Long-term predictions using recurrent neural networks for state changes in polymerization reactors

Long-term predicting methods using neural networks (NN) are discussed for state changes in polymerization reactors. The temperature at the outlet of a continuous bulk polystyrene polymerization reactor is the present target of predictions using a layered neural network and some recurrent neural networks (RNN). Some structural problems in a general RNN are indicated, and two improvements (H -RNN with additional processing in hidden layer units, M-RNN with an additional calculating module of a hidden layer in a layered NN) are proposed on data processing and arranging by hidden layer units in RNN. As to the above networks, each predictive performance can be comparatively evaluated using mean square error. The predictive performance of H-RNN and M-RNN is superior to that of a layered NN in the initial stage of predictions, or in the state change with maximum or minimum points. In particular, long -term predictive performance is widely satisfied by M-RNN where the combined structure of RNN with hidden units of layered NN is built to improve accuracy in initial stage of predictions.