Journal of Hebei University(Natural Science Edition) >

2018 , Vol. 38 >Issue 4: 423 - 431

DOI: https://doi.org/10.3969/j.issn.1000-1565.2018.04.013

Prediction of denitrification system inlet nitrogen oxide based on neural network online learning

  • JIN Xiuzhang ,
  • ZHANG Shaokang
Expand
  • School of Control and Computer Engineering, North China ElectricPower University, Baoding, 071003, China

Received date: 2017-09-07

  Online published: 2018-07-25

Fold

Abstract

To solve the problem that Static soft sensing model of denitrification system inlet nitrogen oxides can notsatisfy the requirement of variable load,a soft sensor model based on neural network online learning was proposed.Finding the optimum solution of Static neural network by particle swarm optimization,combining prediction error with current prediction error to update weights,thresholds and learning rates online,which can satisfy the requirement of variable load,the model was verified with power plant actual data of power plant.Results show that neural network online learning model is high accuracy and good realtime,with strong generalization ability,which is able to predict the inlet nitrogen oxides,and provide a effective method for online measuremet and monitoring of inlet nitrogen oxides in NOx reduction control system.

Key words： inlet nitrogen oxides; kernel principal component analysis; particle swarm optimization; prediction error; neural network online learning

Cite this article

JIN Xiuzhang , ZHANG Shaokang . Prediction of denitrification system inlet nitrogen oxide based on neural network online learning[J]. Journal of Hebei University(Natural Science Edition), 2018 , 38(4) : 423 -431 . DOI: 10.3969/j.issn.1000-1565.2018.04.013

