大范围工况热工过程的多模型预测控制  

Multiple model-based predictive control for the plant-wide thermal processes

作　　者：潘天红[1] 乐艳[1] 李少远[1] 

机构地区：[1]上海交通大学自动化研究所,上海200030

出　　处：《系统工程与电子技术》2004年第10期1439-1443,共5页Systems Engineering and Electronics

基　　金：国家"863"高技术计划基金资助课题(2002AA412130)

摘　　要：充分考虑大多数复杂热工控制对象非线性特性与运行工况密切相关的实际特点,采用多模型动态矩阵控制方法,并将该方法应用于某电厂300MW机组锅炉过热汽温对象,在典型工况下通过试验数据获得其局部三阶子模型集,基于每个局部子模型分别设计子DMC控制器。通过跟踪实际工况变化来对子控制器加权以获得合适的控制增量。实验结果表明该方法对参数突变适应快,可取得令人满意的控制效果。By taking into account the practical features that most thermal processes have nonlinear characteristics dependent closely on operating regime, a predictive dynamic matrix control method based on multiple models is presented. The method has been used in the control process of a 300 MW unit boiler superheater. In typical operating regimes, the third order local model sets are formulated with the experimental data. Based on each local model of the set, a local DMC controller is designed. Appropriate control increment is obtained by weighting of the local controllers according to the change of the real operating condition. The results show that this control method can rapidly adapt to the sudden changes in the process parameters and bring about good control effects.

关 键 词：多模型 预测控制 动态矩阵控制 热工过程 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]