LU Jun, ZHANG Guangcai, XU Dangqi, et al. Application of modified BP neural network in optimization of low NOx combustion for a pulverized coal boiler[J]. Thermal power generation, 2015, 44(4): 106-111.
LU Jun, ZHANG Guangcai, XU Dangqi, et al. Application of modified BP neural network in optimization of low NOx combustion for a pulverized coal boiler[J]. Thermal power generation, 2015, 44(4): 106-111.DOI:
改进BP神经网络在低氮燃烧优化中的应用
摘要
基于燃煤锅炉低氮燃烧优化试验数据
建立了预测NOx排放浓度和锅炉效率的改进BP神经网络模型
通过样本学习
该网络可以精确描述锅炉运行参数与NOx排放浓度和锅炉效率之间的非线性映射关系。仿真结果表明
网络的NOx排放浓度和锅炉效率预测值的相对误差分别小于3.92%和7.6%
能够描述NOx排放浓度和锅炉效率随SOFA风门开度和运行氧量变化的规律
而且预测精度高、泛化能力强
可为燃煤锅炉低氮燃烧优化提供指导。
Abstract
On the basis of field test data during optimization of low NOxcombustion for a pulverized coal boiler
a modified BP neural network for prediction of NOxemission and boiler efficiency was established.After learning the training samples
the BP neural network can describe the non-linear relationship between the operation parameters and NOxemission and boiler efficiency.The simulation results show that the relative error of prediction values of NOxemission and boiler efficiency is less than 3.92% and 7.6%
respectively
indicating the neural network can describe the change rules of NOxemission and boiler efficiency against the SOFA air damper opening degree and oxygen content.This modified BP neural network has high prediction accuracy and stable generalization ability
which can provide guidance for optimization of low NOxcombustion for pulverized coal boilers.