论文部分内容阅读
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。
The entire intermittent production process is expressed as a trapezoidal neural network group, using the actual operating data to learn the entire network group. At the same time, the on-line optimization strategy with rolling operation is studied. This method completely avoids the difficulties of real-time identification of nonlinear system and optimization model. The simulation results of fermentation process show that our method is effective.