基于改进LSTM与注意力机制的SCR入口NO<sub><i>x</i></sub>生成浓度预测

赵旭东; 李军; 王林; 高林; 弓林娟; 高艺轩; 梁永吉; 杨大锚; 王文毓

doi:10.19805/j.cnki.jcspe.2025.250403

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于改进LSTM与注意力机制的SCR入口NO_x生成浓度预测

更新时间：2026-10-10

- 基于改进LSTM与注意力机制的SCR入口NO_x生成浓度预测
- Prediction of NO_x Concentration at the SCR Inlet Based on Improved LSTM with Attention Mechanism
- 动力工程学报 2025年45卷第12期页码：2044-2055
- 作者机构：
  
  1. 西安热工研究院有限公司,陕西,西安,710054
  2. 高效灵活煤电及碳捕集利用封存全国重点实验室,北京,100031
  3. 北京交通大学(威海) 经济管理学院,山东,威海,264401
  4. 华能铜川照金煤电有限公司,陕西,铜川,727100
- 作者简介：
  
  赵旭东(2001—),男,山东德州人,硕士研究生,研究方向为智慧电厂,E-mail:zhaoxudong@tpri.com.cn
- 基金信息：
  
  国家重点研发计划资助项目(2022YFB4100700);中国华能集团有限公司重点科技项目(HNKJ24-H136,HNKJ25-H22)
- DOI：10.19805/j.cnki.jcspe.2025.250403
  中图分类号：
- 网络首发：2025-12-16，
  
  纸质出版：2025-12-16
- 稿件说明：
移动端阅览
赵旭东,李军,王林,高林,弓林娟,高艺轩,梁永吉,杨大锚,王文毓. 基于改进LSTM与注意力机制的SCR入口NO_x生成浓度预测动力工程学报, 2025, 45(12): 2044-2055 https://doi. DOI: 10.19805/j.cnki.jcspe.2025.250403.

org/10.19805/j.cnki.jcspe.2025.250403
赵旭东,李军,王林,高林,弓林娟,高艺轩,梁永吉,杨大锚,王文毓. 基于改进LSTM与注意力机制的SCR入口NO_x生成浓度预测动力工程学报, 2025, 45(12): 2044-2055 https://doi. DOI: 10.19805/j.cnki.jcspe.2025.250403. DOI：

org/10.19805/j.cnki.jcspe.2025.250403 DOI：

摘要

为解决火电厂燃煤锅炉NO

生成机理复杂

选择性催化还原(SCR)系统喷氨导致的难以及时、精准调控的问题

提出了一种基于改进的长短时记忆网络(LSTM)与注意力机制的SCR入口NO

生成浓度预测方法。首先

通过锅炉燃烧NO

生成机理分析确定初始特征变量集

在原始运行数据经过预处理后

利用最大信息系数确定各变量的延迟时间

并构建时滞补偿数据集;其次

基于灰色关联度分析(GRA)对影响NO

生成的关键变量进行排序

从而实现特征选择。最后

提出了一种融合改进LSTM与注意力机制的新型时间序列预测模型ADS-Forecaster

采用双路径特征编码及自注意力融合解码

实现NO

生成浓度实时预测。基于某600 MW亚临界燃煤机组运行数据的实验结果表明:与基准模型RNN、LSTM及xLSTM相比

ADS-Forecaster模型在泛化能力和预测精度方面具有优势。

Abstract

To address the issues of complex NO

generation mechanisms in coal-fired boilers of thermal power plants

and the difficulty in timely and precise regulation caused by ammonia injection into selective catalytic reduction (SCR) systems

a prediction method for NO

generation concentration at the SCR system inlet based on an improved long short-term memory network (LSTM) and attention mechanism was proposed. Firstly

the initial feature variable set was determined through mechanistic analysis of NO

generation during boiler combustion. Following the preprocessing of the raw operational data

the maximum information coefficient was used to ascertain

the delay time for each variable

thereby constructing a time-lag compensated data set. Secondly

based on grey relational analysis (GRA)

the key variables affecting NO

formation were ranked to achieve feature selection. Finally

a novel time-series forecasting model ADS-Forecaster integrated the improved LSTM with an attention mechanism was proposed

which used dual-path feature encoder and self-attention fusion decoder to realize real-time prediction of NO

concentration. The test results based on operational data from a 600 MW subcritical coal-fired unit show that compared with benchmark models (RNN

LSTM and xLSTM)

ADS-Forecaster model achieves superior generalization capability and prediction accuracy.

