摘要
PVC树脂的颗粒特性对聚合物的成型加工具有重要意义,而目前缺少可靠的在线测量传感器件,通常采用取样离线分析,滞后时间太长,影响先进控制技术的有效应用。另外,聚合过程还呈现出高度的非线性特性。针对这些问题,采用广义回归神经网络对PVC树脂颗粒特性进行预测,验证了该方法的有效性,为保证PVC聚合过程先进控制技术的应用提供了有效方法。
The particle character of PVC resin is important to the molding and processing of polymer. Because reliable on - line measurement sensors are short, samples are usually off - line analyzed, which results in too long hesteres'is time and interferes with the effective application of advanced control technology. Moreover, highly nonlinear characteristics occur during polymerization. In order to solve these problems, the generalized regression nerve networks have been adopted to predict the particle character of PVC resin. The results have testified the efficiency of the method, and it ensures the application of advanced control technology in PVC polymerization.
出处
《聚氯乙烯》
CAS
2006年第4期32-34,共3页
Polyvinyl Chloride
基金
国家自然科学基金资助项目(60374003)
关键词
PVC
预测
广义回归神经网络
PVC
prediction
generalized regression nerve networks