PDMS composite membrane using an artificial neural network. The evaporation process can be used to separate many liquids. The amounts achieved from the experimental data were compared to the modeling data, and the results were analyzed. In this study, the Perceptron Multilayer Neural Network (MLP) is powered by a Levenberg-Marquardt diffusion learning and operation algorithm with two forward inputs and outputs. Tansig Activation algorithm was used for hidden layer and purelin algorithm was used for output layer. In addition, 5 neurons are defined here for the hidden layer. After processing the data, 70% was allocated for learning, 15% for credit and the remaining 15% for experience. The results obtained by the above method had good accuracy.