针对SCR入口NOx浓度单一预测模型无法满足在不同工况下保持预测精度的问题,提出了一种基于蜣螂优化(dung beetle optimizer,DBO)集成模型加权融合的预测SCR入口NOx浓度的动态模型。首先使用CatBoost与LightGBM的混合模型在筛选辅助变量的同时,求取辅助变量的迟延时间和阶次信息,并根据以上信息确定预测模型的输入变量;然后建立由LightGBM,XGBoost与CatBoost组成的集成模型,并使用蜣螂优化算法对预测结果进行加权融合;最后将DBO-集成加权融合动态预测模型与3种单模型和蜣螂算法优化2种模型加权融合的预测模型进行对比。结果证明DBO综合加权融合动态预测模型的评价指标优于其他模型,具有更高的预测精度、实时性和适应性,能够更好地满足不同工况下的NOx浓度预测要求。
Abstract
Aiming at the problem that a single prediction model for SCR inlet NOx concentration could not maintain prediction accuracy under different operating conditions, a dynamic model for predicting SCR inlet NOx concentration based on weighted fusion of Dung Beetle Optimizer (DBO) ensemble model was proposed. Firstly, a hybrid model of CatBoost and LightGBM was used to filter auxiliary variables while obtaining the delay time and order information of the auxiliary variables, and the input variables of the prediction model were determined based on the above information. Then, an integrated model consisting of LightGBM,XGBoost, and CatBoost was established, and the prediction results were weighted and fused using the dung beetle optimization algorithm. Finally, the DBO integrated weighted fusion dynamic prediction model was compared with three single models and the two weighted fusion prediction models optimized by the dung beetle algorithm. The evaluation indicators of the DBO integrated weighted fusion dynamic prediction model were superior to other models, with higher prediction accuracy, real-time performance, and adaptability, which could better meet the requirements of NOx concentration prediction under different working conditions.
关键词
化学计量;NOx排放预测 /
蜣螂优化算法 /
CatBoost /
LightGBM /
XGBoost /
集成模型
Key words
stoichiometry;NOx emission prediction /
dung beetle optimization algorithm /
CatBoost /
LightGBM;XGBoost /
Integrated model
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］袁家海, 苗若兰, 张健. 碳中和背景下我国火电行业低碳发展展望 ［J］. 中国国情国力, 2022 (12): 9-13.
YUAN J H, MIAO R L, ZHANG J. Prospects for the low-carbon development of Chinas thermal power industry in the context of carbon neutrality ［J］. Chinas national conditions and strength, 2022 (12): 9-13.
［2］周洪煜, 张振华, 张军, 等. 超临界锅炉烟气脱硝喷氨量混结构-径向基函数神经网络最优控制 ［J］. 中国电机工程学报, 2011, 31 (5): 108-113.
ZHOU H Y, ZHANG Z H, ZHANG J, et al. Optimal control of ammonia injection volume mixed structure Radial basis function neural network for Supercritical steam generator flue gas denitrification ［J］. Chinese Journal of Electrical Engineering, 2011, 31 (5): 108-113.
［3］谭增强, 李元昊, 牛拥军,等. 火电厂NOx的均布优化及预测研究 ［J］. 环境工程, 2023, 41 (S1): 349-353.
TAN Z Q, LI Y H, NIU Y J, et al. Optimization and Prediction of NOx Distribution in Thermal Power Plants ［J］. Environmental engineering, 2023, 41 (S1): 349-353.
［4］姚宁, 金秀章, 李阳峰. 基于改进鲸鱼算法优化Bi-LSTM的脱硝系统NOx建模 ［J］. 华北电力大学学报 (自然科学版), 2022, 49 (6): 76-83.
YAO N, JIN X Z, LI Y F. NOx modeling of denitrification system optimized by Bi-LSTM based on improved whale algorithm ［J］. Journal of North China Electric Power University (Natural Science Edition), 2022, 49 (6): 76-83.
［5］邢红涛, 郭江龙, 张颖, 等. 基于变量选择和XGBoost组合模型的NOx排放预测 ［J］. 自动化与仪器仪表, 2021(7): 21-25.
XING H T, GUO J L, ZHANG Y, et al. NOx emission prediction based on variable selection and XGBoost combination model ［J］. Automation and Instrumentation, 021, (7): 21-25.
［6］吴康洛, 黄俊, 李峥辉, 等. 基于MIC-CFS-LSTM的SCR出口NOx浓度动态预测 ［J］. 洁净煤技术, 2023, 29 (6): 142-150.
WU K L, HUANG J, LI Z H, et al. Dynamic prediction of NOx concentration at SCR outlet based on MIC-CFS-LSTM ［J］. Clean Coal Technology, 2023, 29 (6): 142-150.
