The predictive model of bubble point based on neural network

摘要

The calculation of bubble point plays an important role in the chemical process of separation. Traditional methods are quite complicated, and they are also time-consuming tasks. In the paper, the bubble points in trays of a methanol distillation column are first calculated by process simulation software named Design II. Then, some of the data are used to train a backpropagation (BP) neural network and a radial basis function (RBF) neural network respectively. Finally neural networks are used to predict the left bubble points. The result indicates that the predicted data are in good agreement with the experimental data obtained by Design II, and the speed of the RBF neural network is better than that of the BP neural network.