References

[1] 崔超,杨婷婷,刘吉臻,等.基于BP网络变量选择与LSSVM结合的锅炉脱硝经济性建模[J].中国电机工 程学报,2015,35(24):6413-6420.DOI:10.13334/j.0258-8013.pcsee.2015.24.017. CUI C,YANG T T,LIU J Z,et al.Denitration cost modeling for boilers based on BP variable selection and LSSVM[J].Proceedings of the CAEE,2015,35(24):6413-6420.DOI:10.13334/j.0258-80 13.pcsee.2015.24.017.
[2] 武宝会,崔利.火电厂SCR烟气脱硝控制方式及其优化[J].热力发电,2013,42(10):116-119.DOI:10.3969/j.issn.1002-3364.2013.10.116. WU B H,CUI L.SCR flue gas denitrification control and optimization in thermal power plants[J].Thermal Power Generation,2013,42(10):116-119.DOI:10.3969/j.issn.1002-3364.2013.10.116.
[3] 韩良云,陆金桂.神经网络球磨机出力软测量模型的建立[J].动力工程学报,2015,35(11):901-905+933. HAN L Y,LU J G.Modeing for soft sensing of coal mill output based on neural network[J].Chinese Journal of Power Engineering,2015,35(11):901-905+933.
[4] 杨耀权,张新胜.LSSVM动态软测量模型在磨煤机一次风量预测方面的应用[J].动力工程学报,2016,36(03):207-212. YANG Y Q,ZHANG X S.Soft-sensing of pri mary air flow in a coal mill based on LSSVM[J].Chinese Journal of Power Engineering,2016,36(03):207-212.
[5] 金秀章,郝兆平,佟纯涛,等.KPCA和LSSVM在SCR脱硝反应器入口NO_x含量软测量中的应用[J].工业仪表与自动化装置,2016,45(05):124-128. JIN X Z,HAO Z P,TONG C T,et al.The application of KPCA and LSSVM in the soft-sensing of SCR denitration reactor inlet NOx content[J].Industrial instrumentation automation,2016,45(05):124-128.
[6] QIAO J F,ZHANG Y,HAN H G.Fast unit pruning algorithm for feedforward neural network design[J].Applied Mathematics and Computation,2008,205(02):622-627.DOI:10.1016/j.amc.2008.05.049.
[7] 金秀章,张少康,尹子剑,等.时序-神经网络模型在磨煤机一次风量测量中的应用[J].自动化仪表,2017,38(01):46-49.DOI:10.16086/j.cnki.issn1000-0380.201701011. JIN X Z,ZHANG S K,YIN Z J,et al.Application of time sequence Neural Network Model in Measuremet of Primary Air Flow of Pulverizer[J].Process automation instrumentation,2017,38(01):46-49.DOI:10.1 6086/j.cnki.issn1000-0380.201701011.
[8] 胡学聪,李柠,李少远,等.基于Matlab的SCR烟气脱硝仿真平台设计[J].系统仿真学报,2010,22(0 1):71-74.DOI:10.16182/j.cnki.joss.2010.01.004. HU X C,LI N,LI S Y,et al.Development of simulation platform based on matlab for SCR flue gas denitration[J].Journal of system simulation,2010,22(01):71-74.DOI:10.16182/j.cnki.joss.2 010.01.004.
[9] 金秀章,韩超.KPCA-LSSVM在磨煤机一次风量预测中的应用[J].自动化仪表,2015,36(03):62-67.DOI:10.16086/j.cnki.issn1000-0380.201503017. JIN X Z,HAN C.Application of KPCA-LSSVM in prediction of the primary air flow of pulverizer[J].Process automation instrumentation,2015,36(03):62-67.DOI:10.16086/j.cnki.issn1000-038 0.201503017.
[10] 黄宴委,彭铁根.基于核主元分析的非线性动态故障诊断[J].系统仿真学报,2005,17(09):2291-2294.D OI:10.16182/j.cnki.joss.2005.09.064. HUANG Y W,PENG T G.Fault diagnosis for nonlinear dynamic system using kernel principal component analysis[J].Journal of system simulation,2005,17(09):2291-2294.DOI:10.16182/j.cnki.joss.2005.09.064.
[11] 王玉昆,党金梗.基于KPCA与ILSFA-LSSVM的浮选加药控制模型[J].控制工程,2017,24(02):326-3 30.DOI: 10.14107/j.cnki.kzgc.140465. WANG Y K,DANG J G.Flotation dosing control modelbased on KPCA and ILSFA-LSSVM[J].Co- ntrol engineering,2017,24(02):326-330.DOI: 10.14107/j.cnki.kzgc.140465.
[12] 洪文鹏,陈重.基于自适应粒子群优化BP神经网络的氨法烟气脱硫效率预测[J].动力工程学报,2013,33(04):290-295. HONG W P,CHEN Z.Efficiency prediction of ammonia flue gas desulfurization based on adaptive PSO-BP model[J].Chinese Journal of Power Engineering,2013,33(04):290-295.
[13] 向怀坤,李伟龙,谢秉磊.粒子群优化神经网络的交通事件检测算法研究[J].计算机测量与控制,2016,24(02):171-174.DOI:10.16526/j.cnki.11-4762/tp.2016.02.047. XIANG H K,LI W L,XIE B L.Research on traffic incident detection algorithm based on particle swarm optimizer neural network[J].Computer measurement&control,2016,24(02):171-174.DOI:10.16526/ j.cnki.11-4762/tp.2016.02.047.
[14] 乔俊飞,李瑞祥,柴伟,等.基于PSO-ESN神经网络的污水BOD预测[J].控制工程,2016,23(04):463-467.DOI:10.14107/j.cnki.kzgc.140445. QIAO J F,LI R X,CHAI W,et al.Prediction of BODbased on PSO-ESN Neural Network[J].Control Engineering of China,2016,23(04):463-467.DOI:10.14107/j.cnki.kzgc.140445.
[15] 张英堂,马超,李志宁,等.基于快速留一交叉验证的核极限学习机在线建模[J].上海交通大学学报,2014,48(05):641-646. ZHANG Y T,MA C,LI Z N,et al.Online modeling of kernel extreme learning machine based on fast leave-one-out crossvalidation[J].Journal of Shanghai Jiaotong University,2014,48(05):641-646.
[16] 李云路,王大志,宁一,等.基于复变自适应神经网络的电网相位估计方法[J].东北大学学报(自然科学版),2017,38(1):6-10.DOI:10.3969/j.issn.1005-3026.2017.01.002. LI Y L,WANG D Z,NING Y,et al.Phase estimation method for power system based on complex adaptive neural network[J].Journal of Northrastern University Natural Science,2017,38(01):6-10.DOI:10.3969/j.issn.1005-3026.2017.01.002.
Options
Outlines

/

〈 〉