超超临界机组模型辨识及预测控制仿真研究  

A Study of Model Identification and Model Predictive Control for Ultra-Supercritical Unit

作　　者：胡勇[1] 阎威武[1] 王国良[1] 

机构地区：[1]上海交通大学自动化系,上海200240

出　　处：《计算机仿真》2014年第7期121-126,共6页Computer Simulation

基　　金：国家自然科学基金(60974119)

摘　　要：超超临界机组是一个大而复杂的系统,具有机-炉耦合、动态非线性、大延时等特性。传统控制系统难以满足在节能减排及动态响应上的更高要求。为此采用模型辨识结合模型预测控制的方法对其进行建模和控制。建立了机组的闭环系统运行模型,在分析机组特点及预测控制器预测模型的需求的基础上,在模型辨识过程中,给出了一种子空间法和传统SISO(single input single output)辨识方法相结合的方法,以取得最终预测控制器所需的单位阶跃预测模型。实时仿真结果表明了所采用策略及方法的有效性。The Ultra-supercritical unit is a large and complicated system which characterizes Boiler-turbine cou- pling, dynamic nonlinearity and large time delay. Traditional control system is hard to satisfy the high requirement of energy-saving and emission reduction, friendly environment and system dynamic response. This paper discusses Model Identification and Model Predictive Control for Uhra-supercritical Unit. Based on the analysis of features of u- nit and the prediction model, a model identification method, which is combined SMI method with classical SISO system identification method, is proposed to obtain the final predictive model in the model identification process. Finally, the result of real time simulation demonstrates the effectiveness of the method.

关 键 词：子空间模型辨识 闭环辨识 模型预测控制 超超临界机组 

分 类 号：TM743[电气工程—电力系统及自动化]