中文摘要：

针对工业双环管丙烯聚合装置熔融指数（MI）在线测量的困难，进行了聚丙烯熔融指数的在线最小二乘建模与自适应预报方法的研究。首先结合聚合反应机理分析，以Hammerstein模型描述聚丙烯熔融指数的动态特性，其次，利用渐消记忆递推最小二乘法在线更新Hammerstein模型参数。在此基础上，通过降维观测器的设计提出并实现了工业双环管聚丙烯装置的一种熔融指数的在线最小二乘建模与自适应预报方法。最后，利用工业数据验证了此算法的可行性和有效性。

英文摘要：

The online least-square modeling and adaptive prediction of polypropylene melt index （MI） were investigated to overcome the difficulties in measurement of MI in industrial bi-tubular propylene polymerization plants. Firstly, with combination of the analysis of the pmpylene polymerization mechanism, the Hammerstein model was used to represent the dynamics of melt index of polypropylene, and secondly, the fading memory recursive least-square method was employed to on-line update the parameters of the Hammerstein model. This then was exploited to design a reduced-dimensional state estimator, which established an online least-square modeling and adaptive predictor of the polypropylene melt index for industrial bi-tubular polypmpylene plants. Finally, the process data was used to demonstrate the effectiveness of the method presented here.