DESARROLLO DE UN PROGRAMA PARA ESTUDIAR EL COMPORTAMIENTO DE UNA COLUMNA DE FRACCIONAMIENTO ETANO/ETILENO DE UNA PLANTA DE OLEFINAS (Development of a program for studying the behavior of an ethane - ethylene splitter of an olefins plant)

This paper presents the development of programs for the study of the behavior of a fractionating column of Ethane/Ethylene, which belongs to an olefin plant. We used a neural network (ANN) multilayer type with forward propagation algorithm and partial least squares (PLS) for this purpose. The training of the neural network and the PLS model obtaining were performed with experimental data of the plant. These experimental data were divided into a test set and a validation set. It was shown that the neural network was much higher than PLS algorithm for predicting the production of ethylene from operating parameters of the plant. Additionally, these results were compared with simulation results achieved by a commercial package, resulting in a good agreement between predictions of the neural network model and the commercial package ones.