金秀章,乔鹏,史德金

• 1:华北电力大学控制与计算机工程学院

摘要(Abstract)：

针对火电厂选择性催化还原（selective catalytic reduction,SCR）烟气脱硝系统中，由于影响入口NO_x质量浓度因素过多及系统大迟延大惯性，导致入口NO_x质量浓度难以准确及时测量的问题，提出了利用最大相关-最小冗余算法（max-relevance and min-redundancy,mRMR）结合贝叶斯优化算法（Bayesian optimization,BO）优化Stacking集成模型的SCR烟气脱硝系统入口NO_x质量浓度动态软测量模型。针对动态NO_x生成过程中静态单一模型预测精度降低及辅助变量与入口NO_x质量浓度时间异步的问题，利用mRMR-BO结合模型进行辅助变量筛选，Copula熵（copula entropy,CE）确定辅助变量迟延，BO结合模型确定辅助变量阶次，将TCN及LASSO利用Stacking法集成，使用含有迟延时间及阶次信息的辅助变量构建动态Stacking集成软测量模型。仿真结果显示：集成模型较TCN及LASSO单一网络的均方根误差、平均绝对误差、平均绝对百分比误差最小；动态集成模型对比静态集成模型，预测精度更高，能够实现对入口NO_x质量浓度的准确软测量。

