文中利用B样条神经网络可以逼近连续函数的良好性质,构造合适的网络模型从而得到相对分子质量分布模型的线性表示方法,根据最优化方法中的梯度法,求得最优控制序列。应用在苯乙烯聚合反应的仿真实验中,收到了很好的控制效果,证明了方法的可行性。
Abstract
The characteristics of B-spline functions and the structure of B-spline neural networks was introduced and a new way was presented to build the neural networks and to get a linear express of molecular weight distribution. An optimization algorithm based on the linear express was used to calculate the optimal control sequences. A new control procedure was utilized on a simulated object and a well control led result was got that demonstrated the feasibility of this new way.
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1 ] Vicente M , Benamor S , Gugliotta L M. Control of molecular weight distribution in emulsion polymerization using on-line reaction calorimeter [ J ] . Indus and Eng Chem Res , 2001 , 40(1) : 218 - 227
[2 ] Ellis M F , Taylor T W , Jensen K F. On2line molecular weight distribution estimation and control in batch polymerization[J ] . AIChE Journal , 1994 , 40(3) :445 - 462
[3 ] Timothy J . Crowley , Kyu YongChoi. Discrete optimal control of MWD in a batch free radical polymerization process[J ] . Ind Eng Chem Res , 1997 , 36 : 3676 - 3682
[4 ] Scali C ,Ciari R ,Bello T , et al . Optimal temperature for the control of the product quality in batch polymerization :Simulation and experimental results [J ] . Journal of Applied Polymer Science , 1995 , 55(6) : 945 - 959
[5 ] 程正兴,李水根. 数值逼近与常微分方程数值解[M] .西安:西安交通大学出版社,2000 ,88 - 98
[6 ] 孙增圻. 智能控制理论与技术[M] . 北京:清华大学出版社,广西:广西科学技术出版社, 1997 ,163 - 169
[7 ] Girosi F , Poggio T. Networks and the best approximation property[J ] . Biol Cybern , 1990 , l63 :169 - 176
[8 ] Wang H. Bounded dynamic stochastic distributions modelling and control [ M ] . London : Springer-Verlag Ltd ,2000 ,17 - 38
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}