Use of neural networks for modeling of olefin polymerization in high pressure tubular reactors

Neural network computing is one of the fastest growing fields of artificial intelligence due to its ability to “learn” nonlinear relationships. This article presents the approach of back propagation neural networks for modeling of free radical polymerization in high pressure tubular reactors. Industrial data were used to train the network for prediction of the temperature profile along the reactor, as well as polymer properties such as density, melt flow index, and molecular weight averages. Comparisons were made between the neural network and mechanistic model predictions published in the literature. Results showed the promising capability of a neural network as an alternative approach to model polymeric systems. © 1994 John Wiley & Sons, Inc.