基于鲁棒最小二乘支持向量机的聚丙烯熔融指数预报

Melt Index Prediction of Polypropylene Based on Weighted Least Squares Support Vector Machines

可靠地预报熔融指数在聚丙烯生产过程中至关重要。在最小二乘支持向量机采用的误差平方和惩罚函数可能会导致不稳健的预报值基础上,进一步提出了基于鲁棒最小二乘支持向量机的聚丙烯熔融指数软测量模型。工业实例研究表明该方法拟合精度高、泛化能力强,具有广阔的应用前景。

Reliable estimation of melt index （MI） is crucial for the production of polypropylene. Considering the use of a SSE cost function without regularization, as in the case with LS-SVM, might lead to less robust estimates, the weighted LS-SVM soft-sensor model of propylene polymerization process is further presented. The research results show that the proposed method provides promising prediction reliability and accuracy. This method is supposed to have extensive application prospect in propylene polymerization processes.