摘要
VCM的转化率是PVC聚合过程中一个重要的产品质量指标.由于PVC聚合过程较为复杂,且受现场条件的限制,VCM的转化率难以实时在线监测.因此,本文建立了一种基于Isomap-CS-Elman的软测量建模方法.利用Isomap算法对高维输入变量进行特征降维,确定了软测量模型的辅助变量,再采用CS智能优化算法优化Elman神经网络模型的结构参数,实现了输入输出变量之间的非线性映射.实验结果表明:与传统的Elman神经网络模型相比,所提出的模型具有更高的预测精度,较好地预测了VCM的转化率,满足了PVC聚合过程的实时控制要求.
The conversion rate of VCM is an important product quality index during the PVC polymerization process.Due to the complexity of the PVC polymerization process and the limitations of field conditions,the conversion rate of VCM is difficult to monitor online in real time.Therefore,a soft-sensor modeling method based on ISOMAP-CS-Elman is established in this paper.Firstly,the Isomap algorithm is used to reduce the characteristic dimensions of the high-dimensional input variables,and the auxiliary variables of the soft-sensor model are determined.Secondly,the CS optimization algorithm is used to optimize the structural parameters of the Elman neural network model,and the nonlinear mapping between input and output variables is realized.The experimental results show that compared with the traditional Elman neural network model,the proposed model has higher prediction accuracy,so the VCM conversion rate can be well predicted and the real-time control requirements of PVC polymerization process can be satisfied.
作者
张毅蒙
张国光
高淑芝
ZHANG Yimeng;ZHANG Guoguang;GAO Shuzhi(Shenyang University of Chemical Technology,Shenyang 110142,China)
出处
《沈阳化工大学学报》
CAS
2023年第4期362-368,共7页
Journal of Shenyang University of Chemical Technology
基金
辽宁省自然科学基金项目(20170540725)
辽宁省高端人才建设项目-辽宁省特聘教授(〔2018〕3533)。
