酒精发酵系统的鲁棒预测控制器设计

Robust Model Predictive Controller on An Alcohol Fermentation System

针对具有系统状态测量延迟和输入输出约束的酒精发酵系统,研究了其鲁棒预测控制器的设计问题。介绍了酒精连续发酵过程及其数学模型,通过在平衡工作点附近的线性化处理得到系统的线性化模型;考虑到发酵系统线性化模型的参数变化,采用多胞结构来描述发酵过程。以此为基础,借助LMI(Linear Matrix Inequality)工具将鲁棒预测控制器的在线优化问题转化为求解一组由线性矩阵不等式构成的优化问题。另外,针对系统部分状态的实时不可测性,通过二次规划求解进行状态估计,利用估算值为优化计算提供有效信息,解决了系统状态测量时延给预测控制器处理带来的困难。仿真结果证明了该方法的有效性。

For an alcohol fermentation system with measurement-delayed state and input/output constraints,the problem of robust model predictive controller design is studied.A continuous alcohol fermentation process and its mathematical model are introduced.By linearizing the mathematical model at the equilibrium working point,a linear system model is obtained,which adopts the polytopic description to deal with the parameter variations of system.Robust MPC is later converted into a linear matrix inequality optimization problem thanks to LMI method.The system state measurement delay,which is considered as a difficulty for LMI method,is dealt with by state estimation using quadratic programming method.Simulation results illustrate the effectiveness of the proposed approach.