关键词(KeyWords)： NO_x动态建模;最大相关-最小冗余;贝叶斯优化;Stacking集成模型

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 金秀章,乔鹏,史德金

DOI: 10.19666/j.rlfd.202302378

参考文献(References)：

• [1]金秀章,丁续达,赵立慧.传递熵变量选择的非线性系统时序预测模型[J].中国电机工程学报,2018,38(增刊1):192-200.JIN Xiuzhang,DING Xuda,ZHAO Lihui.Variables selection for nonlinear system time series prediction model by transfer entropy[J].Proceedings of the CSEE,2018,38(Supply.1):192-200.
• [2]付忠广,高学伟,李闯,等.基于改进模糊聚类与IPSO-SVM的燃煤电站NOx排放多模型预测[J].动力工程学报,2019,39(5):387-393.FU Zhong guang,GAO Xuewei,LI Chuang,et al.Multimodel NOx emission prediction based on IGASA-FCMand IPSO-SVM for coal-fired power plants[J].Journal of Chinese Society of Power Engineering,2019,39(5):387-393.
• [3]赵征,李悦宁,袁洪.燃煤机组NOx生成量动态软测量模型[J].动力工程学报,2020,40(7):523-529.ZHAO Zheng,LI Yuening,YUAN Hong.Dynamic model for soft sensing of NOx generation in coal-fired units[J].Journal of Chinese Society of Power Engineering,2020,40(7):523-529.
• [4]ZHOU J,LIANG H.Prediction of the NOx emissions from thermal power plant based on support vector machine optimized by chaos optimization algorithm[C]//IEEEInternational Conference on Information&Financial Engineering,2010.
• [5]LI Y,LI F.NOx prediction method based on deep extreme learning machine[C]//2018 3rd International Conference on Computational Intelligence and Applications (ICCIA),2018.
• [6]SUN X,JIA X,HOU P,et al.NOx emission prediction for a coal-fired power process using gated recurrent unit approach[C]//2020 Chinese Automation Congress (CAC),2020.
• [7]TANG Z,ZHANG H.Modeling NOx emission of coalfired boiler with differential evolution optimized least square support vector machine[C]//2018 Chinese Control And Decision Conference (CCDC),2018.
• [8]CUI J,JIA X.Prediction of NOx generation process based on a nonlinear ma model[C]//第31届中国控制与决策会议,2019.
• [9]MA J,SUN Z.Mutual information is copula entropy[J].Tsinghua Science&Technology,2011,16(1):51-54.
• [10]袁铁江,郭泽林,方彤.基于运行数据时空特征和Stacking集成学习的PEMFC故障诊断[J/OL].中国电机工程学报:1-10.(2022-07-08)[2022-11-28].https://kns.cnki.net/kcms/detail/11.2107.tm.20220707.1344.003.html.YUAN Tiejiang,GUO Zelin,FANG Tong.PEMFC fault diagnosis based on operation data temporal and spatial characteristics and stacking ensemble learning[J/OL].Proceedings of the CSEE:1-10.(2022-07-08)[2022-11-28].https://kns.cnki.net/kcms/detail/11.2107.tm.20220707.1344.003.html.
• [11]BAI S,KOLTER J Z,KOLTUN V.An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[J].Computation and Language 2018.DOI:10.48550/ar Xiv.1803.01271.
• [12]王万良,胡明志,张仁贡,等.改进TCN和LSTM的泸水河流域月径流量预测模型[J].计算机集成制造系统,2022,28(11):3558-3575.WANG Wanliang,HU Mingzhi,ZHANG Rengong,et al.Monthly runoff prediction model of Lushui river basin based on improved TCN and LSTM[J].Computer Integrated Manufacturing Systems,2022,28(11):3558-3575.
• [13]HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2016.
• [14]郑茗友,王伟,赵文杰,等.基于PSO-Lasso算法的电站湿法脱硫出口SO2浓度预测[J].华北电力大学学报(自然科学版),2022,49(1):112-117.ZHENG Mingyou,WANG Wei,ZHAO Wenjie,et al.Prediction of SO2 concentration at WFGD outlet of power station based on PSO-Lasso algorithm[J].Journal of North China Electric Power University,2022,49(1):112-117.
• [15]BERGSTRA J,BARDENET R,KéGL B,et al.Algorithms for hyper-parameter optimization[C]//Advances in Neural Information Processing Systems,2011.
• [16]金秀章,史德金,乔鹏.基于沙地猫群优化-最小二乘支持向量机的动态NOx排放预测[J/OL].中国电机工程学报:1-9.(2022-11-04)[2022-11-28].https://doi.org/10.13334/j.0258-8013.pcsee.222144.JIN Xiuzhang,SHI Dejin,QIAO Peng.Dynamic NOx emission prediction based on sandcat swarm optimizationleast squares support vector machine[J/OL].Proceedings of the CSEE:1-9.(2022-11-04)[2022-11-28].https://doi.org/10.13334/j.0258-8013.pcsee.222144.
• [17]胡2)华,姚圣,曾成.面向康复训练的多通道m RMR-PSO肌电特征选择算法[J].电子测量技术,2022,45(11):72-77.HU Ronghua,YAO Shen,ZENG Cheng.Multi-channel m RMR-PSO s EMG feature selection algorithm for rehabilitation training[J].Electronic Measurement Technology,2022,45(11):72-77.
• [18]马剑.Copula熵:理论和应用[R/OL].(2022-10-18)[2022-12-01].http://eprint.las.ac.cn/abs/202105.00070.MA Jian.Copula entropy:theory and applications[R/OL].(2022-10-18)[2022-12-01].http://eprint.las.ac.cn/abs/202105.00070.
• [19]金秀章,于静,刘岳.基于人工鱼群-径向基神经网络的NOx预测模型[J].动力工程学报,2021,41(7):551-557.JIN Xiuzhang,YU Jing,LIU Yue.NOx prediction model based on artificial fish-radial basis neural network[J].Journal of Chinese Society of Power Engineering,2021,41(7):551-557.
• [20]杜鹏,包晓安,胡逸飞,等.基于卡尔曼滤波的无线传感网时空数据融合算法[J].电子科技,2022,35(6):21-27.DU Peng,BAO Xiaoan,HU Yifei,et al.Research on spatio-temporal data fusion algorithm of wireless sensor network based on Kalman filter[J].Electronic Science and Technology,2022,35(6):21-27.