State Key Laboratory of Alternate Electrical Power System With Renewable Energy Sources,North China ElectricPower University
Department of Energy and Power Engineering,Tsinghua University

为响应碳达峰,碳中和目标,我国循环流化床锅炉大规模参与深度调峰运行,导致锅炉NOx排放浓度波动范围大,控制效果不佳,难以满足污染物超低排放需求,因此对深度调峰NOx排放浓度进行精准建模预测有重要意义。以即燃碳模型为基础,深度剖析炉内NOx生成和还原机理,建立炉内即燃碳燃烧模型、O2动态平衡模型、CO软测量模型、NOx生成与还原模型,完成SNCR入口NOx浓度机理计算;选取给煤量、床温、烟气温度及含氧量、一二次风量、尿素溶液流量作为NOx排放浓度的输入变量,将SNCR入口NOx浓度计算值作为拓展输入变量,对所有输入变量与NOx排放浓度进行相关性分析和迟延补偿,完成数据集重构;采用长短期记忆神经网络对重构数据集进行训练和预测,并将鲸鱼优化算法用于长短期记忆神经网络的参数优化,建立循环流化床锅炉深度调峰NOx排放浓度机理———数据混合预测模型。仿真验证表明混合预测模型不同工况下预测性能和泛化能力好,能够实现循环流化床锅炉变负荷时NOx排放浓度的实时预测,相较其他预测模型的各项误差性能指标均显著提升,平均绝对误差δMAE达2.14mg/m3,平均相对百分误差δMAPE达5.68%,决定系数R2达0.9021。混合预测模型能精准预测循环流化床锅炉深度调峰下NOx排放浓度,为循环流化床锅炉超低排放智能控制系统的设计提供参考。

In response to the goal of Carbon peak Carbon neutral,China′s circulating fluidized bed boilers participate in deep peaking operation on a large scale,resulting in large fluctuation ranges of NO emission concentration in boilers,poor control effect,and difficultyin meeting the demand for ultra-low emission of pollutants,so it is important to accurately model and predict the NO emission concentration in deep peaking. Based on the instantaneous carbon model,the NO generation and reduction mechanism in the furnace was deeply analyzed,and the instantaneous carbon combustion model,O dynamic balance model,CO soft measurement model,NO generation and reduction model were established to complete the calculation of the mechanism of the NO concentration at the entrance of the SNCR. Theamount of coal feed,bed temperature,flue gas temperature and oxygen content,the first and second airflow,and the flow rate of the urea solution were selected as the input variables for the NO emission concentration, and the NO emission concentration was predicted bythe SNCR inlet model. The SNCR inlet NO concentration was used as an extended input variable,and the data set was reconstructedby correlation analysis and delay compensation between all input variables and NO emission concentration. The reconstructed data set wastrained and predicted by using long and short-term memory neural network,and whale optimization algorithm was used for the optimizationof parameters of the long and short-term memory neural network to establish a NO emission concentration model,the mechanism-data hybrid prediction model,for deep peaking of circulating fluidized bed boilers. The simulation validation shows that the hybrid prediction model has good prediction performance and generalization ability under different working conditions,and is able to realize real-time predictionof NO emission concentration in circulating fluidized bed boilers at variable loads,and significantly improves all the error performance indexes compared with other prediction models,with an average absolute error δ up to 2.14 mg/ m,an average relative percentage errorδ up to 5.68%,and a coefficient of determination R up to 0.902 1. The hybrid prediction model can accurately predict the NO emission concentration under deep peaking in circulating fluidized bed boilers,which provides a reference for the design of the ultra-low emission intelligent control system of circulating fluidized bed boilers.

国家重点研发计划资助项目(2022YFB4100304)

张鹏新,高明明,郭炯楠,等.基于混合建模方法循环流化床锅炉深度调峰NOx排放预测[J].洁净煤技术,2024,30(9):85-94.

ZHANG Pengxin,GAO Mingming,GUO Jiongnan,et al.Prediction of NOx emissions from deep peaking circulating fluidizedbed boilers based on a hybrid modelling approach[J].Clean Coal Technology,2024,30(9):85-94.