关键词

Keywords

references

DIRIK M. Prediction of NO x emissions from gas turbines of a combined cycle power plant using an ANFIS model optimized by GA[J ] . Fuel, 2022, 321: 124037.

孙胡彬, 杨建国, 金宏伟, 等. 基于贝叶斯优化-随机森林回归的燃煤锅炉NO x 预测模型[J ] . 动力工程学报, 2023, 43(7): 910-916. SUN Hubin, YANG Jianguo, JIN Hongwei, et al. NO x prediction model for coal-fired boiler based on Bayesian optimization-random forest regression[J ] . Journal of Chinese Society of Power Engineering, 2023, 43(7): 910-916.

林伟俊, 李敏, 赵畅, 等. 基于延迟特性的NO x 浓度软测量与喷氨优化[J ] . 动力工程学报, 2023, 43(10): 1352-1362. LIN Weijun, LI Min, ZHAO Chang, et al. NO x concentration soft measurement and ammonia injection optimization based on delay feature[J ] . Journal of Chinese Society of Power Engineering, 2023, 43(10): 1352-1362.

王婷. 模型预测控制在电厂选择性催化还原脱硝系统中的应用研究[D]. 杭州: 浙江大学, 2015.

王焕敏, 王沈振, 唐亮, 等. 基于强化学习的SCR脱硝系统优化控制[J]. 动力工程学报, 2024, 44(12): 1916-1922, 1934. WANG Huanmin, WANG Shenzhen, TANG Liang, et al. Optimal control of SCR denitrification system based on reinforcement learning[J]. Journal of Chinese Society of Power Engineering, 2024, 44(12): 1916-1922, 1934.

LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444.

WU Zheng, ZHANG Yue, DONG Ze. Prediction of NO x emission concentration from coal-fired power plant based on joint knowledge and data driven[J ] . Energy, 2023, 271: 127044.

温鑫, 钱玉良, 彭道刚, 等. 基于GA-GRNN数据挖掘的SCR脱硝系统建模优化研究[J]. 上海电力大学学报, 2020, 36(2): 161-167. WEN Xin, QIAN Yuliang, PENG Daogang, et al. Research on modeling optimization of SCR denitration system based on GA-GRNN data mining[J]. Journal of Shanghai University of Electric Power, 2020, 36(2): 161-167.

杨国田, 王英男, 谢锐彪, 等. 基于变量选择的深度置信神经网络锅炉烟气NO x 排放预测[J ] . 热力发电, 2020, 49(4): 34-40. YANG Guotian, WANG Yingnan, XIE Ruibiao, et al. Prediction of NO x emission from boiler based on DBN with variable selection[J ] . Thermal Power Generation, 2020, 49(4): 34-40.

余廷芳, 张浩杰. 基于SVM和RBF神经网络的CFB NO x 生成预测模型[J ] . 计算机仿真, 2020, 37(9): 209-213, 316. YU Tingfang, ZHANG Haojie. NO x generation prediction model of coal-fired boiler based on SVM and RBF neural network[J ] . Computer Simulation, 2020, 37(9): 209-213, 316.

印江, 王尚尚, 李丽锋, 等. 基于IPSO-BP算法的CFB锅炉NO x 浓度预测[J ] . 自动化与仪表, 2021, 36(2): 58-63. YIN Jiang, WANG Shangshang, LI Lifeng, et al. NO x concentration prediction of CFB boiler based on IPSO-BP algorithm[J ] . Automation & Instrumentation, 2021, 36(2): 58-63.

TANG Zhenhao, WANG Shikui, CHAI Xiangying, et al. Auto-encoder-extreme learning machine model for boiler NO x emission concentration prediction[J ] . Energy, 2022, 256: 124552.

董渊博, 茅大钧, 章明明. 基于CNN-LSTM的燃气轮机NO x 排放预测研究[J ] . 热能动力工程, 2021, 36(9): 132-138. DONG Yuanbo, MAO Dajun, ZHANG Mingming. Research on NO x emission prediction of gas turbine based on CNN-LSTM[J ] . Journal of Engineering for Thermal Energy and Power, 2021, 36(9): 132-138.

WANG Yingnan, YANG Guotian, XIE Ruibiao, et al. An ensemble deep belief network model based on random subspace for NO x concentration prediction[J ] . ACS Omega, 2021, 6(11): 7655-7668.

HAO Xiaochen, XU Qingquan, SHI Xin, et al. Prediction of nitrogen oxide emission concentration in cement production process: a method of deep belief network with clustering and time series[J]. Environmental Science and Pollution Research, 2021, 28(24): 31689-31703.

