Modeling of Separation of Ethanol-Water Mixtures with the Help of Artificial Neural Network

Abstract

In this study, the water-ethanol separation was investigated using composite polydimethylsiloxane (PDMS) membranes in the pervaporation process with the help of the Artificial Neural Network. In the pervaporation process can be used to separate many liquids. Due to the presence of azeotropic in some materials and problems of conventional methods such as distillation, pervaporation process is used. In this study, experimental results are modeled and predicted using a neural network. The Multi Layers Perceptron (MLP) neural network was implemented with 2 inputs and outputs along with the Levenberg-Marquardt function and Propagation learning algorithm. Tang Activation algorithm was used for the hidden layer, and the pure algorithm was utilized for the output layer. Furthermore, for the hidden layer, 5 neurons were defined. At this method, 70 percent of data were allocated for learning and 15% of data were allocated for the experience and the remaining 15% of data were used for learning. The results of this study were accurate. The error percentage graphs for the actual values of the separation factor and the output of the flux were compared with the values obtained from the modeling through PDMS membranes to evaluate the efficiency of the pervaporation process in ethanol separation from water.