［7］唐振浩, 张佳宁, 沈涛. 一种基于卷积特征NOx排放浓度深度预测模型 ［J］. 中国电机工程学报, 2023, 43 (21): 8356-8366.
TANG Z H, ZHANG J N, SHEN T. A deep prediction model for NOx emission concentration based on convolutional features ［J］. Chinese Journal of Electrical Engineering, 2023, 43 (21): 8356-8366.
［8］GUO L, ZHANG S, HUANG Q. NOx prediction of gas turbine based on Dual Attention and LSTM［C］//2022 34th Chinese Control and Decision Conference (CCDC), Hefei, China, 2022.
［9］ZHAO Z, MA Y. Research on Multi-step Prediction of Inlet NOx Concentration Based on VMD-ARIMA Model［C］//2021 40th Chinese Control Conference (CCC), Shanghai, China, 2021.
［10］贾颖, 赵峰, 李博, 等. 贝叶斯优化的XGBoost信用风险评估模型 ［J］. 计算机工程与应用, 2023: 1-15.
JIA Y, ZHAO F, LI B, et al. Bayesian optimized XGBoost credit risk assessment model ［J］. Computer engineering and Application, 2023: 1-15.
［11］ZHANG W, WU C, TANG L, et al. Efficient time-variant reliability analysis of Bazimen landslide in the Three Gorges Reservoir Area using XGBoost and LightGBM algorithms ［J］. Gondwana Research, 2023, 123: 41-53.
［12］SHAKEEL A, CHONG D, WANG J. District heating load forecasting with a hybrid model based on LightGBM and FB-prophet ［J］. Journal of Cleaner Production, 2023, 409: 137130.
［13］孙胡彬. 基于数据挖掘的电站锅炉NOx预测与配煤优化研究［D］. 浙江大学, 2022.
［14］蔡明, 孙杰, 李培德, 等. 三种机器学习算法在回归应用中的对比分析 ［J］. 智能计算机与应用, 2022, 12 (8): 165-170.
CAI M, SUN J, LI P D, et al. Comparative analysis of three machine learning algorithms in regression applications ［J］. Intelligent Computers and Applications, 2022, 12 (8): 165-170.
［15］WEI N, ZHANG Q, ZHANG Y, et al. Super-learner model realizes the transient prediction of CO2 and NOx of diesel trucks: Model development,evaluation and interpretation ［J］. Environment International, 2022, 158: 106977.
［16］符鑫隆, 林姗, 牛辉, 等. 基于CatBoost的患者住院优先级预测模型 ［J］. 信息化研究, 2023, 49 (1): 43-47.
FU X L, LIN S, NIU H, et al. Patient hospitalization priority prediction model based on CatBoost ［J］. Information Technology Research, 2023, 49 (1): 43-47.
［17］XUE J, SHEN B. Dung beetle optimizer: a new meta-heuristic algorithm for global optimization ［J］. Journal of Supercomputing, 2023, 79 (7): 7305-7336.
［18］潘劲成, 李少波, 周鹏, 等. 改进正弦算法引导的蜣螂优化算法 ［J］. 计算机工程与应用, 2023, 59 (22): 92-110.
PAN J C, LI S B, ZHOU P, et al. Improved Sine Algorithm Guided Dung Beetle Optimization Algorithm ［J］. Computer engineering and Application, 2023, 59 (22): 92-110.
［19］程楠楠. 基于混合特征选择模型CatBoost-LightGBM的违约风险预测研究 ［J］. 现代信息科技, 2021, 5 (14): 116-120.
CHENG N N. Research on default risk prediction based on mixed Feature selection model CatBoost-LightGBM ［J］. Modern Information Technology, 2021, 5 (14): 116-120.
［20］肖鹏飞. 基于LightGBM与Catboost融合的二手车估价模型研究［D］. 武汉:中南财经政法大学, 2021.
［21］金秀章,陈佳政,李阳峰. 基于ARIMA-OSELM的火电厂SCR入口NOx浓度预测建模研究［J］. 计量学报, 2023, 44(9): 1458-1466.
JIN X Z, CHEN J Z, LI Y F. SCR Inlet of Thermal Power Plant Based on ARIMA-OSELM Prediction of NOx Concentration［J］. Acta Metrologica Sinica, 2023, 44(9): 1458-1466.
［22］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 Z, LIY N, YUAN H. Dynamic Soft Sensor Model for NOx Generation in Coal-fired Units ［J］. Journal of Power Engineering, 2020, 40 (7): 523-529.
PDF(3110 KB)
