摘要
在工业生产中,聚氯乙烯树脂的颗粒平均粒径测试,通常采用批次结束后取样,在实验室离线分析得到,一般导致几小时的滞后,影响了先进控制技术的有效应用.另外,聚合过程呈现出高度的非线性特性.针对这些特点,采用广义回归神经网络对PVC树脂颗粒特性进行预测研究,应用结果验证了该方法的有效性.
Particle feature of Polyvinyl chloride （PVC） is very important for polymer molding process. However, it is usually available at the end of each batch and measured off-line, normally hours delay in the quality control laboratory. This affects effective implement of the higher level of the automation in the process. In addition, polymerization process has higher nonlinear characteristic. For these characteristics, this paper uses generalized regression neural network to model Particle feature of Polyvinyl chloride （PVC） quality prediction model. The results prove the effectiveness of the used method.
出处
《江南大学学报（自然科学版）》
CAS
2006年第4期408-410,共3页
Joural of Jiangnan University (Natural Science Edition)
基金
国家自然科学基金项目(60374003)
关键词
聚氯乙烯
预测
广义回归神经网络
polyvinyl chloride （PVC）
prediction
generalized regression neural network
