摘要
将整个间歇生产过程表达成梯形的神经网络群,利用实际操作数据对整个网络群进行学习。同时研究了具有滚动运算特征的在线优化策略,该法完全避免了非线性系统实时辩识和建立优化模型的困难。对发酵生产过程的仿真结果,表明本文方法是有效的。
Based on the analysis to the batch process, a stair-type structure of neural network model is introduced and a new strategy for optimizing control of the batch process is proposed. The new method can carry over some difficulties of the real-time nonlinear identification and modelling procedure. A simulation of the batch fermentation process shows that this approach is effective.
关键词
神经网络
发酵
间歇过程
neural network
fermentation
batch process
optimizing control
modelling
process control
