摘要
为建立精确的选择性催化还原(selective catalytic reduction,SCR)脱硝系统动态模型,提出了一种基于时间差分(time difference,TD)的动态建模方法。常规动态建模方法一般是在模型输入端增加历史输入和输出信息,会增加模型的维数并影响其训练速度和泛化能力。TD方法通过求解输入、输出的一阶差分并建立变化量间的数据驱动模型,能在不增加维数的前提下实现动态建模。使用某燃煤电站的变负荷运行数据,针对正常运行以及测量仪表反吹两种工作状态,基于TD方法和最小二乘支持向量机建立了SCR脱硝系统的动态模型并与静态及常规动态建模方法进行了对比。结果表明,基于TD方法的动态模型维数低、训练速度快、建模精度高,且在测量仪表反吹时仍能保持较高的精度。
To establish an accurate selective catalytic reduction(SCR)denitration system model,a dynamic modeling method based on time difference(TD)method was proposed.The conventional dynamic modeling method usually adds historical input and output information to the input end of the model,which will increase the dimension of model and influence its training speed and generalization.By solving the first difference of input and output data and establishing data-driven model of variations,TD method can realize dynamic modeling without increasing dimensions.By adopting the variable-load operation data of a coal-fired power plant,a dynamic model of the SCR denitrification system was established on the basis of TD method and least squares support vector machine under the normal operation and measuring instruments blowback conditions.Then the method was compared with static and conventional dynamic modeling methods.The results show that the dynamic model based on TD method has low dimension,fast training speed,high modeling precision and great dynamic performance,and it can maintain high accuracy when measuring instrument blowback occurs.
作者
李海军
夏静
史恒惠
刘长良
王梓齐
LI Haijun;XIA Jing;SHI Henghui;LIU Changliang;WANG Ziqi(Technical Information Center of Henan Electric Power Co.,Ltd.,Zhengzhou 450001,China;State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources,North China Electric Power University,Beijing 102206,China;School of Control 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
2019年第1期89-94,共6页
Journal of North China Electric Power University:Natural Science Edition
基金
北京市自然科学基金资助项目(4182061)
关键词
时间差分
SCR脱硝系统
动态建模
数据驱动模型
最小二乘支持向量机
time difference
SCR denitrification system
dynamic modeling
data-driven model
least squares support vector machine