基于Model-on-Demand的预测控制研究

在控制领域中，存在着大量的非线性时变系统，其中带有本质非线性环节的系统尤其难以得到好的控制效果。研究基于局部建模思想的Model-on-Demand（MOD）辨识算法及其与控制理论的结合，对于非线性系统的辨识及控制具有十分重要的意义。本文的研究主要集中在以下几个方面：比较广泛而深入地研究了MOD 辨识算法;对MOD 算法和广义预测控制的结合过程进行了大量研究，并详细推导出算法公式，提出了完整的实现过程;最后对污水生化处理过程的典型控制回路进行仿真。
关键词：非线性系统；局部建模；广义预测控制；
Abstract: There are many nonlinear time varying systems in control domain, and those systems having intrinsical nonlinear are harder to be controlled. Studying model-on-demand (MOD) which based on local modeling and its combination with control theory are significant to identification and control of nonlinear system. The main contents andconclusions are summarized as follows:The MOD algorithm is learned relatively and broadly based on vast literature. In this thesis, we have done many researches on the combination of model-on-demand and generalized predictive control, gets some formula meanwhile. At last a
application of them to wastewater treatment processes is studied.
Key words: nonlinear systems , local modeling, Generalized Predictive Control