钱虹,张俊,徐邦智

• 1:上海电力大学自动化工程学院
• 2:上海市电站自动化技术重点实验室

摘要(Abstract)：

针对当前燃煤-煤气锅炉煤气掺烧量不确定情况下对NO_x排放量预测不够准确的问题，提出一种基于注意力机制组合在线预测模型。首先，通过最大信息系数法与皮尔逊相关系数法相结合确定模型的特征变量；其次，对线性相关特征变量采用滑动时间窗口在线构建向量自回归模型（VAR），实现多维时序线性相关变量输入下对NO_x排放量的预测，而对于非线性相关特征变量通过构建在线循环极限学习机（OR-ELM）模型在线学习非线性相关变量在时序上的关系对NO_x排放量进行预测；最后，采用注意力机制对2个预测模型进行动态赋权以实现趋势预测。采用实际运行数据对该模型验证，结果表明，所构建的VAROR-ELM组合在线预测模型能够准确预测10min后的NO_x排放量变化趋势，并在不同负荷段对NO_x质量浓度进行准确预测；综合预测精度及预测时间，所构建的组合预测模型比其他单一预测模型的预测效果更好。

关键词(KeyWords)： 最大信息系数;注意力机制;组合预测;在线学习;NO_x排放

Abstract:

Keywords:

基金项目(Foundation): 上海市自然科学基金（19ZR1420700）~~

作者(Author): 钱虹,张俊,徐邦智

DOI: 10.19666/j.rlfd.202212226

参考文献(References)：

• [1]李伟. 300 MW高炉煤气煤粉混燃锅炉NOx特性数值模拟及研究[D].保定:华北电力大学, 2013:1.LI Wei. Study and numerical simulation of 300 MW blast furnace gas and pulverized coal blended combustion boiler on the optimizing of NOx[D]. Baoding:North China Electric Power University, 2013:1.
• [2]赵晶.火电厂SCR烟气脱硝系统建模与控制策略研究[D].保定:华北电力大学, 2018:1.ZHAO Jing. Research on modeling and control strategy of SCR flue gas denitrification system in thermal power plants[D]. Baoding:North China Electric Power University, 2018:1.
• [3]毛剑宏.大型电站锅炉SCR烟气脱硝系统关键技术研究[D].杭州:浙江大学, 2011:1.MAO Jianhong. Key technology study of SCR flue gas deNOx system in large power plant boiler[D]. Hangzhou:Zhejiang University, 2011:1.
• [4]谷丽景,李永华,李路.电站锅炉燃烧优化混合模型预测[J].中国电机工程学报, 2015, 35(9):2231-2237.GU Lijing, LI Yonghua, LI Lu. Hybrid model prediction of utility boiler combustion optimization[J]. Proceedings of the CSEE, 2015, 35(9):2231-2237.
• [5] IRFAN M F, MJALLI F S, KIM S D. Modeling of NH3-NO-SCR reaction over CuO/gamma-Al2O3 catalyst in a bubbling fluidized bed reactor using artificial intelligence techniques[J]. Fuel, 2012, 93(1):245-251.
• [6]钱虹,张超凡,柴婷婷.基于随机森林的SCR脱硝系统出口NOx浓度预测研究[J].热能动力工程, 2021,36(3):122-129.QIAN Hong, ZHANG Chaofan, CHAI Tingting. Research on outlet NOx concentration prediction of SCR denitration system based on random forest algorithm[J]. Journal of Engineering for Thermal Energy and Power, 2021, 36(3):122-129.
• [7]王伟同,范海东,梁成思,等.基于随机森林算法的对冲锅炉出口NOx排放量预测模型研究[J].热力发电,2022, 51(4):96-104.WANG Weitong, FAN Haidong, LIANG Chengsi, et al.Predictive modeling of NOx outlet of hedged boiler based on random forest[J]. Thermal Power Generation, 2022,51(4):96-104.
• [8]许壮,康英伟.基于随机森林的火电机组SCR脱硝反应器建模[J].动力工程学报, 2020, 40(6):486-491.XU Zhuang, KANG Yingwei. Modeling of SCR denitrification reactor in a thermal power unit based on random forest algorithm[J]. Journal of Chinese Society of Power Engineering, 2020, 40(6):486-491.
• [9] LIU Q, HU H L. Prediction model of the NOx emissions based on long short-time memory neural network[C]//2020 Chinese Automation Congress(CAC)in Shanghai,China:IEEE, 2020:2795-2797.
• [10]高常乐,司风琪,任少君,等.基于LSTM的烟气NOx浓度动态软测量模型[J].热能动力工程, 2020, 35(3):98-104.GAO Changle, SI Fengqi, REN Shaojun, et al. Dynamic soft sensing model of NOx concentration in flue gas based on LSTM[J]. Journal of Engineering for Thermal Energy and Power, 2020, 35(3):98-104.
• [11]杨国田,张涛,王英男,等.基于长短期记忆神经网络的火电厂NOx排放预测模型[J].热力发电, 2018,47(10):12-17.YANG Guotian, ZHANG Tao, WANG Yingnan, et al.Prediction model for NOx emissions from thermal power plants based on long-short-term memory neural network[J].Thermal Power Generation, 2018, 47(10):12-17.
• [12]潘红光,裴嘉宝,苏涛,等.基于LSTM的燃煤电厂NOx排量软测量[J].西安科技大学学报, 2022, 42(2):362-370.PAN Hongguang, PEI Jiabao, SU Tao, et al. LSTM-based soft sensor of NOx emissions from coal-fired power plants[J]. Journal of Xi’an University of Science and Technology, 2022, 42(2):362-370.
• [13]钱虹,柴婷婷,张超凡.基于深度循环神经网络的SCR烟气脱硝系统出口NOx排放预测研究[J].热能动力工程, 2020, 35(8):77-84.QIAN Hong, CHAI Tingting, ZHANG Chaofan. NOx emission prediction model of SCR flue gas denitrification system based on deep recurrent neural network[J]. Journal of Engineering for Thermal Energy and Power, 2020,35(8):77-84.
• [14]谢锐彪,李新利,王英男,等.基于粒子群优化的双向门控循环神经网络燃煤电厂NOx排放预测[J].热力发电, 2021, 50(10):87-94.XIE Ruibiao, LI Xinli, WANG Yingnan, et al. NOx emission prediction of coal-fired power plants based on PSO and Bi-GRU[J]. Thermal Power Generation, 2021,50(10):87-94.
• [15] TANG Z, WANG S, CAO S, et al. Dynamic prediction model for NOx emission at the outlet of SCR system based on extreme learning machine[C]//2020 Chinese Automation Congress(CAC)in Shanghai, China:IEEE, 2020:3226-3229.
• [16]程琳,赵文杰.基于GA-ELM的SCR脱硝系统动态建模[J].热力发电, 2019, 48(6):29-33.CHENG Lin, ZHAO Wenjie. Dynamic modeling of SCR denitrification system based on GA-ELM[J]. Thermal Power Generation, 2019, 48(6):29-33.
• [17]唐振浩,王世魁,曹生现,等.基于混合数据驱动算法的SCR氮氧化物排放量动态预测模型[J].中国电机工程学报, 2022, 42(9):3295-3307.TANG Zhenhao, WANG Shikui, CAO Shengxian, et al.Dynamic prediction model for NOx emission of SCR system based on hybrid data-driven algorithms[J].Proceedings of the CSEE, 2022, 42(9):3295-3307.
• [18] LIU X, LIN Z, FENG Z. Short-term offshore wind speed forecast by seasonal ARIMA-A comparison against GRU and LSTM[J]. Energy, 2021, 227:120492.
• [19] LI Y J, YUAN X l, XU J Y, et al. Medium-term forecasting of cold, electric and gas load in multi-energy system based on VAR model[C]//2018 13th IEEE Conference on Industrial Electronics and Applications(ICIEA). Wuhan:IEEE, 2018:1676-1680.
• [20]崔建国,李鹏程,崔霄,等.基于ARIMA-LSTM的飞机液压泵性能趋势预测方法[J].振动,测试与诊断,2021, 41(4):735-740.CUI Jianguo, LI Pengcheng, CUI Xiao, et al. Aircraft hydraulic pump performance trend prediction method based on ARIMA-LSTM[J]. Journal of Vibration,Measurement&Diagnosis, 2021, 41(4):735-740.
• [21]王英伟,马树才.基于ARIMA和LSTM混合模型的时间序列预测[J].计算机应用与软件, 2021, 38(2):291-298.WANG Yingwei, MA Shucai. Time series forecasting based on ARIMA and LSTM hybrid model[J]. Computer Applications and Software, 2021, 38(2):291-298.
• [22]杨锡运,刘欢,张彬,等.基于熵权法的光伏输出功率组合预测模型[J].太阳能学报, 2014, 35(5):744-749.YANG Xiyun, LIU Huan, ZHANG Bin, et al. A combination method for photovoltaic power forecasting based on entropy weight method[J]. Acta Energiae Solaris Sinica, 2014, 35(5):744-749.
• [23]陈刚,杨轶.煤粉和高炉煤气混烧锅炉燃烧问题的分析及改造[J].电站系统工程, 2003, 19(2):36-38.CHEN Gang, YANG Yi. Technical reform of combustion system in pulverized coal and BFG combustion boiler[J].Power System Engineering, 2003, 19(2):36-38.
• [24]郑峰,王睿,宦克为,等.粳米近红外光谱模型中基于拉依达准则的异常值剔除方法[C]//第十七届全国分子光谱学学术会议论文集, 2012:135-136.ZHENG Feng, WANG Rui, HUAN Kewei, et al.Eliminating abnormal values in keng-rice NIRS modeling based on PauTa criterion[C]//Proceedings of the 17th National Conference on Spectroscopy and Spectral Analysis, 2012:135-136.
• [25]陈昭.时序非平稳性ADF检验法的理论与应用[J].广州大学学报(自然科学版), 2008(5):5-10.CHEN Zhao. ADF test for non-stationary time series:theory and application[J]. Journal of Guangzhou University(Natural Science Edition), 2008(5):5-10.
• [26] LU G, YU E, WANG Y, et al. A novel hybrid machine learning method(OR-ELM-AR)used in forecast of PM2.5concentrations and its forecast performance evaluation[J].Atmosphere, 2021, 12(1):78.
• [27]王璇.基于注意力机制双向长短期记忆网络的短期负荷滚动预测方法[D].天津:天津大学, 2019:1.WANG Xuan. Short-term load rolling forecasting method based on attention mechanism and Bi-directional long short-term memory network[D]. Tianjin:Tianjin University, 2019:1.