连续搅拌釜式生化反应器的非模型控制策略

MODEL-FREE CONTROL STRATEGIES FOR CONTINUOUS FERMENTERS

摘要: 本文提出了SIMO系统一种新的基于神经网络的非模型控制方法（MFC）．它包括线性控制策略和控制作用受限的非线性控制策略，讨论了非模型控制算法的收敛性和非模型控制系统的稳定性，将其用于某连续搅拌釜式发酵器的产率最优反馈控制．仿真研究获得了满意的结果．MFC方法无需对象模型，具有收敛快、自学习、强抗扰和鲁棒性好等特点，为难以建模的时变、非线性复杂生化过程的控制提供一条新的途径．

Abstract: A new model-free control (MFC) is proposed for the single-input and multi-output system.MFC consists of linear control strategy and nonlinear control strategy with control variables constraints.The convergence of MFC algorithm and the stability of MFC system are studied.MFC is used for direct optimal productivity control of continuous stirred tank fermenter using the dulation rate as the manipulated input.Excellent simulation results are obstained by using MFC without requiring the process model.