用动态神经网络解决粘度预测的方法

A solution to the measurement lag in prediction of viscosity using dynamic neural network

针对聚酯生产中粘度计测量滞后的问题， 报告了用动态神经元网络进行粘度预测的方法， 着重讨论了神经网络输入变量的选择。

In order to solve the problem of measurement lag caused by the viscometer in PET production, a method for viscosity prediction of PET using dynamic neural networks is put forward, with special emphasis on selecting input variables, constructing dynamic ANNs and realizing on line correction.