考虑输入变量时滞的NO_x生成量动态建模

Dynamic modeling of NO_x production considering input variable time-delay

为提高选择性催化还原(SCR)脱硝系统的控制品质与经济性,需要建立精确的SCR脱硝反应器入口NO_x生成量动态模型。本文基于模糊树模型建立了SCR反应器入口NO_x生成量动态模型,提出采用复相关系数的时滞联合估计方法,对影响NO_x生成量的模型输入变量进行时滞估计。将该方法应用于某600 MW燃煤机组的运行数据,并对输入变量采用相关系数法和复相关系数法进行时滞估计对比。结果表明:模糊树模型的建模精度较高、泛化能力强;在模型输入变量进行时滞估计时,相较于相关系数法,采用复相关系数法对模型精度和泛化能力提升程度更高。

In order to improve the control quality and economy of selective catalytic reduction(SCR)denitration system,it is necessary to establish an accurate dynamic model of NOx production at inlet of the SCR denitrification reactor.On the basis of fuzzy tree model,this paper builds up a dynamic model for NOx production at the SCR reactor inlet,and proposes using a joint time-delay estimation method based on multiple correlation coefficients to estimate the time-delay of the input variables which affect the NOx production.The method was applied to the operation data of a 600 MW coal-fired unit,and the time-delay of input variables were estimated by correlation coefficient method and multiple correlation coefficient method.The experimental results show that,the fuzzy-tree model has high accuracy and generalization ability,the time-delay estimation of input variables based on multiple correlation coefficient method improves more accurate and generalization ability for the model than the correlation coefficient method.