摘要
SCR脱硝系统被控对象具有大迟延和大惯性,采用常规PID难以满足对出口NO_(x)含量的精准控制。为提高控制品质,提出了基于阶梯式广义预测的SCR脱硝系统优化控制。首先采用粒子群算法辨识了NO_(x)排放量的数学模型,其次基于模型设计了阶梯式广义预测控制器,最后将基于联合互信息的LSSVM动态软测量模型得到的NO_(x)生成量作为前馈,构建了基于软测量模型的前馈-反馈控制系统。结果表明:基于阶梯式广义预测的脱硝优化控制能很好的减少出口NO_(x)的波动,将动态软测量模型得到的NO_(x)生成量预测值作为前馈引入SCR脱硝控制系统,克服了测量数据存在的滞后问题,可以及时准确的控制喷氨量,实现脱硝系统的稳态经济运行。
The controlled object of SCR denitration system has large delay and inertia,it is difficult to control the NO_(x) content accurately by using conventional PID.In order to improve the control quality,the optimal control of SCR denitration system based on stepped generalized prediction was proposed.Firstly,the mathematical model of NO_(x) emission was identified by particle swarm optimization.Secondly,a stepped generalized predictive controller was designed based on the model.At last,LSSVM dynamic soft-sensing model based on joint mutual information was used to obtain NO_(x) generation amount.And the amount was taken as feedforward to construct a feedforward-feedback control system based on soft-sensing model.The results show that the optimal control of denitration based on stepwise generalized prediction can reduce the fluctuation of NO_(x) at the outlet.The predicted value of NO_(x) production obtained by dynamic soft-sensor model was introduced into SCR denitration control system as feedforward,which overcomes the lag problem of measurement data and can timely and accurately control the ammonia injection amount,so as to realize the steady-state economic operation of denitration system.
作者
赵征
王金
马毅杰
李悦宁
ZHAO Zheng;WANG Jin;MA Yijie;LI Yuening(School of Cuntrol and Computer Engineering,North China Electric Power University,Baoding 071003,China)
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2022年第6期68-75,共8页
Journal of North China Electric Power University：Natural Science Edition
基金
中央高校基本科研业务费专项资金资助项目(2017MS133)
北京市自然科学基金面上项目(3202027).
关键词
SCR
粒子群算法
阶梯式广义预测控制
动态软测量
SCR
particle swarm optimization
stepped generalized predictive control
dynamic soft sensing
