Simulación y validación de un sistema de destilación para la separación de Azeótropos de Benceno – Isopropanol – agua en procesos para la industria química

Abstract

The objective of this work was to create an artificial neural network (ANN) which is able to predict accurate or approximate data based on a prediction of the behavior of a benzene-isopropanol-water ternary separation system by changing specifications in the inlet stream and thus obtaining the fractions of the main components in each of the distillates, in the first column the main fraction is benzene, in the second column the main fraction is isopropanol and in the third column the fraction is water, this due to the results present in the reference simulation for this process by Tsai and Chien (2021). The design, training and validation of the artificial neural network was carried out in MATLAB software using data provided by the DWSIM software, validating the simulation results with high impact publications using Aspen HYSYS, one of the most widely used at industrial level. The neural network was developed from a database with 325 data for training and 25 data for validation, its architecture has 4 inputs, 65 hidden neurons and 3 outputs using the Bayesian Regularization (BR) algorithm for training, obtaining a Mean Squared Error (MSE) of 1.13867 E-02 and a Linear Regression coefficient (R) of 0.959099. The ANN was validated by means of a comparative statistical analysis reaching a confidence percentage of 95%, it was demonstrated that there is no significant discrepancy. The simulation achieved a benzene fraction of 0.9979 in the first distillate, isopropanol fraction of 0.9967 in the second distillate and water fraction of 0.9993 in the third distillate, demonstrating the viability of the industrial process for the separation.