摘要
锅炉主蒸汽温对象的复杂非线性动态特性使得常规设计的固定参数比例-积分-微分(PID)串级控制很难适应工况变化和保证各种环境下的控制品质。为此,采用基于SOM的RBF-ARX模型框架,通过运行数据离线辨识的方法获取主汽温对象全局动态模型参数,在该模型基础上设计具有滚动优化和反馈校正功能的广义预测控制(GPC)器。220 t/h高压煤粉锅炉模型上的仿真结果和130 t/h煤粉炉上的工程应用均显示该方法得到的全局动态模型具有较高的预测精度,设计的GPC控制器具有较强的工况变化适应能力,能够适应多煤种变化和负荷变化。
The drum boiler main steam temperature is characterized by complex nonlinear dynamics, so control with conventional PID cascade control is rarely satisfactory. Therefore, an RBF-ARX control model based on the SOM was developed to describe the main steam temperature dynamics with its parameters identified offline from field data. A generalized predictive control （GPC） method with reeursive optimization and feedback was then developed based on the SOM-RBF-ARX model. Simulation of a 220 t/h drum boiler and tests on a 130 t/h drum boiler in the field show that the model accurately predicts and tracks variations of the working conditions due to fuel and demand changes.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第2期228-231,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目(50323002)
关键词
锅炉
主汽温控制系统
全局动态模型
广义预测控制
boiler
main steam temperature control system
global dynamic model
generalized predictive control
