Industrial polymerization process quality prediction based on CPSO-LSTM-RNN

Due to the industrial polymer process is a complex and nonlinear process with the characteristic of multi-variables, hysteresis, large inertia and strong coupling, the main production targets are difficult to measure accurately, and there are fluctuations in data information during the production, so many production data need to be analyzed and processed. Therefore, this paper proposes the CPSO-LSTM-RNN algorithm to predict the yield of industrial polymer process. Firstly, the LSTM-RNN model is established and the model is trained with the data of the production process. Then, the CPSO algorithm is used to obtain the optimal hyperparameters of the model. Finally, the validity of the model is verified by a set of industrial data of the polymerization process.