WAQAS SAIF-UL-ALLAH M, KHAN J, AHMED F, et al. Computationally inexpensive 1D-CNN for the prediction of noisy data of NO x emissions from 500 MW coal-fired power plant[J ] . Frontiers in Energy Research, 2022, 10: 945769.

许子明, 姜浩, 赵文杰. 基于IHHO-LSTM的SCR脱硝反应器出口NO x 浓度预测[J ] . 电力科学与工程, 2023, 39(8): 71-78. XU Ziming, JIANG Hao, ZHAO Wenjie. Prediction of NO x concentration at the outlet of SCR denitration reactor based on IHHO-LSTM[J ] . Electric Power Science and Engineering, 2023, 39(8): 71-78.

郭浩楠, 钱进, 朱道兴, 等. 基于MI-GA-LSTM的锅炉NO x 排放快速预测模型[J ] . 热能动力工程, 2023, 38(8): 103-109. GUO Haonan, QIAN Jin, ZHU Daoxing, et al. Fast prediction model for NO x emission of the utility boiler furnace via MI-GA-LSTM[J ] . Journal of Engineering for Thermal Energy and Power, 2023, 38(8): 103-109.

刘建军, 赵旭, 张卫东, 等. 基于CNN(1D)-LSTM模型的电站锅炉SCR入口NO x 浓度预测[J ] . 电子测量技术, 2023, 46(13): 59-65. LIU Jianjun, ZHAO Xu, ZHAN G Weidong, et al. Prediction of NO x concentration at SCR inlet of utility boiler based on CNN(1D)-LSTM model[J ] . Electronic Measurement Technology, 2023, 46(13): 59-65.

HU Zongyang, FANG Jiuwen, ZHENG Ruixiang, et al. Efficient model predictive control of boiler coal combustion based on NARX neutral network[J]. Journal of Process Control, 2024, 134: 103158.

钱虹, 王海心, 徐邦智. 基于DRNN-LSTM-IPSO的锅炉经济运行优化目标值[J]. 科学技术与工程, 2024, 24(7): 2749-2758. QIAN Hong, WANG Haixin, XU Bangzhi. Target value of boiler economic operation optimization based on DRNN-LSTM-IPSO[J]. Science Technology and Engineering, 2024, 24(7): 2749-2758.

XIE Peiran, GAO Mingming, ZHANG Hongfu, et al. Dynamic modeling for NO x emission sequence prediction of SCR system outlet based on sequence to sequence long short-term memory network[J ] . Energy, 2020, 190: 116482.

冯旭刚, 文作银, 章家岩, 等. 基于KPCA-AGRU神经网络的火电机组NO x 排放预测[J ] . 重庆工商大学学报(自然科学版), 2023, 40(6): 18-24. FENG Xugang, WEN Zuoyin, ZHANG Jiayan, et al. Prediction of NO x emissions from thermal power units based on KPCA-AGRU neural network[J ] . Journal of Chongqing Technology and Business University (Natural Science Edition), 2023, 40(6): 18-24.

BECK M, PPPEL K, SPANRING M, et al. xLSTM: extended long short-term memory[C]//Proceedings of the 38th International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc., 2024: 3417.

SU Jianlin, AHMED M, LU Yu, et al. RoFormer: enhanced transformer with rotary position embedding[J]. Neurocomputing, 2024, 568: 127063.

ZHANG Zijun. Improved Adam optimizer for deep neural networks[C]//2018 IEEE/ACM 26th International Symposium on Quality of Service (IWQoS). Banff: IEEE, 2018: 1-2.

张美娜. 基于灰色关联分析的多目标智能优化算法[D]. 长春: 长春师范大学, 2022.

浏览量

265

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于特征相似日和CNN-BiLSTM-Attention模型的风电短期出力预测

基于CNN-BiLSTM-SSA的锅炉再热器壁温预测模型

基于模态分解和LSTM-IDBO-GRU的光伏功率预测研究

基于多尺度注意力域适应网络的跨域转子故障迁移诊断方法

融合时空特性的改进Informer风电功率短期预测

基于改进LSTM与注意力机制的SCR入口NOx生成浓度预测

Prediction of NOx Concentration at the SCR Inlet Based on Improved LSTM with Attention Mechanism

DOI：10.19805/j.cnki.jcspe.2025.250403

摘要

Abstract

关键词

Keywords

references

基于改进LSTM与注意力机制的SCR入口NO_x生成浓度预测

Prediction of NO_x Concentration at the SCR Inlet Based on Improved LSTM with Attention Mechanism