论文部分内容阅读
建立聚合反应切片平均分子量的预测模型对锦纶帘子布的生产有重要的意义。该文采用改进的遗传算法 (GA)和BP算法相结合的混合学习算法来训练神经网络 ,并采用多元逐步回归法对输入层节点数进行了优化 ,建立了聚合反应切片平均分子量在线预测的神经网络模型。在某化工厂聚合反应中的应用表明 ,该模型比基于最小二乘法的预测模型收敛速度快、预测精度高、网络的泛化能力强。
The prediction model of the average molecular weight of the polymerization reaction chip is of great significance for the production of nylon cord fabric. In this paper, a hybrid learning algorithm combining genetic algorithm (GA) and BP algorithm is used to train the neural network, and the number of nodes in the input layer is optimized by multivariate stepwise regression. The neural network with the average predicted molecular weight of polymerization line Network model. The application in polymerization reaction of a chemical plant shows that the model has faster convergence rate, higher prediction accuracy and better generalization ability than the prediction model based on least square method.