一种基于子空间辨识的预测控制器设计及其性能评价方法  

Directly Performance Assessment of Model Predictive Controller based on Subspace Identification

作　　者：刘伟[1] 田学民[1] 

机构地区：[1]中国石油大学(华东)信息与控制工程学院,东营257061

出　　处：《科技通报》2010年第5期726-729,736,共5页Bulletin of Science and Technology

基　　金：国家863计划项目(2007AA04Z193);山东省自然科学基金项目(Y2007G49)

摘　　要：在模型预测控制器的设计过程中模型辨识和控制器设计是两个独立的步骤;传统的LQG基准需要一个解析的参数模型来设计LQG控制器,并将其应用到实际过程中才能计算性能评价基准。为了在控制器投入实际应用之前,同时实现控制器的设计和性能评价基准的构造,本文提出一种基于子空间辨识的预测控制器设计及性能评价方法。该方法仅依赖过程历史输入输出数据,利用子空间辨识得到的子空间矩阵,设计预测控制器和构造性能评价指标。最终,在一个统一的框架下实现了预测控制器的设计和性能评价。针对Wood-Berry过程的仿真研究验证了该方法的可行性和有效性。In the steps of Model Predictive Controller （MPC） design,the model identification and the control of the system are considered as two separate problems. The conventional LQG benchmark procedure requires an explicit parametric model to design a LQG benchmark. To design controller and construct performance assessment benchmark before the control action,in this paper,we propose a MPC design and assessment method based on subspace identification. Just with the historic inputs and outputs,it designs the MPC and constructs the performance assessment benchmark only using the subspace matrix. As a result, the controller design and assessment are realized in a unified framework. The simulation results on the Wood-Berry model validate the feasibility and effectiveness of proposed method.

关 键 词：子空间辨识 预测控制 LQG基准 性能评价 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]